JP7656662B2 - Point cloud information conversion device, point cloud information conversion method, and program - Google Patents

Description

The present invention relates to a point cloud information conversion device, a point cloud information conversion method, and a program.

Technology is known for estimating the shape of an object based on point cloud data acquired by LiDAR (Light Detection and Ranging) measurement or the like. For example, a DSM (Digital Surface Model) and a DEM (Digital Elevation Model) that indicate the shape of features and the earth's surface are generated based on point cloud data acquired by airborne laser surveying. In estimating shape based on point cloud data, it is necessary to appropriately classify the point cloud according to the type of object indicated by the point cloud data. Patent Document 1 describes a method for extracting a road surface candidate point cloud that indicates a road surface from point cloud data.

JP 2018-173749 A

Point clouds are classified by setting a class for the type of object (e.g., ground surface, building, noise, etc.) that each constituent point represents for each of the constituent points that make up the point cloud data. The classes are set automatically by a device such as that described in Patent Document 1, and an operator checks whether the classes have been set appropriately. The classes may also be set entirely manually by an operator. In either case, the point cloud data must be handed over between the administrator of the point cloud data and the operator who sets the classes.

Point cloud data has a large data volume, and it takes a long time to transmit it over a network. Therefore, there is a need to reduce the time required to send and receive data between the operator who sets the point cloud classes.

For example, in general, there is a spatial uneven distribution of class information of constituent points contained in point cloud data, such as the fact that a constituent point of a building class has many other constituent points of the building class in the vicinity of the other constituent points. The inventors have focused on this spatial uneven distribution of class information and invented a method to reduce the amount of data transmitted and received by omitting class information for some constituent points.

The present invention aims to provide a point cloud information conversion device, a point cloud information conversion method, and a program that can reduce the time required to send and receive information indicating a point cloud class.

The point cloud information conversion device according to an embodiment of the present invention is characterized by having an acquisition unit that acquires, in a predetermined coordinate space, point cloud information including coordinate information indicating the positions of a plurality of constituent points that indicate the shape of the earth's surface and features, and class information indicating the classification result of each constituent point based on the type of thing that each constituent point indicates, a setting unit that sets a plurality of areas in the coordinate space, and a conversion unit that selects, from the plurality of areas, a single-class area in which all of the constituent points included therein have the same class, and converts the class information into area class information that does not include class identification information of each constituent point included in the single-class area, and in which the identification information of the single-class area corresponds to the identification information of the class of the constituent points included in the single-class area.

The conversion unit preferably selects a multi-class area containing two or more types of class of constituent points from among the multiple areas, identifies the area most frequent class having the largest number of corresponding constituent points from among the classes of the constituent points contained in the multi-class area, and converts the class information such that the class identification information of the constituent points of the area most frequent class from among the constituent points contained in the multi-class area is omitted.

It is also preferable that the setting unit sets the multiple regions such that the multiple regions include at least one single-class region.

In addition, it is preferable that the setting unit identifies the most frequent class corresponding to the largest number of constituent points from among the classes of the multiple constituent points, calculates an approximation plane of the constituent points of the most frequent class, and sets the multiple regions by calculating a plane parallel to the approximation plane as a separation plane that separates the multiple regions.

In addition, it is preferable that the setting unit calculates, as the separation plane, a plane that passes through the component point of a class other than the most frequent class that is closest to the approximation plane in a predetermined coordinate axis direction of the coordinate space and is parallel to the approximation plane.

In addition, when the inclination of the calculated separation plane satisfies a predetermined condition, it is preferable that the setting unit calculates a horizontal plane as the separation plane instead of the calculated separation plane.

It is also preferable that the point cloud information conversion device further includes a classification unit that classifies the constituent points included in the point cloud information into a plurality of groups based on the coordinate information of each constituent point, the setting unit sets a plurality of regions in each group, and the conversion unit identifies a single-class region for each group and converts the class information of the constituent points included in the single-class region into region class information.

The point cloud information conversion method according to an embodiment of the present invention is characterized by including the steps of: acquiring point cloud information in a predetermined coordinate space in which coordinate information indicating the positions of a plurality of constituent points indicating the shape of the Earth's surface and features is associated with class information indicating the classification result of each constituent point based on the type of object each constituent point indicates; setting a plurality of regions in the coordinate space; selecting a single-class region from among the plurality of regions in which all constituent points included therein have the same class; and converting the class information into region class information that does not include class identification information of each constituent point included in the single-class region and in which the identification information of the single-class region is associated with the identification information of the class of the constituent points included in the single-class region.

A program according to an embodiment of the present invention is characterized in that it causes a computer to execute the following operations: acquire point cloud information in a specified coordinate space in which coordinate information indicating the positions of a plurality of constituent points indicating the shape of the Earth's surface and features is associated with class information indicating the classification result of each constituent point based on the type of object each constituent point indicates; set a plurality of regions in the coordinate space; select from the plurality of regions a single-class region in which all constituent points contained therein have the same class; and convert the class information into region class information that does not include class identification information of each constituent point contained in the single-class region and in which the identification information of the single-class region is associated with the identification information of the class of the constituent points contained in the single-class region.

The point cloud information conversion device, point cloud information conversion method, and program of the present invention make it possible to reduce the time required to send and receive information indicating the class of a point cloud.

1 is a schematic diagram showing an example of a schematic configuration of a transmission/reception system 1. FIG. FIG. 2 is an exemplary functional block diagram of a management device 2. FIG. 2 is an exemplary functional block diagram of the editing device 3. 4 is a diagram showing an example of the data structure of a point cloud information table T1. 13 is a diagram illustrating an example of the data structure of a class information table T2. 11 is a sequence diagram illustrating an example of the flow of a class information setting process. FIG. 11 is a flow diagram showing an example of the flow of a conversion process. FIG. 13 is a schematic diagram for explaining an example of classification of composition points. FIG. 13 is a schematic diagram for explaining a result of rearrangement of configuration points. FIG. 11 is a schematic diagram for explaining coding class information. FIG. 11 is a flowchart showing an example of the flow of an area setting process. FIG. 11 is a schematic diagram for explaining normalization of coordinate information. 11A and 11B are schematic diagrams for explaining calculation of a lower separation surface and an upper separation surface. 13 is a schematic diagram for explaining calculation of an additional separation surface. FIG. FIG. 2 is a diagram illustrating an example of the data structure of converted data D generated by a conversion process. 11 is a diagram showing an example of the data structure of information E regarding the classes of configuration points included in a multiple class area. FIG. FIG. 11 is a flow diagram showing an example of the flow of an association process. 11A and 11B are schematic diagrams for explaining calculation of a lower separation surface and an upper separation surface having mutually different inclinations.

Various embodiments of the present invention will be described below with reference to the drawings. Please note that the technical scope of the present invention is not limited to these embodiments, but extends to the inventions described in the claims and their equivalents.

FIG. 1 is a diagram showing an example of a schematic configuration of a transmission/reception system 1 according to an embodiment of the present invention. The transmission/reception system 1 transmits and receives class information between a management base that manages point cloud information and an editing base that sets class information indicating the classification result of the constituent points included in the point cloud information. The point cloud information includes coordinate information indicating the positions of multiple constituent points that indicate the shape of the earth's surface and buildings in a specified coordinate space. The class information indicates the classification result of each constituent point based on the type of object that each constituent point indicates. The transmission/reception system 1 has a management device 2 installed at the management base and an editing device 3 installed at the editing base. The management device 2 and the editing device 3 communicate with each other via a network N, which is the Internet or an intranet. The management device 2 and the editing device 3 are examples of a receiving device and a transmitting device.

