JP5140263B2 - Tree vertex recognition method, tree vertex recognition program, and tree vertex recognition apparatus - Google Patents

Description

  The present invention relates to a tree vertex recognition apparatus for recognizing the vertex of a single tree from a large number of tree data (DSM, DEM) obtained by laser measurement.

  The Kyoto Protocol entered into force in February 2005, and aims to reduce CO2 emissions by -3.9% in forest carbon absorption. In addition, forest parameters such as the number of trees, canopy area, and tree height are necessary to calculate the amount of CO2 absorbed by the forest.

  For example, when specifying the number of Karamatsu in a specific area in a mountainous area, the vertices of Karamatsu trees in the area such as a photographic image are visually identified, and the total number of these is obtained.

  In some cases, the number of trees is estimated using a laser range finder (hereinafter referred to as an aviation laser).

  Specifically, surface layer grid data (hereinafter referred to as DSM: Digital Surface Model) and ground grid data (hereinafter referred to as DEM: Digital Elevation Model) are created using data acquired by aviation laser measurement. Furthermore, the tree height grid data was calculated by taking the difference between DSM and DEM data, and single tree extraction was performed using the tree height or the maximum value of the tree height as a discrimination index.

  On the other hand, patent documents described below are disclosed for calculating carbon dioxide of trees.

  Japanese Patent Laying-Open No. 2005-52045 (Patent Document 1) evaluates the influence of vegetation based on the exchange process of radiation, energy, water and carbon dioxide in a soil-plant-atmosphere circulation system.

Japanese Patent Laying-Open No. 2004-33149 (Patent Document 2) automatically investigates the diameter, number, and arrangement of standing trees in a forest stand.
JP 2005-52045 A JP 2004-33149 A

  However, it takes a very long time to visually determine the number of trees from a photographic image. In particular, as the area becomes larger, it takes more time. In addition, experience is required to correctly identify each tree from a photographic image.

  In addition, the conventional tree extraction method using DEM and DSM has a complicated tree crown shape. Therefore, a plurality of local maximum values are likely to exist in one tree crown, and single tree extraction processing is performed using the conventional method. And there is a disadvantage that it is extracted excessively compared to the actual number.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to obtain a tree recognition apparatus that automatically recognizes the vertices of a single tree correctly.

The tree vertex recognition method of the present invention includes a DEM database storing DEM grid data obtained by emitting a laser from above to an area including a forest,
A DSM database storing grid data of the local DSM;
First storage means for storing tree height grid data;
Second storage means for storing averaged tree height grid data;
Third storage means for storing feature data of the crown three-dimensional shape;
Fourth storage means for storing feature data of the crown three-dimensional shape after filtering;
Fifth storage means for storing tree vertex candidates;
Providing a sixth storage means for the upper crown of the data is stored,

On the computer,
Reading the grid data of the DEM and the grid data of the DSM, setting the difference between the z values as a tree height, and storing the tree height grid data together with the x and y coordinate values of the grid data in the first storage means When,
Averaging the tree height grid data in a predetermined range centered on each data, and storing the averaged tree height grid data in the second storage means;
For each of the averaged tree height grid data, sequentially obtaining a ridge valley degree within a predetermined range, and storing it in the third storage means together with the x and y coordinate values as feature data of the canopy solid shape; ,
A filter that sequentially replaces each data with the ridge valley degree that is the maximum within a predetermined range is applied to the feature data of the crown three-dimensional shape, and the fourth storage means is used as the feature data of the crown three-dimensional shape after the filter. The step of storing in
Comparing the filtered crown crown shape feature data and the crown crown shape feature data, extracting the same as the tree vertex candidates and storing them in the fifth storage means;
Storing, in the sixth storage means, data on the crown crown shape characteristic data that is equal to or higher than a threshold value according to the state of the forest, as data on the crown top;
In the upper crown data, an adjacent cluster is set as one crown upper area, and the largest one of the tree vertex candidates existing in the upper crown area is determined as a tree vertex;
And summarized in that cause I rows and outputting the coordinates of the display or the tree vertexes on the display unit by the trees apex in a predetermined shape of the image.

  As described above, according to the present invention, the vertex of a single tree can be automatically recognized from the DEM and DSM obtained by laser measurement.

  For this reason, since it is not necessary to search a wide forest etc. for every tree manually, it is applicable to forest property management.

  In the present embodiment, the crown is made to stand out and reproduced, and the vertex of the single tree is recognized. In order to recognize the vertices of this single tree, the crown shape is discriminated by newly creating “ridge valley degree” data.

  The ridge valley degree is calculated as follows.

  In this embodiment, using the tree height grid data (0.5 m), which is the difference between DEM and DSM, the search is performed in 8 directions (= all directions) and the ground opening (φ1) and the underground opening ( Find φ2). The search range was 4m, 1m, 2m, 5m, and 10m, and was determined to be 2m that could properly represent the crown shape. Then, the bisector of the angle between the ground opening and the underground opening representing the rough topography in the search range 2 m is defined as the ridge valley degree (φ3).

  Embodiments of tree recognition will be described below with reference to the drawings.

