JP7613093B2 - Information processing device and information processing program - Google Patents

Description

The present invention relates to an information processing device and an information processing program.

For example, Patent Document 1 describes a biometric authentication device that determines whether performance has deteriorated due to dirt. This biometric authentication device includes an authentication unit that uses an imaging unit to capture an image of the biometric subject to authentication and performs biometric authentication by comparing this image data with pre-prepared registered data, a dirt detection unit that detects the amount of dirt on the imaging unit, a dirt determination condition calculation unit that determines dirt determination conditions based on a history of the amount of dirt detected by the dirt detection unit and a matching result history in the authentication unit, and a dirt determination unit that determines whether there is dirt that will cause a deterioration in authentication performance based on the dirt determination conditions and the amount of dirt detected at that time.

Patent Document 2 also describes a determination system that notifies the authenticity of an object and the authenticity of the owner of the object. This determination system has a storage means that associates and stores the registered identification information of the registered object and a registered image of a specific location of the registered object, an acquisition means that acquires a designated image of a location designated by the user from images of the presented object and acquires from the storage means a registered image associated with the same registered identification information as the presented identification information of the presented object, a determination means that determines whether the acquired designated image matches the registered image, and a notification means that notifies the determination result of the determination means.

JP 2010-129045 A Patent No. 6481754

There is a system that uniquely identifies an object by capturing an image of a random pattern on the surface of the object and comparing a registered image that represents the captured random pattern with a match image that is used for matching.

In this case, the quality of the registered image may be poor due to deterioration of the imaging device that captures the object, changes in the imaging environment, the material of the object, etc., which may affect the recognition performance. However, depending on the level of image quality, the number of registered images for one object, the level of image quality of other areas of one object, etc., it is desirable to register an image even if its image quality is somewhat poor.

The present disclosure aims to provide an information processing device and an information processing program that, when registering a registration image that is an image showing a random pattern on the surface of an object to be uniquely identified, can register the registration image as a registration image even if the image quality of the registration image is poor, as long as the image can be uniquely identified.

To achieve the above object, the information processing device according to the first aspect includes a processor, which extracts multiple regions to be used when performing unique identification from an image representing an object to be uniquely identified, and determines whether or not to register each of the multiple regions as a registered image based on an index value representing the image quality of each of the multiple extracted regions and two or more thresholds determined by at least one of the characteristics of the object and the characteristics of a photographing device that photographs the object.

In addition, the information processing device according to the second aspect is an information processing device according to the first aspect, in which the two or more thresholds include a first threshold for determining, for an index value representing the image quality, a normal value indicating that the image quality is normal and a first abnormal value indicating that the image quality is abnormal, and a second threshold for determining, for the index value representing the image quality, the first abnormal value and a second abnormal value that is more abnormal than the first abnormal value, and a first range of equal to or greater than the first threshold and less than the second threshold or equal to or greater than the second threshold and less than the first threshold is a range that does not degrade the identification performance of unique identification, and a second range of equal to or greater than the second threshold or less than the second threshold is a range that degrades the identification performance of unique identification.

In addition, in the information processing device according to the third aspect, in the information processing device according to the second aspect, the processor determines that, if the index value representing the image quality is within the first range, the area corresponding to the index value representing the image quality is to be registered as a registered image, and, if the index value representing the image quality is within the second range, the processor determines that the area corresponding to the index value representing the image quality is not to be registered as a registered image.

In addition, in the information processing device according to the fourth aspect, in the information processing device according to the second or third aspect, the two or more threshold values are determined by evaluating and stratifying the matching scores of a sample image group including a plurality of images that satisfy predetermined shooting conditions.

In addition, in the information processing device according to the fifth aspect, in the information processing device according to any one of the second to fourth aspects, the processor registers the two or more thresholds for each combination of the object and the photographing device, and acquires the two or more corresponding thresholds by specifying the identification information of the object, the identification information of the photographing device, and the identification information of the photographing location of the object, which are associated in advance.

In addition, the information processing device according to the sixth aspect is an information processing device according to any one of the first to fifth aspects, in which the index value representing the image quality is at least one of an index value indicating the degree of blur in the region, an index value indicating the degree of shaking in the region, an index value indicating the brightness in the region, and an index value indicating randomness in the region.

In addition, the information processing device according to the seventh aspect is an information processing device according to the sixth aspect, in which the index value indicating the degree of blur is expressed as the standard deviation of Laplacian filter values, the index value indicating the degree of shaking is expressed as the difference between the maximum and minimum values of the standard deviations of Sobel filter values in four directions, the index value indicating brightness is expressed as the average luminance, and the index value indicating randomness is expressed as a value obtained by cutting out a central 1/4-size portion of the image, finding correlation values with an image of the same size starting from each coordinate of the image, subtracting the average value from the maximum value of the correlation value group, and dividing the result by the standard deviation.

In addition, the information processing device according to the eighth aspect is an information processing device according to any one of the first to seventh aspects, in which the multiple regions are extracted for each image obtained by photographing each of multiple locations of the object.

Furthermore, in order to achieve the above object, the information processing program according to the ninth aspect causes a computer to extract multiple regions to be used when performing unique identification from an image representing an object to be uniquely identified, and determine whether or not to register each of the multiple regions as a registered image based on an index value representing the image quality of each of the multiple extracted regions and two or more thresholds determined by at least one of the characteristics of the object and the characteristics of a photographing device that photographs the object.