FIG. 2 is an exemplary functional block diagram of the management device 2. The management device 2 is, for example, a computer such as a server, a PC (Personal Computer), a tablet terminal, or a smartphone. The management device 2 transmits point cloud information to the editing device 3. The management device 2 also receives class information indicating the class of the constituent points included in the point cloud information from the editing device 3, and associates the class information with the constituent points. The management device 2 has a storage unit 21, a communication unit 22, and a processing unit 23.

The storage unit 21 is configured to store programs and data, and includes, for example, a semiconductor memory. The storage unit 21 stores, as programs, operating system programs, driver programs, application programs, etc., used in processing by the processing unit 23. Programs are installed in the storage unit 21 from computer-readable, non-transient portable storage media such as CD-ROMs (Compact Disc Read Only Memory) and DVD-ROMs (Digital Versatile Disc Read Only Memory).

The communication unit 22 is configured to enable the management device 2 to communicate with other devices, and includes a communication interface circuit. The communication interface circuit is a communication interface circuit for a wired LAN (Local Area Network) or a wireless LAN. The communication unit 22 receives data from other devices and supplies it to the processing unit 23, and transmits data supplied from the processing unit 23 to other devices.

The processing unit 23 is configured to comprehensively control the operation of the management device 2, and includes one or more processors and their peripheral circuits. The processing unit 23 includes, for example, a CPU (Central Processing Unit). The processing unit 23 may include a GPU (Graphics Processing Unit), a DSP (Digital Signal Processor), an LSI (Large Scale Integration), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), etc. The processing unit 23 controls the operation of each component and executes various processes so that the various processes of the management device 2 are executed in an appropriate order based on the programs stored in the memory unit 21.

The processing unit 23 has a transmitting unit 231, a receiving unit 232, and a matching unit 233 as functional blocks. These functional blocks are functional modules that are realized by the processing unit 23 executing a program. Each of these units may be implemented in the management device 2 as a dedicated processing circuit.

FIG. 3 is an exemplary functional block diagram of the editing device 3. The editing device 3 is, for example, a computer such as a PC, a tablet terminal, or a smartphone. The editing device 3 receives point cloud information from the management device 2. The editing device 3 also sets class information indicating the class of the constituent points included in the point cloud information, and transmits the class information to the management device 2. The editing device 3 has a memory unit 31, a communication unit 32, a display unit 33, an operation unit 34, and a processing unit 35.

The storage unit 31 is configured to store programs and data, and includes, for example, a semiconductor memory. The storage unit 31 stores, as programs, operating system programs, driver programs, application programs, etc., used in processing by the processing unit 35. Programs are installed into the storage unit 31 from computer-readable, non-transitory, portable storage media such as CD-ROMs and DVD-ROMs.

The communication unit 32 is configured to enable the editing device 3 to communicate with other devices, and includes a communication interface circuit. The communication interface circuit is a communication interface circuit such as a wired LAN (Local Area Network) or a wireless LAN. The communication unit 32 receives data from other devices and supplies it to the processing unit 35, and transmits data supplied from the processing unit 35 to other devices.

The display unit 33 is configured to display an image, and includes a display such as a liquid crystal display or an organic EL (Electro-Luminescence) display. The display unit 33 displays an image based on the display data supplied from the processing unit 35.

The operation unit 34 is configured to accept operations on the editing device 3, and includes input devices such as a keyboard, a keypad, and a mouse. The operation unit 34 may include a touch panel integrated with the display unit 33. The operation unit 34 supplies a signal corresponding to the accepted operation to the processing unit 35.

The processing unit 35 is configured to comprehensively control the operation of the editing device 3, and includes one or more processors and their peripheral circuits. The processing unit 35 includes, for example, a CPU. The processing unit 23 may include a GPU, a DSP, an LSI, an ASIC, an FPGA, etc. The processing unit 35 controls the operation of each component and executes various processes so that the various processes of the editing device 3 are executed in an appropriate order based on the programs stored in the memory unit 31.

The processing unit 35 has a receiving unit 351, a generating unit 352, a transmitting unit 353, a classifying unit 354, a setting unit 355, and a converting unit 356 as functional blocks. These functional blocks are functional modules that are realized by the processing unit 35 executing a program. These functional blocks may be implemented in the editing device 3 as dedicated processing circuits.

Figure 4 is a diagram showing an example of the data structure of the point cloud information table T1. The point cloud information table T1 stores point cloud information. The point cloud information is acquired by LiDAR measurement and stored in the management device 2. The point cloud information is, for example, data that conforms to the LAS file format, but is not limited to this example and may be data in any format, such as text data.

The point cloud information stores a constituent point ID that identifies each of the multiple constituent points, and the coordinate information of each constituent point. Each constituent point corresponds to a reflection point of the laser irradiated in the LiDAR measurement. The constituent point ID is information that identifies each constituent point. The coordinate information is information that indicates the position of the constituent point in the coordinate space. The coordinate space is a space that conforms to a planar rectangular coordinate system or an arbitrary mathematical coordinate system (orthogonal coordinate system). The multiple constituent points included in the point cloud information are ordered based on a predetermined criterion. In the example shown in FIG. 4, the multiple constituent points are ordered based on the constituent point ID.

Figure 5 is a diagram showing an example of the data structure of class information table T2. Class information table T2 stores class information indicating the class of each of multiple constituent points. The class is information indicating the type of object indicated by each constituent point. For example, constituent points associated with class information of the ground and building indicate the shapes of the ground and building, respectively, in coordinate space. Furthermore, constituent points associated with class information of noise are caused by noise such as multipath of the laser or reflection from clouds or water vapor in the air, and do not indicate the shape of a specific object.

The class information is ordered in the same order as the constituent points included in the point cloud information table T4. This allows the management device 2 that has acquired the class information to associate class information with each constituent point included in the point cloud information, even if the class information does not include constituent point IDs and coordinate information. In the example shown in Figure 5, the ground class, building class, building class and noise class are set for the constituent points with constituent point IDs of 1, 2, 3 and 4, respectively.

Figure 6 is a sequence diagram showing an example of the flow of class information setting processing executed by the transmission/reception system 1. The class information setting processing is processing for associating class information with point cloud information. The class information setting processing is realized by the processing units of the management device 2 and the editing device 3 working in cooperation with other components of each device.

First, the transmission unit 231 of the management device 2 transmits point cloud information to the editing device 3 via the communication unit 22 (step S101). The point cloud information is pre-stored in the storage unit 21. The point cloud information stored in the storage unit 21 is an example of first point cloud information.

Next, the receiving unit 351 of the editing device 3 receives the point cloud information via the communication unit 32 (step S102). The receiving unit 351 stores the received point cloud information in the storage unit 31. The point cloud information stored in the storage unit 31 is an example of second point cloud information.

Next, the generating unit 352 of the editing device 3 generates class information (step S103). For example, the generating unit 352 automatically sets a class for each of the multiple constituent points included in the point cloud information according to a predetermined algorithm. The generating unit 352 may display the constituent points included in the point cloud information on the display unit 33, and manually set a class for each constituent point based on an operator's operation on the operation unit 34. The generating unit 352 generates class information indicating the class set for each constituent point, and stores the class information in the storage unit 31.

Next, a conversion process is executed (step S104). The conversion process is a process for generating converted data based on the class information. The converted data is data obtained by losslessly compressing the class information. The details of the conversion process will be described later.

Next, the transmission unit 353 of the editing device 3 transmits the converted data to the management device 2 via the communication unit 32 (step S105).

Next, the receiving unit 232 of the management device 2 receives the converted data via the communication unit 22 (step S106). The receiving unit 232 stores the received converted data in the memory unit 21.

Next, the association unit 233 of the management device 2 executes an association process (step S107). The association process is a process of associating class information with a plurality of configuration points based on the conversion data. The details of the association process will be described later. This completes the class information setting process.

Figure 7 is a flow diagram showing an example of the flow of the conversion process. The conversion process is executed by the editing device 3 in step S104 of the class information setting process.