  FIG. 1 is a schematic configuration diagram of a tree vertex recognition apparatus according to the present embodiment. As shown in FIG. 1, a database 10 storing grid data (DEM: Digital Elevation Model) of ground in a specific area and a database 11 storing grid data (DSM: Digital Surface Model) of a surface layer (forest) of the specific area. A database 12 storing a digital photograph of a specific area, a display unit 13, a tree height calculation unit 14, a file (memory) 15 in which tree height data is stored, a smoothing filter unit 16 (3 pixels × 3 Pixel), a ridge valley degree calculation unit 17, a file 18 in which the ridge valley degree is stored, a display processing unit 19, a local maximum filter unit 20, a tree vertex candidate extraction unit 21, and tree vertex candidate data are stored. File 22, tree crown upper extraction unit 23, file 24 in which tree crown upper data is stored, tree vertex determination unit 25, and tree vertex data are stored. File 26, output unit 30 and the like.

  The database 10 stores grid data (DEM: Digital Elevation Model: grid-like data with an interval of 0.5 m in the present embodiment) of the ground in a specific area. This grid data is stored in association with X, Y, and Z (ground height Za) of the ground. In other words, the laser data is read, and the ground is restored by generating a TIN for the contour map connecting the same elevation values.

  The database 11 stores grid data (DSM: Digital Surface Model: 0.5 m interval in the present embodiment) of the surface layer (forest) in a specific area. This grid data (in this embodiment, grid-like data with an interval of 0.5 m) stores X, Y, and Z (Zb: tree surface height) of the tree surface in association with each other. That is, the laser data is read and the surface layer is restored by generating a TIN for the contour map connecting the same elevation values.

  That is, in these grid data, an X value and a Y value corresponding to latitude and longitude and a Z value obtained by laser measurement are assigned to an XY plane rectangular coordinate system having a lattice interval of 0.5 m.

  The tree height calculation unit 14 obtains a difference between the DSM of the database 10 and the DEM of the database 11 and assigns the difference to the X and Y coordinates of the file 15 (memory).

  The smoothing filter unit 16 is also referred to as an average filter. The tree height grid data Di of the file 15 is XY coordinate plane 3 × 3 range centered on each data (1. 5m × 1.5m) and the average is stored in the file 26.

  When the tree height grid data Ddi (Dd1, Dd2,...) Is stored in the file 26, the ridge valley degree calculation unit 17 stores any averaged tree to which the tree height grid data Ddi is added. Search range of arbitrary radius of XY coordinate plane centering on each data from high grid data Ddi (hereinafter referred to as tree high grid data Ddi) (search range is 2m which can properly express the characteristics of the solid shape of tree crown) The degree of ridge valley is obtained at every 2 m range below and stored in the file 18.

  The local maximum value filter 20 sequentially applies a 3 × 3 range filter to the data Dei of the ridge valley degree of the file 18 and stores the result in the file 27.

  The tree vertex candidate extraction unit 21 compares the crown three-dimensional feature data Dfi (XY, degree of ridge valley) of the file 27 with the Dei (Xpi, Ypi, degree of ridge valley) of the file 18, and the same data is used as the vertex candidate Dgi. As a file 22.

  The crown top extraction unit 23 extracts the data Dei at the top of the canopy by providing a threshold corresponding to the state of the forest in the ridge valley degree data (crown solid shape feature data Dei) of the file 18, and extracts this data Data Dhi (Dh1, Dh2,...) Is stored in the file 24.

  The tree vertex determining unit 25 determines the data Dhi (Dh1, Dh2,...) And the vertex candidate Dgi (Dg1, Dg2,. This is saved in the file 28. When a plurality of temporary vertices exist at the top of the tree crown, the highest tree height at the top of the tree crown is determined as a single tree vertex and stored in the file 29 (Pi).

  The display processing unit 19 displays the determined vertex on the screen, for example, with a circle (identification such as a triangle, x, number, etc.). At this time, it may be displayed so as to overlap with the photographic image.

  The output unit 30 attaches a tree number to the coordinates (XY) of the vertex Pi of the file 29 and displays it on a printer (not shown) or a screen. At this time, the tree height may also be displayed.

(Description of operation)
The operation of the tree vertex specifying apparatus configured as described above will be described below. In the present embodiment, the display unit 13 displays the DSM of the database 10, the DEM of the database 11 and the digital photograph in an overlapping manner.

  FIG. 2 is a flowchart for explaining the schematic operation of the tree vertex recognition apparatus of the present embodiment.

  In the present embodiment, the tree height calculation unit 14 reads data (DSM: Digital Surface Model) acquired by aviation laser measurement (S1). Further, the grid data of the ground (hereinafter referred to as DEM: Digital Elevation Model) is read (S2).

  Then, the tree height calculation unit 14 takes the difference Zci (Zci, Zci...) Between the DSM and DEM data and stores it in the file 15 as tree height grid data Di (S3).

  A smoothing filter unit 16 (also referred to as “Average filter 16”) sequentially averages the tree height grid data Di in a 3 × 3 range and stores it in the file 26 (tree height grid data Ddi (Dd1, Dd2,...). )) (S4).