First, the classification unit 354 acquires point cloud information and class information from the second storage unit 31 (step S201).

Next, the classification unit 354 classifies the constituent points included in the point cloud information into a plurality of groups according to a predetermined rule (step S202). For example, the classification unit 354 classifies the constituent points included in the point cloud information into a plurality of groups based on the position information of each constituent point.

8 is a schematic diagram for explaining an example of classification of constituent points. The classification unit 354 specifies a distribution range R in the horizontal direction of a plurality of constituent points P included in the point cloud information. In the example shown in FIG. 8, the horizontal direction is, for example, a direction along a plane including the X direction and the Y direction in an arbitrary mathematical coordinate system (orthogonal coordinate system). In this case, the classification unit 354 specifies, from among the plurality of constituent points, a constituent point having a minimum X coordinate, a constituent point having a maximum X coordinate, a constituent point having a minimum Y coordinate, and a constituent point having a maximum Y coordinate. The classification unit 354 specifies, as the distribution range R, a rectangular area that extends from the position of the minimum X coordinate to the position of the maximum X coordinate in the X direction and extends from the position of the minimum Y coordinate to the position of the maximum Y coordinate in the Y direction. In the example shown in FIG. 8, a rectangular area whose width in the X direction is Wx and whose width in the Y direction is Wy is specified as the distribution range R.

The classification unit 354 sets a number of buckets, each of which has a predetermined size and shape, in the distribution range. In the example shown in FIG. 8, each bucket is a rectangular area with a width in the X direction of bx and a width in the Y direction of by, and each bucket is arranged in the X direction and the Y direction within the distribution range. The size of each bucket is set in a range in which the shape of the features and the ground surface do not change abruptly within the bucket. For example, bx and by are each set to 10 m or more and 100 m or less. In the example shown in FIG. 8, N buckets are arranged in the X direction and M buckets are arranged in the Y direction. N is the smallest integer equal to or greater than the value obtained by dividing Wx by bx, and M is the smallest integer equal to or greater than the value obtained by dividing Wy by by.

The sorting unit 354 orders the multiple buckets based on their positional relationship. For example, the multiple buckets are ordered so that adjacent buckets in the X or Y direction are adjacent to each other. In the example shown in FIG. 8, the buckets are ordered so that adjacent buckets in the X direction are adjacent to each other. That is, in the example shown in FIG. 8, bucket B1 located at the upper left end of the distribution range R is the first bucket, and buckets B2, ..., BN located to the right of it are ordered so that they are adjacent to each other. Buckets B(N+1), which is adjacent to bucket B1 in the Y direction, are ordered next to BN located at the upper right end of the distribution range R. The subsequent buckets are ordered according to the same rules.

The classification unit 354 determines which of a plurality of buckets each constituent point is included in based on the horizontal position of each constituent point. That is, the classification unit 354 determines in which bucket the horizontal position specified by the X and Y coordinates of each constituent point is located. The classification unit 354 classifies each constituent point into a plurality of groups so that constituent points included in the same bucket are included in the same group. Note that in general point cloud data, each bucket includes several hundred or more constituent points, but for ease of viewing, only a portion of the constituent points are illustrated in FIG. 8.

Returning to FIG. 7, the classification unit 354 then rearranges the constituent points included in the point cloud information based on the coordinate information in accordance with a predetermined rule (step S203). For example, the classification unit 354 classifies the constituent points included in each group into a plurality of small groups by setting small buckets inside each group in the same manner as in step S202. The size of the small buckets is set according to the measurement density of the LiDAR measurement so that the number of constituent points included in each small group is approximately several tens or less. For example, the small buckets are set so that the width in the X direction and the width in the Y direction are 1 m or more and 2 m or less, respectively. The classification unit 354 orders the plurality of small buckets based on their positional relationship, similar to the ordering of the buckets in step S202.

For each small group, the sorting unit 354 sorts the constituent points included in the small group based on the coordinate information. For example, the constituent points included in the small group are sorted in ascending order of X coordinates. If there are multiple constituent points with the same X coordinate, those constituent points are sorted in ascending order of Y coordinates. If there are multiple constituent points with the same X coordinate and Y coordinate, those constituent points are sorted in ascending order of Z coordinates. For example, the sorting unit 354 sorts the constituent points included in the small groups by applying quick sort.

The classification unit 354 rearranges the constituent points included in each group by linking the rearranged constituent points included in each small group within the same group based on the order of each small group. The classification unit 354 rearranges the constituent points included in the point cloud information by linking the rearranged constituent points included in each group based on the order of each group. That is, the classification unit 354 rearranges the constituent points by applying a combination of two-stage bucket sorting and quick sorting. By applying quick sorting, which has a computational complexity of O(NlogN), only to constituent points included in the same small bucket in bucket sorting, which has a computational complexity of O(N), the constituent points are rearranged faster than when quick sorting is applied to all constituent points.

Figure 9 is a schematic diagram for explaining the result of sorting the constituent points. The multiple constituent points included in the point cloud information are sorted based on the order of the groups to which each constituent point belongs. The order of the groups is the order of the buckets set in step S202 corresponding to each group. The multiple constituent points belonging to the same group are sorted based on the order of the small groups to which each constituent point belongs. The order of the small groups is the order of the small buckets corresponding to each small group. The multiple constituent points belonging to the same small group are sorted in ascending order of the X, Y, and Z coordinates.

Returning to FIG. 7, the classification unit 354 then rearranges the class information to correspond to the rearranged configuration points (step S204).

Next, the setting unit 355 selects a group to be processed from among the multiple groups set in step S202 (step S205). For example, the setting unit 355 selects a group to be processed based on the order of the groups.

Next, the setting unit 355 executes an area setting process (step S206). The area setting process is a process for setting multiple areas in the coordinate space based on the coordinate information of the constituent points included in the group to be processed. The area setting process will be described in detail later.

Next, the conversion unit 356 selects an area to be processed from among the multiple areas that have been set (step S207).

Next, the conversion unit 356 determines whether the area to be processed is a single-class area or a multiple-class area (step S208). A single-class area is an area in which, among the configuration points included in the group to be processed, all of the configuration points included are of the same class. A multiple-class area is an area in which, among the configuration points included in the group to be processed, the configuration points included are of two or more different classes.

If the area to be processed is a single-class area (step S208-Yes), the conversion unit 356 converts the class information of the configuration points included in the single-class area into area class information (step S209). The area class information is information in which the identification information of the single-class area is associated with the identification information of the class of the configuration points included in the single-class area.

In other words, when the area to be processed is a single-class area, all of the constituent points contained in the area to be processed are of the same class, so there is no need to store class information for each constituent point. By omitting the class information for each constituent point and instead storing only the identification information for the class of the constituent points contained in the area to be processed, the amount of data is reduced.

If the area to be processed is a multiple class area (step S208-No), the conversion unit 356 converts the class information of the configuration points included in the multiple class area into coded class information (step S210). The coded class information is information obtained by converting the class information of the configuration points included in the multiple class area so that the identification information of the area's most frequent class is omitted. The area's most frequent class is the class that corresponds to the largest number of configuration points among the classes of configuration points included in the group to be processed and included in the multiple class area to be processed.

Figure 10 is a schematic diagram for explaining the encoding class information. The conversion unit 356 extracts class information of the constituent points included in the group to be processed and included in the area to be processed from the class information of the constituent points included in the point cloud information. At this time, the conversion unit 356 extracts the class information CL so that the order of the class information rearranged in step S204 is maintained.

The conversion unit 356 identifies the area most frequent class and the non-area most frequent class based on the extracted class information CL. The non-area most frequent class is a class that is different from the area most frequent class among the classes of the constituent points that are included in the group to be processed and are included in the multi-class area to be processed. In the example shown in FIG. 10, the building class is identified as the area most frequent class, and the ground class and noise class are identified as the non-area most frequent classes.