  Next, the ridge valley degree calculation part 17 calculates the ground opening degree and the underground opening degree for every 2 m range (Ddi) (S5). Next, the ridge valley degree calculation part 17 uses the underground opening degree and the ground opening degree to create a “ridge valley degree” data, thereby obtaining a ridge valley image that can distinguish the characteristics of the three-dimensional shape of the crown. 18 (S6).

  The ridge valley degree is calculated as follows. Using the grid data Ddi of the tree height, search is made in eight directions (= all directions) to obtain the ground opening (φ1) and the underground opening (φ2).

  The search range was 4m, 1m, 2m, 5m, and 10m, and was determined to be 2m that could properly represent the crown shape.

  Then, the bisector of the angle between the ground opening and the underground opening that represents the rough topography in the search range 2 m is defined as the ridge valley degree (φ3) (see FIG. 9). The data Dei of the ridge valley degree is displayed by the display processing unit 19 (ridge valley degree image).

  Further, the local maximum value filter 20 is sequentially applied to the data Dei of the ridge valley degree (S7), and is stored in the file 27.

  At this time, not only the size of the tree crown but also the minimum area size is fixed to a 3 × 3 range, and “tree candidate points” are extracted instead of tree vertices, and tree vertex candidates Dgh (dg1,. Save as (..) (S8). The 3 × 3 range is fixed because, for example, when it is 5 × 5, an experimental result has been obtained that the upper part of the crown where there are no tree vertex candidates is obtained.

  On the other hand, using the property that “ridge valley degree> 0” expresses ridge topography and “ridge valley degree <0” expresses valley topography, the tree vertex candidate extraction unit 21 converts the forest into the ridge valley degree data. The upper part of the tree crown is extracted by providing a threshold value corresponding to the state (S9), and the upper crown extraction part 24 is labeled (S10).

  Further, the tree vertex determining unit 25 obtains the highest tree height from the tree vertex candidate points for each of the tree top candidate points and the crown upper area, determines that point as the tree vertex (S11), and extracts this as the tree vertex. (S12). The determination of the tree vertex will be described later.

  FIG. 3 is a flowchart for explaining processing for obtaining the ridge valley degree of the tree vertex recognition apparatus of the present embodiment.

  In the present embodiment, an orthophoto image is displayed on the display unit 13 (S30).

  The tree height calculation unit 14 reads DSM (X, Y, Z) and DEM (X, Y, Z) (S31), obtains a difference Zci (S32), and obtains the tree height grid data Di (D1, D2). D2... Are stored in the file 15.

  As shown in FIG. 4, the tree height data Di is data in which coordinates (X, Y: plane rectangular coordinates) are associated with the tree height Zci.

  Next, the smoothing filter unit 16 sequentially averages (3 × 3 range) the grid data Di of a large number of tree heights in the file 15 with the filter 16 (S33), and the grid data Ddi of the tree height is stored in the file 26. save.

  Next, the ridge valley degree calculation part 17 discriminate | determines the characteristic of the solid shape of a tree by calculating | requiring a ridge valley degree sequentially in 2 m range in each Ddi (S34).

  The ridge valley degree is searched in 8 directions (may be performed at 1m, 5m, 10m depending on the type of tree) to obtain the ground opening and the underground opening, and the difference between the ground opening and the underground opening is 2 It is obtained by dividing by (S34). That is, the feature of the three-dimensional shape of the crown (leaves, branches) of the tree is required.

  Using the tree height grid data Ddi of the file 26, as shown in FIG. 5, search is made in eight directions (= all directions) to obtain the ground opening (φ1) and the underground opening (φ2). The search range is determined to be 2m, which can properly represent the three-dimensional shape of the tree crown.

  FIG. 6 is a plot of tree height grid data Ddi in an arbitrary search range on an axis composed of distance and height. Also, the black dots in FIG. 6 indicate grid data Ddi at intervals of 0.5 m (2 m range is not shown).

That is, FIG. 6 represents a rough crown shape when one direction is viewed within a range of an arbitrary radius from the sample point of interest. In FIG. 6, φ1 is the ground opening (ground angle), and φ2 is the underground opening (ground angle). 7 (a) and 7 (b), the angle between the bisector of the angle between the ground opening and the underground opening and the distance axis (horizontal axis) is defined as the ridge valley degree (φ3). ing. Further, the black dots in FIG. 7 indicate grid data Ddi at intervals of 0.5 m (2 m range is not shown).
In this embodiment, the cord range is set to 2 m, the ridge valley degree (φ3) is obtained, and stored in the file 18 as ridge valley degree data (hereinafter, crown three-dimensional shape feature data Dei (Dei: De1, De2...)) (S36). ).

FIG. 8 is an explanatory diagram representing the crown shape of the locus Ai with the tree height as the Y axis and the distance as the X axis. FIG. 9 is an explanatory diagram expressing the features of the three-dimensional shape of the crown of the locus Ai with the ridge valley degree as the Y axis and the distance as the X axis.

  That is, as shown in FIG. 9, when the features of the three-dimensional shape of the crown of the locus Ai are expressed by the ridge valley degree, the boundary between the crowns appears more conspicuously than in FIG. The numbers (1) to (6) in FIG. 8 and FIG. 9 indicate the crown area read from the photograph with the naked eye.

  In order to improve the ease of recognition as a single tree, the difference between the values of the tree vertices and the edge part is large, the values show the same tendency at the tree vertices, the values show the same tendency at the edges, etc. It is done. As shown in FIG. 9, it can be seen that the “ridge valley degree” data of the present method satisfies the above-mentioned conditions than the conventional “height” data. Thereby, it can contribute to the accuracy improvement at the time of creating a crown crown polygon automatically.

  In FIG. 9, it can be seen that the range from the point where the ridge valley degree has a negative value to the point where it suddenly increases in the positive direction and then suddenly becomes a negative value is the crown portion of the single tree.

  FIG. 11 is a diagram for explaining a method for selecting vertex candidates and tree vertices, and shows that a candidate having the highest tree height is selected at the top of each tree crown. The determination of the tree vertex will be described later in detail.

  Next, the ridge valley degree calculation part 17 preserve | saves the calculated | required ridge valley degree in the file 18 (XY is matched) as the characteristic data Dei of a crown solid shape (refer FIG. 10) (S36).

  The above-mentioned ground opening is searched for each element (Ddi) in eight directions from the element, and the maximum inclination angle is obtained within the consideration distance. The average of the maximum inclination angles in each of the 8 directions is expressed as the angle from the zenith. Note that the value of the ground opening depends on the range to be considered and must be set according to the scale of the object to be read.

  The underground opening is searched for each element in eight directions from the element, and the minimum inclination angle is obtained within the considered distance. The average of the minimum inclination angles in each of the 8 directions is expressed as the angle from the zenith. As with the ground opening, the value depends on the range to be considered and must be set according to the scale of the object to be interpreted.

  The ridge valley degree is calculated by subtracting the underground opening from the calculated ground opening and dividing by 2. Ridge terrain has a positive value and valley terrain has a negative value.

  And it progresses to the tree vertex extraction process of FIG. In step S7 of the tree vertex extraction process, the 3 × 3 local maximum value filter 20 (using a window of 3 × 3 range) is applied to the crown solid shape feature data Dei (De1, De2,...) Of the file 18. This is saved in the file 27 (Dfi).

  In step S8, the tree vertex candidate extraction unit 21 uses the crown solid feature data Dfi (see FIG. 11) of the peak point existing in the crown area ((1), (2),...) As a vertex candidate (hereinafter referred to as Dgi). Extracted and stored in the file 22 (see FIG. 12). Specifically, the value of the data in the file 18 and the data of the same coordinate value in the file 27 are defined as the same value in the file 22 as “1” and the others as “0” (see FIG. 12).

  Here, the description of the processes of the local maximum filter unit 20 and the tree vertex candidate extraction unit 21 will be supplemented.

  That is, the local maximum filter unit 20 (also referred to as Local max filter 20) applies a window of 3 × 3 range to the feature data Dei (De1, De2,...) Of the crown three-dimensional shape of the file 18, and becomes the maximum in this window. The process of replacing the value with the median value of the window is sequentially performed. These are stored in the file 27 as the canopy solid shape feature data Dfi.

  Then, the tree vertex candidate calculation unit 21 compares the crown three-dimensional shape feature data Dfi (XY, degree of ridge valley) with Dei (XY, degree of ridge valley) of the file 18, and the same data is stored in the file 22 as the vertex candidate Dgi. Saved.

  On the other hand, in step S9, the upper crown extraction unit 23 uses the property that “ridge valley degree> 0” represents ridge topography and “ridge valley degree <0” represents valley topography. The data Dei at the upper part of the canopy is extracted by setting a threshold value corresponding to the forest condition in the degree data (feature data Dei of the three-dimensional shape of the crown), and this is filed as data Dhi (Dh1, Dh2,...) 24 (see FIG. 13).

  The data Dhi in the upper part of the tree crown of FIG. 13A is obtained by extracting Dei exceeding the threshold. For example, the data Dei in the area of number 1 is (Xh5, Yh5) is “1”, (Xh6, Yh6) Is represented as “1” (Xh10, Yh10) as “1”.

  FIG. 13B specifically shows FIG. 13A, and the vertex exceeding the threshold is set to “1”. The range where “1” continues is the upper part of the tree crown.

  In step S10, Xhi and Yhi of the vertex candidate Dhi in the file 24 are converted into polygons (coordinates Xhi and Yhi and color information), and the display processing unit 19 labels the label ring (the collected area (upper)). (See FIG. 14).

    Next, in step S11, the tree vertex determining unit 25 superimposes and matches the data Dhi (Dh1, Dh2...) On the top of the tree of the file 24 and the vertex candidate Dgi (Dg1, Dg2. Is determined as a temporary vertex. When there are a plurality of temporary vertices in the upper area of the tree crown, the area with the highest tree height in the upper area of the tree crown (the areas of numbers 1, 2,... In FIGS. 8 and 9). It is determined as a vertex of a single tree, and this is stored in the file 29 (Pi). The display processing unit 19 displays the determined vertex on the screen (see FIG. 15). That is, for example, the vertex candidate Dgi is displayed as shown in FIG. 16, but the vertex-determined image is as shown in FIG.