The conversion unit 356 counts the types of non-region most frequent classes and encodes the class information of the non-region most frequent classes with a number of bits according to the counting result. In the example shown in FIG. 10, the class information CL includes three types of classes: ground class, building class, and noise class. Of the three classes, the building class is identified as the region most frequent class. Therefore, the non-region most frequent classes include two types of classes: ground class and noise class. The class information of the non-region most frequent class is encoded with 1 bit so that these classes can be identified. In the example shown in FIG. 10, the class information of the ground class and the noise class are encoded with bit strings 0 and 1, respectively.

The conversion unit 356 also calculates the maximum number of consecutive classes of the same type for each type of class. In the example shown in FIG. 10, the maximum number of consecutive classes of the ground, building, and noise are 3, 4, and 2, respectively. The conversion unit 356 determines the number of bits to be used when encoding the number of consecutive classes of the same type based on the calculated maximum number. At this time, the conversion unit 356 determines the number of bits for the area-most frequent class so that an integer greater than or equal to 0 and less than or equal to the maximum number can be identified. Furthermore, the conversion unit 356 determines the number of bits for the non-area-most frequent class so that an integer greater than or equal to 1 and less than or equal to the maximum number can be identified. In the example shown in FIG. 10, the building class, which is the area-most frequent class, is encoded with 3 bits so that an integer greater than or equal to 0 and less than or equal to 4 (i.e., 0, 1, 2, 3, and 4) can be identified. The ground class, which is the non-area-most frequent class, is encoded with 2 bits so that an integer from 1 to 3 (i.e., 1, 2, or 3) can be identified. The noise class, which is the non-region most frequent class, is coded with one bit so that integers greater than or equal to 1 and less than or equal to 2 (i.e., 1 or 2) can be identified. The conversion unit 356 codes the consecutive numbers according to the determined number of bits.

The conversion unit 356 generates coded class information by encoding class information of the configuration points included in the multiple class area so that the consecutive number of the area most frequent class, the identification information of the non-area most frequent class, and the consecutive number of the non-area most frequent class are repeated in this order. In the example shown in FIG. 10, for ease of viewing, a hyphen (-) is shown between the identification information of the non-area most frequent class and the consecutive number in the converted data, and a slash (/) is shown between the consecutive number of the area most frequent class and the identification information of the non-area most frequent class, and between the consecutive number of the non-area most frequent class and the consecutive number of the area most frequent class. As shown in FIG. 10, when different types of non-area most frequent classes are consecutive, a bit string of 000 (underlined) is inserted between them in the converted data, indicating that the consecutive number of the area most frequent class is 0 (i.e., there is no area most frequent class between the non-area most frequent classes).

The encoded class information generated in this manner includes identification information for the non-region most frequent class, but does not include identification information for the region most frequent class. In other words, the conversion unit 356 converts the class information for the constituent points included in the multi-class region so that identification information for the region most frequent class is omitted.

Returning to FIG. 7, after step S209 or S210, the conversion unit 356 determines whether all areas have been selected as areas to be processed (step S211).

If all areas have not been selected as areas to be processed (step S211-No), the conversion process returns to step S207, and the conversion unit 356 selects one of the areas that has not yet been selected as areas to be processed as the area to be processed.

If all areas have been selected as areas to be processed (step S211-Yes), the setting unit 355 determines whether all groups have been selected as groups to be processed (step S212).

If all groups have not been selected as the group to be processed (step S212-No), the conversion process returns to step S205, and the setting unit 355 selects one of the groups that has not yet been selected as the group to be processed as the group to be processed.

If all groups have been selected as groups to be processed (step S212-Yes), the conversion unit 356 generates conversion data (step S213). The conversion unit 356 concatenates the area class information or coding class information for the multiple areas set for each group in a predetermined order. For example, the predetermined order is an order based on the positional relationship of the multiple areas. The conversion unit 356 generates conversion data by concatenating the area class information or coding class information concatenated for each group based on the order of the groups. This ends the conversion process.

Figure 11 is a flow diagram showing an example of the flow of the area setting process. The area setting process is executed by the editing device 3 in step S206 of the conversion process.

First, the setting unit 355 determines whether or not the conditions for calculating a separation surface that separates a plurality of regions are satisfied (step S301). The conditions for calculating the separation surface are based on the number of types of classes of the constituent points included in the group to be processed. For example, the condition for calculating the separation surface is that the number of types of classes of the constituent points included in the group to be processed is two or more.

If the conditions for calculating the separation surface are not met (step S301-No), the area setting process ends.

If the conditions for calculating the separation surface are met (step S301-Yes), the setting unit 355 normalizes the coordinate information of the constituent points included in the group to be processed (step S302).

FIG. 12 is a schematic diagram for explaining normalization of coordinate information. The setting unit 355 performs linear coordinate conversion on the X and Y coordinates so that the X and Y coordinates (X, Y) of the four vertices of bucket B corresponding to the group to be processed, which were set in step S201 of the conversion process, become (-1,-1), (-1,1), (1,-1), and (1,1). The setting unit 355 also calculates the average value Zav, minimum value Zmin, and maximum value Zmax of the Z coordinates of the constituent points included in the group to be processed. If the absolute value of the difference between the average value Zav and the minimum value Zmin is greater than the absolute value of the difference between the average value Zav and the maximum value Zmax, the setting unit 355 performs linear coordinate conversion on the Z coordinate so that the average value Zav becomes 0 and the minimum value Zmin becomes -1. If the absolute value of the difference between the average value Zav and the maximum value Zmax is greater than the absolute value of the difference between the average value Zav and the minimum value Zmin, the setting unit 355 performs linear coordinate transformation of the Z coordinate so that the average value Zav becomes 0 and the maximum value Zmax becomes 1.

In this way, the coordinate information of the constituent points is normalized so that the absolute values of the X, Y, and Z coordinates of the constituent points included in the group to be processed are 1 or less. Note that instead of the average value Zav of the Z coordinates of the constituent points, the median value of the Z coordinates of the constituent points may be used.

Returning to FIG. 11, next, the setting unit 355 identifies the most frequent class corresponding to the largest number of constituent points among the classes of constituent points included in the group to be processed, and calculates an approximation plane of the constituent points of the most frequent class (step S303). For example, the setting unit 355 applies singular value decomposition to a matrix whose elements are coordinate information of the constituent points of the most frequent class, and calculates the singular values and singular vectors of the matrix. The setting unit 355 calculates, as the approximation plane of the constituent points of the most frequent class, the one that passes through the center of gravity of the constituent points of the most frequent class among planes whose normal is the three-dimensional singular vector corresponding to the smallest singular value. The setting unit 355 may calculate the approximation plane using other methods such as principal component analysis.

Next, the setting unit 355 calculates a lower separation surface and an upper separation surface as separation surfaces (step S304). The lower separation surface is a separation surface below the approximation plane, and the upper separation surface is a separation surface above the approximation plane. Here, the lower side is the negative direction of the Z axis, and the upper side is the positive direction of the Z axis.

Figure 13 is a schematic diagram for explaining the calculation of the lower separation surface and the upper separation surface. In Figure 13, the configuration points shown with white circles are configuration points of the building class, which is the most frequent class. The configuration points shown with black circles are configuration points of the ground surface class, which is a non-most frequent class different from the most frequent class. The configuration points shown with triangles are configuration points of the noise class, which is a non-most frequent class.

The setting unit 355 identifies the configuration point that is the smallest distance in the Z direction from the approximation plane AP among the configuration points of the non-most frequent class below the approximation plane AP. The setting unit 355 calculates the plane that passes through the identified configuration point among the planes parallel to the approximation plane as the lower separation plane. In the example shown in FIG. 13, the configuration point PL of the ground surface class that is the smallest distance D in the Z direction from the approximation plane AP is identified, and the lower separation plane LP that passes through the configuration point PL is calculated.