  The explanation of the above-described vertex determination will be supplemented. For example, as shown in FIG. 18A, in the file 24, the first column is “00110011100”, the second column is “011111011110”, the third column is “001110001000”, and the fourth column is “00000100011”. When the upper apex area is obtained, as shown in FIG. 18 (b), a serial number is assigned to each group, with “1” adjacent groups (“0” separated) as one crown top. To go.

  For example, if there are a plurality of vertices in the crown having the same height, the data DQi of the file 28 is displayed once and designated by the operator, and the designated one is determined as the vertex Pi. Save to file 29.

  Next, the explanation about the ridge valley is supplemented. As the characteristics of the ridge valley degree, considering the topography of the sin curve as shown in FIG. 19, the ridge terrain is represented by “ridge valley degree> 0”, the valley terrain is represented by “ridge valley degree <0”, By the way, it is mentioned that “degree of ridge valley = 0”.

  Here, for the sake of easy understanding, if an ideal tree has a shape as shown in FIG. 19, “ridge valley degree> 0” at the top of the tree, and always “ridge valley degree = 0”. It reaches the edge of the canopy by passing through the point, where it is considered that the “ridge valley degree <0”. Therefore, it can be said that the upper part of the canopy can be defined as “degree of ridge valley> 0”.

  Utilizing this fact, in this work, tree vertices are narrowed down as shown in FIG. In other words, the top of the tree crown is extracted by using the value of the ridge valley degree, and the candidate point having the highest tree height (DSM-DEM) is selected as the tree vertex among the candidate tree top points existing in this extraction. It was.

  The explanation is supplemented about the threshold value of the ridge valley degree.

  When the threshold is set for the ridge valley degree and the upper part of the tree crown is extracted, it is understood that the tree 5 cannot be extracted in the case of the threshold α as shown in FIG. On the other hand, in the case of the threshold value β, the trees 5 and 6 are recognized as one. Thus, there is a problem that the tree vertex extracted by the threshold is affected.

Therefore, we decided to examine the optimum threshold for extracting the upper part of the tree crown. Specifically, the threshold value of the ridge valley degree is set to five types “10, 15, 20, 25, 30”, and the one having the best tree vertex extraction result is adopted. Table 1 shows the total results of the tree vertices after narrowing down.

  Looking at Table 1, it can be seen that the result with “degree of ridge valley> 20” was the best. Compared to this, the reason why the number of extracted trees with “ridge ridge degree> 10” is small is that the adjacent tree crown was recognized as one. Also, the reason why the number of extractions with “ridge ridge degree> 30” is small is that there was a crown that could not be extracted because the threshold was set high.

Therefore, it was found that “20” is the optimum threshold for the ridge valley degree for extracting the upper part of the tree crown. The extracted image of the upper part of the tree crown is shown in FIG. 14, and the extracted tree vertex extracted image is shown in FIG.

By calculating the apex of the canopy of a certain forest as described above, a single tree can be specified, so that it can be applied to management of forest assets. For example, when the output unit 30 outputs and prints the vertex data Pi (X, Y, Z) of the file 29, the number and height can be known, so that the value of the forest can be automatically grasped.

  In FIG. 22, the crown polygon is created after calculating the tree vertex. In creating the canopy polygon, a canopy polygon (see FIG. 23) is generated by the Wetershed algorithm using the tree vertices extracted in step S12 and the ridge valley degree of the file 18 (S13, S15).

  The threshold value determination method in step S13 is as follows.

  For example, an arbitrary line (trajectory Ai) is drawn in the orthophoto image, and the ridge valley degree calculation unit creates a cross section of the ridge valley degree and displays it (corresponding to FIG. 9).

  Next, an additional distance from the start of the crossing line (corresponding to the X axis in FIG. 9) is followed to search for the crown of the ortho image on the display portion and the boundary between the crowns.

  Then, by looking at the additional distance of the boundary position, the operator reads the ridge valley degree (calculated and displayed by the ridge valley degree calculation unit) of the additional distance from FIG.

  This process is repeated to read the degree of ridge valley where the upper crown is formed.

  Figure 9 (1) 0 and -50, (2) -50 and -20, (3) -20 and 0, (4) 0 and 0, (5) 0 and 10, (6) Then it became 10 and -10.

  From this result, the ridge valley degree 10 in which (1) to (6) are separated is set as a threshold value.

  If the threshold value is 0, (5) and (6) will be at the top of one tree crown, so 0 is not adopted.

  If the threshold is 30 (2), two upper crowns will be formed, and (5) will not be the upper crown. Accordingly, in FIG. 9, a ridge valley degree of about 10 is suitable as a threshold value. The threshold value thus determined is used.

  Although the above embodiment has been described using aerial photographs, gray scale may be assigned to the tree height grid data, DEM, or DSM.

  For example, the difference image between the ground opening and the underground opening is gray, the slope is put in the red channel, and by creating a pseudo color image, the ridge and the summit part are made whitish, and the valley and the depression are made darker, The steeper slope is expressed in red. With such a combination of expressions, even one image having a stereoscopic effect (also referred to as a stereoscopic red map) may be used.

  For example, when the value of the ground opening is within a range of about 40 degrees to 120 degrees, 50 degrees to 110 degrees is assigned to 255 gradations corresponding to the gray scale. That is, the ridge portion (convex portion) has a larger value of the ground opening, so the color becomes white.

  Further, this embodiment is preferably realized by a computer system VPS as shown in FIG. The VPS 1 includes a central information processing device (CPU) composed of an appropriate combination of workstation, processor, microcomputer, logic, registers, etc., and a keyboard (KB) for inputting control / operation information necessary for the central information processing device, An information input unit including a mouse, an interactive soft switch, an external communication channel, etc., and an information output unit including a display, a printer, an external communication channel, etc. that display and transmit information output from the central information processing device in a broad sense A ROM that stores information such as an operating system and application programs read by the central information processing apparatus, and a ram that stores information to be processed by the central information processing apparatus and information to be written from the central information processing apparatus at any time. (RAM) and the like. It is possible to integrate and subdivide ROM and RAM as appropriate.

  The application program described above includes a tree height calculation unit 14, a ridge valley calculation unit 17, a display processing unit 19, a local maximum filter 20 (also referred to as a Local max filter), and tree vertex candidate extraction according to the present embodiment. A unit 21, an upper crown extraction unit 23, a tree vertex determination unit 25, and an output unit 30.

  In this embodiment, the concept of opening is used. The degree of opening is a quantification of the extent that the point is protruding above the ground and the depth of penetration into the basement. In other words, the ground opening represents the size of the sky that can be seen within the distance L from the sample point of interest, and the underground opening is the basement in the range of distance L when looking up underground with a handstand. It represents the size.

  The opening degree depends on the distance L and the surrounding terrain. In general, the ground opening increases as the point protrudes higher from the surroundings, and takes a larger value at the summit and ridge and smaller at the depression and valley bottom. On the contrary, the underground opening becomes larger as the depth of penetration into the basement is greater, with a large value at the depression and valley bottom, and a small value at the summit and ridge.

  In addition, the opening degree map can extract information suitable for the size of the terrain by specifying the calculation distance, and can display without depending on directionality and local noise.

  In other words, it excels in the extraction of ridge lines and valley lines, and abundant topographical and geological information can be read, and one of the eight directions from the set point A was seen on a certain range of DEM data. An angle vector θi formed by the straight line L1 connecting the point B, which is sometimes the maximum vertex, and the horizontal line is obtained. This angle vector is calculated in eight directions, and the average of these is called the ground opening, and the solid with the air layer pressed against a certain range of DEM data (ground surface: solid) is turned over. Further, the angle formed by the straight line L2 connecting the point C (corresponding to the deepest point) that becomes the maximum vertex when any one of the eight directions from the point A of the inverted DEM data is viewed is obtained. Obtaining and averaging this angle over eight directions is referred to as underground opening ψi.

  That is, the ground opening is a maximum value (vertical) of the slope of the line L1 that generates a topographic cross section for each of the eight directions on the DEM data included in the range from the point of interest to a certain distance and connects each point to the point of interest. (When viewed from the direction). Such processing is performed for eight directions. The angle of inclination is the angle from the zenith (90 degrees for flats, 90 degrees or more for ridges and peaks, and 90 degrees or less for valleys and depressions).

  In addition, the underground opening is generated by generating a topographical section for each of the eight directions in the range from the point of interest of the inverted DEM data to a certain distance, and the maximum value of the slope of the line connecting each point and the point of interest (solid surface solids). In the figure, the minimum value) is obtained when L2 is viewed from the vertical direction. Such processing is performed for eight directions.

  The angle ψi when viewing L2 from the vertical direction in the three-dimensional map of the ground surface is 90 degrees if flat, 90 degrees or less at the ridge or peak, and 90 degrees or more at the valley or depression.

  That is, the ground opening and the underground opening consider two basic points A (iA, jA, HA) and B (iB, jB, HB).

The distance between A and B is
P = {(i _A − i _B ) ² + (j _A − j _B ) ² } ^1/2 … (1)
It becomes.

The relationship between A and B at the sample point is shown with an altitude of 0 m as a reference. The elevation angle θ when viewing the sample point B from the sample point A is θ = tan ⁻¹ {(HB−HA) / P}.
Given in. The sign of θ is positive when (1) HA <HB, and negative when (2) HA> HB.

A set of sample points within the range of the azimuth D distance L from the sample point of interest is described as DSL, and this is referred to as a “DL set of sample points of interest”. here,
DβL: Maximum value of the elevation angle for each element of DSL at the sample point of interest
DδL: The following definition is made as the minimum value of the elevation angle for each element of DSL at the sample point of interest.

Definition 1: The ground angle and underground angle of the D-L set of sample points of interest are each
DφL = 90− DβL
as well as
DψL = 90 + DδL
Means.

  DφL means the maximum value of the zenith angle at which the sky in the direction D can be seen within a distance L from the sample point of interest. Generally speaking, the horizon angle corresponds to the ground angle when L is infinite. Further, DψL means the maximum value of the nadir angle at which the ground in the direction D can be seen within a distance L from the sample point of interest. When L is increased, the number of sample points belonging to DSL increases, so that DβL has a non-decreasing characteristic, while DδL has a non-increasing characteristic. Therefore, both DφL and DψL have non-increasing characteristics with respect to L.