The approximation plane is expressed by the following formula:
The setting unit 355 calculates Zp, which is expressed by the following formula, for each component point of the non-mode class below the approximation plane AP.
Here, Xp and Yp are the X and Y coordinates of the configuration point. The setting unit 355 can specify the configuration point with the maximum Zp among the configuration points of the non-mode class below the approximate plane AP as the configuration point PL with the minimum distance D in the Z direction from the approximate plane AP.

If there are no constituent points of the non-most frequent class below the approximation plane AP, the setting unit 355 does not calculate the lower separation plane.

The setting unit 355 also identifies, among the non-modal class constituent points above the approximate plane AP, the constituent point that has the smallest distance in the Z direction from the approximate plane AP. The setting unit 355 calculates, as the upper separation plane, a plane that passes through the identified constituent point among planes parallel to the approximate plane. In the example shown in FIG. 13, a noise class constituent point PU that has the smallest distance in the Z direction from the approximate plane AP is identified, and an upper separation plane UP that passes through the constituent point PU is calculated.

The setting unit 355 calculates Zp for each component point of the non-most frequent class above the approximation plane AP using the above formula. The setting unit 355 can identify the component point with the smallest Zp among the non-most frequent classes above the approximation plane AP as the component point PU with the smallest distance D in the Z direction from the approximation plane AP.

If there are no constituent points of the non-most frequent class above the approximation plane AP, the setting unit 355 does not calculate the upper separation plane.

By calculating the lower separation surface and the upper separation surface in this manner, the area between the lower separation surface and the upper separation surface becomes a single class area. The setting unit 355 stores the parameters (a, b, c, and d) indicating the calculated lower separation surface and upper separation surface in the memory unit 31.

Returning to FIG. 11, next, the setting unit 355 determines whether or not the conditions for calculating the additional separation surface are satisfied (step S305). The additional separation surface is a separation surface that is different from the lower separation surface and the upper separation surface. The conditions for calculating the additional separation surface are based on the number of types of classes of the constituent points included in the region. For example, the condition for calculating the additional separation surface is that the number of types of classes of the constituent points included in any region is three or more.

If the conditions for calculating the additional separation surface are met (step S305-Yes), the setting unit 355 calculates the additional separation surface for the area where the conditions are met (step S306).

Figure 14 is a schematic diagram for explaining the calculation of an additional separation surface. As in Figure 13, in Figure 14, the configuration points illustrated with white circles are configuration points of the building class, which is the most frequent class. The configuration points illustrated with black circles are configuration points of the ground surface class, which is a non-most frequent class different from the most frequent class. The configuration points illustrated with triangles are configuration points of the noise class, which is a non-most frequent class.

In the example shown in FIG. 14, the area below the lower separation plane LP contains constituent points of three classes: building class, ground class, and noise class. Therefore, an additional separation plane is calculated for the area below the lower separation plane LP. On the other hand, the area above the upper separation plane UP contains only constituent points of the noise class. Therefore, an additional separation plane is not calculated for the area above the upper separation plane UP.

The setting unit 355 identifies the first most frequent class corresponding to the largest number of points and the second most frequent class corresponding to the second largest number of points among the classes of points included in the area where the condition is satisfied. In the example shown in FIG. 14, the ground class is identified as the first most frequent class, and the building class is identified as the second most frequent class.

The setting unit 355 identifies a constituent point that is included in the region where the condition is satisfied and is a non-most frequent class different from the first most frequent class and the second most frequent class, and that has the smallest distance in the Z direction from the lower separation plane. The setting unit 355 calculates a plane that passes through the identified constituent point, among planes parallel to the lower separation plane, as an additional separation plane. In the example shown in FIG. 13, a constituent point PB of the noise class that has the smallest distance D in the Z direction from the lower separation plane LP is identified, and an additional separation plane BP that passes through the constituent point PB is calculated.

When an additional separation surface is calculated for an area above the upper separation surface, the setting unit 355 identifies a configuration point of the non-modal class that has the smallest distance from the upper separation surface along the Z direction, and calculates a plane that passes through the identified configuration point among planes parallel to the approximation plane as the additional separation surface. Similarly, when an additional separation surface is calculated for an area below or above the additional separation surface, the setting unit 355 similarly calculates a plane that passes through a configuration point that has the smallest distance from the additional separation surface along the Z direction among planes parallel to the approximation plane as the further additional separation surface. The setting unit 355 stores parameters indicating the calculated additional separation surface in the storage unit 31. Since the additional separation surface is parallel to the lower separation surface, the parameters indicating the additional separation surface do not need to include a parameter indicating its inclination.

Returning to FIG. 11, after step S306, the region setting process returns to step S305, where the setting unit 355 determines whether the conditions for calculating the additional separation surface are met.

If the conditions for calculating the additional separation surface are not met in step S305 (step S305-No), the area setting process ends.

Figure 15 shows an example of the data structure of the conversion data D generated by the conversion process. The conversion data D correlates group identification information, the number of regions, the parameters of the separation plane, and information on the class of the constituent points included in each region.

In the example shown in FIG. 15, two separation surfaces are calculated for group G1, so three regions are set. In addition, the separation surface, which is a plane, is specified by four parameters, so the four parameters are stored as the parameters of the separation surface. Information regarding the class of the constituent points contained in each region differs depending on whether the region is a single-class region or a multiple-class region. If the region is a single-class region, the region class information generated in step S209 of the conversion process is stored.

Figure 16 is a diagram showing an example of the data structure of information E about the classes of constituent points included in a multi-class area. Information E about the classes of constituent points included in a multi-class area correlates the number of types of classes of constituent points included in a multi-class area, class information of the constituent points, the number of bits for encoding the consecutive number, identification information of the non-area most frequent class, and encoding class information.

The class information of the constituent points included in the multiple class area is ordered in descending order of the number of corresponding constituent points. The number of bits for encoding consecutive numbers is ordered in the same way as the class information of the constituent points. In the example shown in FIG. 16, it is shown that, of the constituent points included in the multiple class area, the building class has the most constituent points and the noise class has the least constituent points. It is also shown that the number of bits for encoding consecutive numbers of the building class, ground class, and noise class are 3, 2, and 1, respectively.

The identification information of the non-region most frequent classes is ordered in the same order as the class information of the constituent points. In the example shown in FIG. 16, the identification information of the non-region most frequent classes is ordered in descending order of the number of corresponding constituent points for classes other than the building class, which is the region most frequent class. In other words, it is shown that the identification information of the ground class is 0, and the identification information of the noise class is 1. The encoding class information was generated in step S210 of the conversion process.

Figure 17 is a flow diagram showing an example of the flow of the association process. The association process is executed by the management device 2 in step S107 of the class information setting process.

First, the association unit 233 classifies the constituent points included in the point cloud information into a plurality of groups according to a predetermined rule (step S401). The association unit 233 classifies the constituent points into a plurality of groups according to the same rule as the classification in step S201 of the conversion process.

Next, the association unit 233 rearranges the constituent points included in the point cloud information based on the coordinate information in accordance with a predetermined rule (step S402). The association unit 233 rearranges the constituent points in accordance with the same rule as the rearrangement in step S202 of the conversion process.

Next, the association unit 233 selects a group to be processed from among the multiple groups (step S403).

Next, the association unit 233 selects an area to be processed from among the multiple areas set for the group to be processed (step S404).

Next, the matching unit 233 extracts, from the rearranged constituent points, constituent points included in the area to be processed (step S405). The matching unit 233 acquires parameters of the separation plane calculated for the group to be processed from the conversion data. The matching unit 233 determines whether each of the constituent points included in the group to be processed is located above or below the separation plane based on the acquired parameters of the separation plane. The matching unit 233 determines whether each constituent point is included in the area to be processed based on the result of the determination.