  High angle in surveying is a concept defined with reference to a horizontal plane passing through the sample point of interest, and does not exactly match θ. In addition, if the ground angle and the underground angle are to be strictly discussed, the curvature of the earth must be taken into consideration, and definition 1 is not necessarily an accurate description. Definition 1 is a concept defined on the assumption that terrain analysis is performed using DEM.

  The ground angle and the underground angle are the concepts of the specified direction D, but the following definition is introduced as an extension of this.

Definition II: The ground opening and the underground opening at the distance L of the sample point of interest are ΦL = (0φL + 45φL + 90φL + 135φL + 180φL + 225φL + 270φL + 315φL) / 8
And ψL = (0ψL + 45ψL + 90ψL + 135ψL + 180ψL + 225ψL + 270ψL + 315ψL) / 8
Means.

  The ground opening represents the extent of the sky that can be seen within the distance L from the sample point of interest, and the underground opening represents the extent of the underground in the range of distance L when looking up underground with a handstand. Represents.

It is a schematic block diagram of the tree vertex recognition apparatus of this Embodiment. It is a flowchart explaining schematic operation | movement of the tree vertex recognition apparatus of this Embodiment. It is a flowchart explaining the process which calculates | requires the ridge valley degree of the tree vertex recognition apparatus of this Embodiment. It is explanatory drawing of the tree height data Di. It is explanatory drawing explaining calculation of a ridge valley degree. It is explanatory drawing explaining calculation of a ridge valley degree. It is explanatory drawing explaining calculation of a ridge valley degree. It is explanatory drawing expressing the crown shape of locus | trajectory Ai by using tree height as a Y-axis and distance as an X-axis. It is explanatory drawing expressing the feature of the three-dimensional shape of the crown of locus | trajectory Ai by making a ridge valley degree into a Y-axis and making distance into an X-axis. It is explanatory drawing of tree high grid data. It is explanatory drawing explaining calculation of the vertex candidate by ridge valley degree. It is explanatory drawing of the vertex candidate Dgi. It is explanatory drawing of the data Dhi of the upper part of a tree crown. It is explanatory drawing explaining the labeling of a tree crown upper part. It is explanatory drawing of a screen about the determined vertex. It is explanatory drawing until a vertex determination. It is explanatory drawing until a vertex determination. It is explanatory drawing until a vertex determination. It is explanatory drawing of a ridge valley degree. It is explanatory drawing of a ridge valley degree. It is explanatory drawing explaining the relationship between a threshold value and a tree crown part. It is a flowchart explaining other embodiment. It is explanatory drawing explaining the polygon of the tree crown obtained by other embodiment. It is a concrete hardware block diagram of this Embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Database 11 Database 13 Display part 14 Tree height calculation part 17 Ridge degree calculation part 21 Tree vertex candidate extraction part 23 Tree crown upper part extraction part 25 Tree vertex determination part 30 Output part

Claims (9)