Next, the association unit 233 associates class information with the extracted configuration points (step S406). The association unit 233 refers to the conversion data and determines whether the area to be processed is a single-class area or a multiple-class area.

If the area to be processed is a single-class area, the association unit 233 obtains area class information corresponding to the area to be processed from the conversion data. The association unit 233 associates the class identification information included in the area class information with each of the extracted constituent points.

When the area to be processed is a multi-class area, the association unit 233 acquires the coding parameters and coding class information corresponding to the area to be processed from the conversion data. The association unit 233 decodes the coding class information based on the coding parameters to generate class information, and associates it with each of the extracted configuration points.

Next, the association unit 233 determines whether all areas have been selected as areas to be processed (step S407).

If all areas have not been selected as areas to be processed (step S407-No), the matching process returns to step S404, and the matching unit 233 selects one of the areas that has not yet been selected as an area to be processed as the area to be processed.

If all areas have been selected as areas to be processed (step S407-Yes), the association unit 233 determines whether all groups have been selected as groups to be processed (step S408).

If all groups have not been selected as the group to be processed (step S408-No), the matching process returns to step S403, and the matching unit 233 selects one of the groups that has not yet been selected as the group to be processed as the group to be processed.

If all groups have been selected as groups to be processed (step S408-Yes), the matching process ends.

As described above, in the transmission/reception system 1, the generation unit 352 of the editing device 3 generates class information that does not include coordinate information of the multiple constituent points and indicates the class of the multiple constituent points. The association unit 233 of the management device 2 associates the class information with the multiple constituent points. This eliminates the need for the editing device 3 to transmit coordinate information to the management device 2, making it possible for the transmission/reception system 1 to shorten the time required to transmit and receive information indicating the class of a point cloud.

The administrator checks whether the class information associated with the configuration point by the management device 2 is appropriate. If the class information is inappropriate, the editing device 3 corrects the class information and transmits it again to the management device 2. In this way, since class information may be repeatedly transmitted from the editing device 3 to the management device 2, the time required to send and receive the class information accounts for a large proportion of the time required for the work of setting the class information. By reducing the time required to send and receive the class information, the work of setting the class information is greatly made more efficient.

The classification unit 354 of the editing device 3 rearranges the constituent points included in the point cloud information according to a predetermined rule, and rearranges the class information to correspond to the rearranged constituent points. The association unit 233 rearranges the constituent points included in the point cloud information according to a predetermined rule, and associates the rearranged class information with the rearranged constituent points. That is, the constituent points are rearranged according to the same rule by the classification unit 354 and the association unit 233.

When an operator operates the editing device 3 to manually set classes to the configuration points, the order of the configuration points may be changed by the operator. The same applies when the operator checks the classes automatically set by the editing device 3. In these cases, in order for the management device 2 to properly associate the configuration points with the class information, the editing device 3 needs to transmit information indicating the changed order of the configuration points to the management device 2. In the transmission/reception system 1, the classification unit 354 and the association unit 233 rearrange the configuration points according to the same pre-set rules, so that even if the order of the configuration points is changed in the editing device 3 or the management device 2, the class information can be associated with the configuration points without transmitting or receiving information indicating the changed order of the configuration points. In other words, even if the order of the configuration points is changed in the management device 2 or the editing device 3, there is no need to transmit or receive information indicating the order of the configuration points, so the transmission/reception system 1 makes it possible to shorten the time required to transmit and receive information indicating the class.

The classification unit 354 and the association unit 233 also classify the multiple configuration points into multiple groups according to a predetermined rule, and rearrange the configuration points for each of the multiple groups. By classifying the configuration points into multiple groups, the number of configuration points that are the subject of a single rearrangement process is reduced, making the rearrangement process more efficient. Therefore, the transmission/reception system 1 makes it possible to shorten the time required for processing associated with the transmission and reception of information indicating a class.

The classification unit 354 also classifies the constituent points classified into groups into a plurality of groups for each group, and rearranges the constituent points included in each of the plurality of small groups based on the coordinate information. By rearranging the constituent points for each small group based on the coordinate information, the transmission/reception system 1 reduces the calculation load compared to rearranging all the constituent points based on the coordinate information.

In addition, in the transmission/reception system 1, the setting unit 355 sets multiple regions in the coordinate space. The conversion unit 356 converts the class information of the constituent points included in the single class region into region class information. This makes it unnecessary to transmit class information for each of the constituent points included in the single class region, and therefore enables the transmission/reception system 1 to reduce the time required to transmit and receive information indicating the class of the point cloud.

The conversion unit 356 also identifies the region's most frequent class from among the classes of the constituent points included in the multiple class region, and converts the class information of the constituent points included in the multiple class region so that the class identification information of the constituent points of the region's most frequent class is omitted. This omits the class information indicating the region's most frequent class, which is the largest number of constituent points included in the multiple class region, making it possible for the transmission/reception system 1 to further reduce the time required to transmit and receive information indicating the class of the point cloud.

The setting unit 355 also sets the multiple regions so that the multiple regions include at least one single-class region. In a single-class region, class information for all constituent points contained therein is omitted, whereas in a multiple-class region, only class information for the constituent points of the most frequent class in the region is omitted. Therefore, a single-class region has a greater effect in reducing the amount of data in class information than a multiple-class region. By setting the multiple regions so that the multiple regions include at least one single-class region, the transmission/reception system 1 can further shorten the time required to transmit and receive information indicating the class of a point cloud.

In addition, in the transmission/reception system 1, the classification unit 354 classifies the constituent points included in the point cloud information into multiple groups based on the coordinate information of each constituent point. The setting unit 355 sets multiple areas for each group. In general, when the target geographical area is small, the topography can be considered to be uniform within that geographical area. In this case, the majority of the constituent points (approximately 80% to 90%) included in the point cloud data corresponding to that geographical area are associated with the same class. By classifying the constituent points into multiple groups based on their coordinate information, the majority of the constituent points included in the same group are in the same class, and many of the constituent points can be included in a single class area. This omits the class information of many of the constituent points, and enables the transmission/reception system 1 to further shorten the time required to transmit and receive information indicating the class of the point cloud.

In addition, in the transmission/reception system 1, the point cloud information is data that conforms to the LAS file format. This makes it possible for the transmission/reception system 1 to reduce the time required to transmit and receive information indicating the class of a point cloud, even in a general-purpose system that uses the LAS file format.

The following modified examples may be applied to the transmission/reception system 1.

In the above-described embodiment, in steps S303 and S304 of the region setting process, the setting unit 355 calculates a lower separation surface and an upper separation surface that are parallel to the approximation plane of the region's most frequent class. This example is not limited to this, and the setting unit 355 may calculate a lower separation surface and an upper separation surface that have mutually different inclinations.

Figure 18 is a schematic diagram for explaining the calculation of a lower separation surface and an upper separation surface that have mutually different inclinations. For each small bucket set in step S202 of the conversion process, the setting unit 355 identifies the uppermost and lowermost constituent points of the area most frequent class contained in the small bucket. In the example shown in Figure 18, the constituent point of the area most frequent class contained in each small bucket and located at the top is identified as Pmax, and the constituent point of the area most frequent class contained in each small bucket and located at the bottom is identified as Pmin.

Next, the setting unit 355 calculates an upper separation surface, which is an approximation plane of the constituent points of the most frequent area class located at the top in each bucket, and a lower separation surface, which is an approximation plane of the constituent points of the most frequent area class located at the bottom in each bucket. For example, the setting unit 355 calculates the upper separation surface and the lower separation surface using singular value decomposition, principal component analysis, etc. In the example shown in FIG. 18, the upper separation surface UP is calculated as the approximation plane of the constituent point Pmax of the most frequent area class located at the top in each small bucket. In addition, the lower separation surface LP is calculated as the approximation plane of the constituent point Pmin of the most frequent area class located at the bottom in each small bucket.

The setting unit 355 determines whether the planar position of the intersection line between the upper separation surface and the lower separation surface is inside the bucket corresponding to the group to be processed. In the example shown in FIG. 18, the planar position of the intersection line IL between the upper separation surface UP and the lower separation surface LP is not inside bucket B. In this case, the setting unit 355 uses the calculated upper separation surface UP and lower separation surface LP.

If the planar position of the intersection line is inside the bucket, the upper separation surface may be located below the lower separation surface, and the area will not be set appropriately. Therefore, in this case, the setting unit 355 recalculates the upper separation surface or the lower separation surface so that the planar position of the intersection line is no longer inside the bucket. For example, the setting unit 355 executes the processes of steps S303 and S304 to recalculate the upper separation surface and the lower separation surface, and adopts these upper separation surface and lower separation surface.

In this way, by calculating upper and lower separation planes with mutually different slopes, it becomes easier for constituent points of the same class to be concentrated in each region. As described above, in a single-class region, the class identification information of all constituent points is omitted. Also, in a multiple-class region, the class identification information of the constituent points of the region's most frequent class is omitted. Therefore, the more constituent points of the same class are concentrated in a specific region, the more data volume is reduced in the conversion process.

The setting unit 355 may translate the calculated upper separation surface and lower separation surface so that at least one single-class area is included in the multiple areas. For example, similar to step S304, the setting unit 355 translates the upper separation surface so as to pass through the lowest one of the non-modal class constituent points located above the predetermined position along the Z direction relative to the upper separation surface. The setting unit 355 also translates the lower separation surface so as to pass through the highest one of the non-modal class constituent points located below the predetermined position along the Z direction relative to the lower separation surface. For example, the predetermined position is a position corresponding to the median value of the Z coordinates of the constituent points included in the group to be processed. In the example shown in FIG. 18, the upper separation surface UP is translated to a plane UP' passing through the non-modal class constituent point PU located lowest along the Z direction relative to the upper separation surface UP. Also, the lower separation surface LP is translated to a plane LP' passing through the non-modal class constituent point PL located highest along the Z direction relative to the lower separation surface LP. This makes it easier for components of the same class to be distributed unevenly across each area.

In the above-described embodiment, the class information is converted into conversion data in the conversion process. This is not a limitation, and the conversion process may be omitted. In this case, the transmission unit 353 transmits class information that does not include coordinate information of the multiple constituent points included in the point cloud information to the management device 2. Even in this way, there is no need to transmit coordinate information, and therefore the time required to transmit and receive information indicating the class of the point cloud is reduced.

In addition, the steps from step S204 onwards in the conversion process may be omitted. That is, the transmission unit 353 may transmit class information rearranged according to a predetermined rule to the management device 2. In this way, the time required to transmit and receive information indicating the class of the point cloud is also reduced. Also, even if the order of the constituent points is changed in the management device 2 or the editing device 3, there is no need to transmit and receive information indicating the order of the constituent points.

In the above-described embodiment, in step S201 of the conversion process, the classification unit 354 classifies the constituent points included in the point cloud information into multiple groups. This is not a limitation, and the constituent points do not have to be classified into multiple groups. In other words, step S201 may be omitted. In cases such as when the distribution range of the constituent points is narrow, even if the constituent points are not classified into multiple groups, constituent points of the same class will be ubiquitous in the same area, and the amount of data in the conversion data will be reduced.

In the above-described embodiment, in steps S202 and S203 of the conversion process, the constituent points and their class information included in the point cloud information are rearranged according to a predetermined rule. This is not a limitation, and the classification unit 354 does not need to rearrange the constituent points and class information. That is, steps S202 and S203 may be omitted. If the order of the constituent points is not changed in the management device 2 or the editing device 3, or if the constituent points included in the point cloud information have been rearranged in advance in the management device 2 according to a predetermined rule, the class information can be appropriately associated with the constituent points without rearranging the constituent points and class information.

In the above-described embodiment, the conversion data generated by the conversion process does not include the constituent point ID. This is not limited to the above example, and the constituent point ID may be associated with the class information included in the conversion data. This makes it easy to associate the constituent points with the class information in the management device 2. Also, in this case, the class information of some of the constituent points is omitted, thereby reducing the time required to send and receive information indicating the class of the point cloud.

In the above-described embodiment, in step S209 of the conversion process, the conversion unit 356 converts the class information of the constituent points included in the multiple class area into encoded class information in which the identification information of the class of the constituent points of the most frequent class in the area is omitted. This example is not limited to this, and the identification information of the class of the constituent points does not have to be omitted. For example, the conversion unit 356 may convert the class information of the constituent points included in the multiple class area into a run-length code in which the identification information of the class of the constituent points and the consecutive number of the class are repeated. Furthermore, the conversion unit 356 may convert the class information of the constituent points included in the multiple class area into encoded class information and a run-length code, respectively, and adopt the encoded class information or the run-length code that has the smaller data volume.

In the above embodiment, in steps S303 and S304 of the region setting process, the setting unit 355 calculates an approximate plane for each group to be processed, and calculates a lower separation surface and an upper separation surface parallel to the approximate plane. This is not limited to the above example, and the setting unit 355 does not need to calculate an approximate plane. For example, the setting unit 355 may calculate a lower separation surface and an upper separation surface parallel to the lower separation surface and the upper separation surface of a group adjacent to the group to be processed for each group to be processed. Even in this way, the topography of adjacent groups is likely to be similar, so that the constituent points can be appropriately classified according to class. In addition, the calculation load for calculating the approximate plane is reduced. In addition, since the coordinate information is normalized in step S302, if the separation surfaces are parallel, the parameters indicating the inclination of each separation surface can be omitted, and the time required for transmitting and receiving information indicating the class of the point cloud is reduced.

Furthermore, when the inclination of the lower separation surface or the upper separation surface is small, the setting unit 355 may calculate a horizontal plane as the separation surface instead of the calculated lower separation surface or upper separation surface. For example, when the inclination of the lower separation surface or the upper separation surface is small, v shown in the following formula is equal to or larger than a threshold value.

This omits the parameters indicating the inclination of each separation plane, thereby reducing the time required to transmit and receive information indicating the class of the point cloud.

In the above-described embodiment, the condition for calculating the additional separation surface in step S305 is based on the number of types of classes of constituent points included in the region. This example is not limited to this, and the condition for calculating the additional separation surface may be based on the number of constituent points included in the region and the number of types of classes included in the region. For example, the condition for calculating the additional separation surface may be that at least one of the following is satisfied: (1) the number of constituent points included in the region is equal to or greater than a threshold value, and the number of types of classes of constituent points included in the region is two or more, and (2) the number of types of classes of constituent points included in the region is three or more.

In this case, if the condition (1) is satisfied, the setting unit 355, like step S301, identifies a region most frequent class and a non-most frequent class from among the classes of the constituent points included in the region, and calculates an additional separation surface. That is, the setting unit 355 calculates an additional separation surface so that a single-class region is separated by the additional separation surface. If the condition (2) is satisfied, the setting unit 355, like step S306, identifies a first region most frequent class, a second region most frequent class, and a non-most frequent class from among the classes of the constituent points included in the region, and calculates an additional separation surface. That is, the setting unit 355 calculates an additional separation surface so that a multi-class region containing only two types of constituent points is separated by the additional separation surface.