A DEM database storing DEM grid data obtained by emitting lasers from above to areas including forests;
A DSM database storing grid data of the local DSM;
First storage means for storing tree height grid data;
Second storage means for storing averaged tree height grid data;
Third storage means for storing feature data of the crown three-dimensional shape;
Fourth storage means for storing feature data of the crown three-dimensional shape after filtering;
Fifth storage means for storing tree vertex candidates;
Providing a sixth storage means for the upper crown of the data is stored,
On the computer,
Reading the grid data of the DEM and the grid data of the DSM, setting the difference between the z values as a tree height, and storing the tree height grid data together with the x and y coordinate values of the grid data in the first storage means When,
Averaging the tree height grid data in a predetermined range centered on each data, and storing the averaged tree height grid data in the second storage means;
For each of the averaged tree height grid data, sequentially obtaining a ridge valley degree within a predetermined range, and storing it in the third storage means together with the x and y coordinate values as feature data of the canopy solid shape; ,
A filter that sequentially replaces each data with the ridge valley degree that is the maximum within a predetermined range is applied to the feature data of the crown three-dimensional shape, and the fourth storage means is used as the feature data of the crown three-dimensional shape after the filter. The step of storing in
Comparing the filtered crown crown shape feature data and the crown crown shape feature data, extracting the same as the tree vertex candidates and storing them in the fifth storage means;
Storing, in the sixth storage means, data on the crown crown shape characteristic data that is equal to or higher than a threshold value according to the state of the forest, as data on the crown top;
In the upper crown data, an adjacent cluster is set as one crown upper area, and the largest one of the tree vertex candidates existing in the upper crown area is determined as a tree vertex;
Trees vertex recognition method characterized by causing I rows and outputting the coordinates of the display or the tree vertexes on the display unit by the trees apex in a predetermined shape of the image. The calculation of the ridge valley degree is
To the computer,
Determining the ground opening and underground opening within the predetermined range for the averaged tree height grid data;
Obtaining the angle between the bisector of the angle between the ground opening and the underground opening and the horizontal line as the ridge valley degree with respect to the obtained ground opening and the underground opening; trees vertex recognition method of claim 1, wherein the calculating by the cause I line. To the computer,
When there are a plurality of maximum tree vertex candidates existing in the upper tree crown area, the maximum one is a temporary vertex, and an image of the temporary vertex is displayed on the display unit; ,
The image of a vertex of the temporary, trees vertices recognition method according to claim 1 or 2, wherein the cause I rows and determining the vertices of the temporary designated by the operator as the tree apex. Preparing a seventh storage means for storing the orthophoto image of the area;
To the computer,
Reading out the orthophoto image of the area from the seventh storage means, and displaying the image of a predetermined shape of the tree apex or the coordinates of the tree apex on the orthophoto image of the area on the display unit trees vertex recognition method according to any one of claims 1, 2 or 3, characterized in that cause I lines and. D that stores DEM grid data obtained by launching lasers from above to areas including forests
EM database,
A DSM database storing grid data of the local DSM;
First storage means for storing tree height grid data;
Second storage means for storing averaged tree height grid data;
Third storage means for storing feature data of the crown three-dimensional shape;
Fourth storage means for storing feature data of the crown three-dimensional shape after filtering;
Fifth storage means for storing tree vertex candidates;
By using the sixth storage means crown overlying data is stored,
On the computer,
Means for reading the grid data of the DEM and the grid data of the DSM, setting the difference between the z values as a tree height, and storing the tree height grid data together with the x and y coordinate values of the grid data in the first storage means ,
Means for averaging the tree height grid data in a predetermined range centered on each data, and storing the averaged tree height grid data in the second storage means;
Means for sequentially obtaining the ridge valley degree within a certain range for each of the averaged tree height grid data, and storing it in the third storage means together with the x and y coordinate values as feature data of the crown solid shape;
A filter that sequentially replaces each data with the ridge valley degree that is the maximum within a predetermined range is applied to the feature data of the crown three-dimensional shape, and the fourth storage means is used as the feature data of the crown three-dimensional shape after the filter. Means to memorize,
Means for comparing the feature data of the crown three-dimensional shape after the filter with the feature data of the crown three-dimensional shape, extracting the same as the tree vertex candidates and storing them in the fifth storage means;
Means for storing, in the sixth storage means, data on the canopy three-dimensional shape that is equal to or higher than a threshold corresponding to the state of the forest, as data on the upper part of the canopy;
Means for determining an adjacent cluster as one crown upper area in the upper crown data, and determining the largest one of the tree vertex candidates existing in the crown upper area as a tree vertex;
A tree vertex recognition program for executing a function as a means for displaying the tree vertex as an image of a predetermined shape on a display unit or outputting the coordinates of the tree vertex. In the computer,
Means for determining the ground opening and underground opening within the certain range for the averaged tree height grid data,
As means for obtaining the ridge valley degree as an angle between the bisector of the angle between the ground opening and the underground opening and a horizontal line with respect to the obtained ground opening and the underground opening. The tree vertex recognition program according to claim 5 , wherein the degree of ridge valley is calculated by executing the function. In the computer,
In the case where there are a plurality of maximum tree vertex candidates present in the upper tree crown area, the maximum one is a temporary vertex, and means for displaying an image of this temporary vertex,
The tree vertex recognition program according to claim 5 or 6, for executing a function as means for determining the temporary vertex designated by an operator as the tree vertex in the image of the tree upper area. Using the seventh memory means for storing the orthophoto image of the region,
In the computer,
As means for reading out the orthophoto image of the area from the seventh storage means and displaying the image of a predetermined shape of the tree apex or the coordinates of the tree apex on the orthophoto image of the area on the display unit The program for recognizing a tree vertex according to any one of claims 5, 6 and 7, for executing the function of D that stores DEM grid data obtained by launching lasers from above to areas including forests
EM database,
A DSM database storing grid data of the local DSM;
First storage means for storing tree height grid data;
Second storage means for storing averaged tree height grid data;
Third storage means for storing feature data of the crown three-dimensional shape;
Fourth storage means for storing feature data of the crown three-dimensional shape after filtering;
Fifth storage means for storing tree vertex candidates;
Sixth storage means for storing data of the upper part of the tree crown;
Means for reading the grid data of the DEM and the grid data of the DSM, setting the difference between the z values as a tree height, and storing the tree height grid data together with the x and y coordinate values of the grid data in the first storage means When,
Means for averaging the tree height grid data in a predetermined range centered on each data, and storing the averaged tree height grid data in the second storage means;
Means for sequentially obtaining a ridge valley degree within a predetermined range for each of the averaged tree height grid data, and storing the obtained data together with the x and y coordinate values in the third storage means as feature data of the crown crown shape; ,
A filter that sequentially replaces each data with the ridge valley degree that is the maximum within a predetermined range is applied to the feature data of the crown three-dimensional shape, and the fourth storage means is used as the feature data of the crown three-dimensional shape after the filter. Means for storing
Means for comparing feature data of the crown three-dimensional shape after the filter and feature data of the crown three-dimensional shape, extracting the same as the tree vertex candidates and storing them in the fifth storage means;
Means for storing, in the sixth storage means, data above the threshold corresponding to the state of the forest in the canopy solid shape feature data as data on the top of the canopy;
Means for determining an adjacent cluster as one crown upper area in the data on the crown top, and determining the largest one of the tree vertex candidates existing in the crown upper area as a tree vertex;
A tree vertex recognition apparatus comprising: means for converting the tree vertex into an image of a predetermined shape and displaying it on a display unit or outputting the coordinates of the tree vertex.