Even if a single-class region is separated by an additional separation surface, if the number of constituent points included in the single-class region is small, the degree of reduction in data volume by converting class information into region class information is small. On the other hand, calculating the additional separation surface increases the calculation load on the editing device 3 and management device 2. The setting unit 355 reduces the calculation load by not calculating the additional separation surface when the number of constituent points included in the region separated by the additional separation surface is small and the degree of reduction in data volume is small.

After calculating the additional separation surface, the setting unit 355 may determine whether or not a condition for adopting the calculated additional separation surface is satisfied. For example, the condition for adopting the additional separation surface is a condition based on the number of constituent points included in each of the two regions separated by the additional separation surface.

In this case, in step S305, the setting unit 355 determines whether the number of types of classes of the constituent points included in the region is two or more. If the number of types of classes of the constituent points included in the region is two or more, in step S306, the setting unit 355 calculates an additional separation surface so that the single-class region is separated. After step S306, the setting unit 355 determines whether the condition for adopting the additional separation surface is satisfied. For example, the setting unit 355 determines whether the number of constituent points included in each of the two regions separated by the additional separation surface is equal to or greater than a threshold. If the number of constituent points included in each of the two regions is equal to or greater than a threshold, the setting unit 355 adopts the additional separation surface, and the region setting process returns to step S305. If the number of constituent points included in at least one of the two regions is less than the threshold, the setting unit 355 does not adopt the additional separation surface. In this case, the setting unit 355 recalculates the additional separation surface so that the multi-class region including only the constituent points of two classes is separated, and adopts the recalculated additional separation surface. Thereafter, the region setting process returns to step S304.

When the number of constituent points included in the area separated by the additional separation surface is small, the degree of reduction in the amount of data by converting the class information into area class information or encoded class information is small. The setting unit 355 adopts an additional separation surface that includes a large number of constituent points in the area separated by the additional separation surface and that reduces the amount of data to a large extent, thereby efficiently shortening the time required to transmit and receive information indicating the class of the point cloud.

In the above-described embodiment, the lower separation surface, the upper separation surface, and the additional separation surface are flat surfaces. However, the present invention is not limited to such an example, and the lower separation surface, the upper separation surface, or the additional separation surface may be a curved surface such as a quadratic surface. This allows the constituent points to be separated more appropriately.

It should be understood by those skilled in the art that various changes, substitutions, and modifications can be made to the present invention without departing from the scope of the present invention. For example, the above-described embodiments and modifications may be implemented in appropriate combinations within the scope of the present invention.

2 Management device 231 Transmission unit 232 Reception unit 233 Correlation unit 3 Editing device 351 Reception unit 352 Generation unit 353 Transmission unit 354 Classification unit 355 Setting unit 356 Conversion unit

Claims (8)

an acquisition unit that acquires point cloud information including coordinate information indicating the positions of a plurality of constituent points that indicate the shape of the earth's surface and features in a predetermined coordinate space, and class information that indicates a classification result of each constituent point based on the type of thing that each constituent point indicates;
A setting unit that sets a plurality of regions in the coordinate space;
a conversion unit that selects a single-class area in which all of the constituent points included therein have the same class from among the plurality of areas, and converts the class information into area class information that does not include identification information of the class of each of the constituent points included in the single-class area, and in which the identification information of the single-class area and the identification information of the class of the constituent points included in the single-class area are associated with each other;
having
The conversion unit is
A multi-class region including points constituting two or more classes is selected from the plurality of regions;
Identifying a region most frequent class having the largest number of corresponding configuration points from among the classes of configuration points included in the multiple class region;
converting the class information such that, among the configuration points included in the multiple class region, class identification information of the configuration point of the most frequent class in the region is omitted;
A point cloud information conversion device comprising : an acquisition unit that acquires point cloud information including coordinate information indicating the positions of a plurality of constituent points that indicate the shape of the earth's surface and features in a predetermined coordinate space, and class information that indicates a classification result of each constituent point based on the type of thing that each constituent point indicates;
A setting unit that sets a plurality of regions in the coordinate space;
a conversion unit that selects a single-class area in which all of the constituent points included therein have the same class from among the plurality of areas, and converts the class information into area class information that does not include identification information of the class of each of the constituent points included in the single-class area, and in which the identification information of the single-class area and the identification information of the class of the constituent points included in the single-class area are associated with each other;
having
The setting unit is
Identifying a most frequent class corresponding to the largest number of component points from among the classes of the plurality of component points;
Calculating an approximate plane of the constituent points of the most frequent class;
a plane parallel to the approximation plane is calculated as a separation plane that separates the plurality of regions, thereby setting the plurality of regions;
A point cloud information conversion device comprising : the setting unit calculates, as the separation plane, a plane that passes through a component point of a class different from the most frequent class that is closest to the approximation plane in a predetermined coordinate axis direction of the coordinate space and is parallel to the approximation plane.
The point cloud information conversion device according to claim 2 . the setting unit calculates a horizontal plane as the separation plane instead of the calculated separation plane when the calculated inclination of the separation plane satisfies a predetermined condition.
The point cloud information conversion device according to claim 2 . an acquisition unit that acquires point cloud information including coordinate information indicating the positions of a plurality of constituent points that indicate the shape of the earth's surface and features in a predetermined coordinate space, and class information that indicates a classification result of each constituent point based on the type of thing that each constituent point indicates;
A setting unit that sets a plurality of regions in the coordinate space;
a conversion unit that selects a single-class area in which all of the constituent points included therein have the same class from among the plurality of areas, and converts the class information into area class information that does not include identification information of the class of each of the constituent points included in the single-class area, and in which the identification information of the single-class area and the identification information of the class of the constituent points included in the single-class area are associated with each other;
a classification unit that classifies the constituent points included in the point cloud information into a plurality of groups based on coordinate information of each constituent point ,
The setting unit sets the plurality of regions for each group,
the conversion unit identifies the single class area for each group, and converts class information of the configuration points included in the single class area into the area class information.
A point cloud information conversion device comprising : the setting unit sets the plurality of regions such that the plurality of regions includes at least one of the single-class regions;
The point cloud information conversion device according to claim 1 . acquiring point cloud information in which coordinate information indicating the positions of a plurality of constituent points indicating the shape of the earth's surface and features is associated with class information indicating a classification result of each constituent point based on the type of thing indicated by each constituent point in a predetermined coordinate space;
A plurality of regions are set in the coordinate space;
a single-class area in which all of the constituent points included therein have the same class is selected from among the plurality of areas, and the class information is converted into area class information that does not include identification information of the class of each of the constituent points included in the single-class area and in which the identification information of the single-class area and the identification information of the class of the constituent points included in the single-class area are associated with each other;
Including,
A multi-class region including points constituting two or more classes is selected from the plurality of regions;
Identifying a region most frequent class having the largest number of corresponding configuration points from among the classes of configuration points included in the multiple class region;
converting the class information such that, among the configuration points included in the multiple class region, class identification information of the configuration point of the most frequent class in the region is omitted;
A point cloud information conversion method comprising : acquiring point cloud information in which coordinate information indicating the positions of a plurality of constituent points indicating the shape of the earth's surface and features is associated with class information indicating a classification result of each constituent point based on the type of thing indicated by each constituent point in a predetermined coordinate space;
A plurality of regions are set in the coordinate space;
a single-class area in which all of the constituent points included therein have the same class is selected from among the plurality of areas, and the class information is converted into area class information that does not include identification information of the class of each of the constituent points included in the single-class area and in which the identification information of the single-class area and the identification information of the class of the constituent points included in the single-class area are associated with each other;
and have the computer execute the
A multi-class region including points constituting two or more classes is selected from the plurality of regions;
Identifying a region most frequent class having the largest number of corresponding configuration points from among the classes of configuration points included in the multiple class region;
converting the class information such that, among the configuration points included in the multiple class region, class identification information of the configuration point of the most frequent class in the region is omitted;
A program that causes a computer to execute the following :

Images