According to the first and ninth aspects, when a registration image is registered, which is an image showing a random pattern on the surface of an object to be uniquely identified, if the image quality of the registration image is poor but unique identification can be performed, the registration image can be registered.

The second aspect has the effect of being able to identify a range that does not degrade the identification performance of unique identification based on the relationship between the index value representing image quality, the first threshold, and the second threshold.

The third aspect has the effect of being able to determine whether or not to register an image as a registered image based on the relationship between the index value representing image quality, the first threshold, and the second threshold.

The fourth aspect has the advantage that two or more thresholds can be determined using the matching scores of a group of sample images.

The fifth aspect has the advantage that it is possible to obtain two or more corresponding thresholds using identification information of the object, identification information of the imaging device, and identification information of the imaging location of the object.

According to the sixth and seventh aspects, the image quality can be evaluated using an index value indicating the degree of blur, an index value indicating the degree of shaking, an index value indicating brightness, and an index value indicating randomness as index values representing the image quality.

The eighth aspect has the advantage that it is possible to obtain multiple images for each of the multiple regions by photographing each of the multiple regions.

FIG. 1 is a diagram illustrating an example of a configuration of a unique identification system according to an embodiment. 2 is a block diagram showing an example of an electrical configuration of the information processing apparatus according to the embodiment; FIG. 2 is a block diagram showing an example of a functional configuration of the information processing device according to the embodiment. 5 is a flowchart showing an example of a process flow according to an information processing program according to the embodiment. 1A and 1B are diagrams illustrating an example of a registration object and an image according to an embodiment. 1A to 1C are diagrams illustrating an image quality evaluation method according to an embodiment. FIG. 11 is a diagram showing an example of an image quality evaluation result table according to the embodiment. FIG. 4 is a diagram illustrating an example of a setting table according to the embodiment. FIG. 11 is a diagram illustrating an example of a notification mode table according to the embodiment.

An example of a form for implementing the technology of the present disclosure will be described in detail below with reference to the drawings. Note that components and processes that perform the same actions, actions, and functions are given the same reference numerals throughout all the drawings, and duplicated descriptions may be omitted as appropriate. Each drawing is merely a schematic illustration to allow a sufficient understanding of the technology of the present disclosure. Therefore, the technology of the present disclosure is not limited to the illustrated examples. Also, in this embodiment, descriptions of configurations that are not directly related to the technology of the present disclosure or well-known configurations may be omitted.

Figure 1 is a diagram showing an example of the configuration of a unique identification system 100 according to this embodiment.

As shown in FIG. 1, the unique identification system 100 according to this embodiment registers an image of a partial area (e.g., about 10 mm square) of the surface of an object in advance as information specific to that object, and uniquely identifies that the objects being matched are the same if the image of the object being matched resembles the image of the registered object. Note that identifying whether or not objects are the same is referred to as unique identification.

As an example of the object-specific information referred to here, random patterns that are difficult to control and form, such as the distribution of plant fibers that make up paper or the dispersion state of metal particles contained in silver paint, are applied. In the unique identification system 100, this random pattern is optically read and used as information. An image representing a fine random pattern is captured using a capture device such as a smartphone, digital camera, or scanner.

The unique identification system 100 shown in FIG. 1 includes an information processing device 10, a photographing device 20, and a photographing device 30. The photographing device 20 photographs the random pattern on the surface of the registration object 21. The photographing device 30 photographs the random pattern on the surface of the matching object 31. The information processing device 10 registers an image representing the random pattern obtained by photographing the surface of the registration object 21 with the photographing device 20 as a registered image in a database (not shown) in advance. Meanwhile, the information processing device 10 acquires an image representing the random pattern obtained by photographing the surface of the matching object 31 with the photographing device 30 as a matching image. The information processing device 10 compares the matching image, which is the target of unique identification, with a registered image registered in advance in the database. Specifically, the feature amounts of the registered image and the matching image are calculated, and the similarity between the registered image and the matching image calculated from the respective feature amounts is used. If the similarity is equal to or greater than a certain value, it is determined that the registered image and the matching image were read from the same object, that is, the matching object 31 is the same.

Figure 2 is a block diagram showing an example of the electrical configuration of the information processing device 10 according to this embodiment.

As shown in FIG. 2, the information processing device 10 according to this embodiment includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, an input/output interface (I/O) 14, a storage unit 15, a display unit 16, an operation unit 17, and a communication unit 18.

The information processing device 10 according to this embodiment is, for example, a general-purpose computer device such as a server computer or a personal computer (PC).

The CPU 11, ROM 12, RAM 13, and I/O 14 are each connected via a bus. The I/O 14 is connected to various functional units including a memory unit 15, a display unit 16, an operation unit 17, and a communication unit 18. These functional units are capable of communicating with the CPU 11 via the I/O 14.

The control unit is composed of the CPU 11, ROM 12, RAM 13, and I/O 14. The control unit may be configured as a sub-control unit that controls part of the operation of the information processing device 10, or may be configured as part of the main control unit that controls the entire operation of the information processing device 10. For example, an integrated circuit such as an LSI (Large Scale Integration) or an IC (Integrated Circuit) chip set is used for part or all of the blocks of the control unit. An individual circuit may be used for each of the above blocks, or a circuit in which some or all of the blocks are integrated may be used. The above blocks may be provided integrally, or some of the blocks may be provided separately. In addition, a part of each of the above blocks may be provided separately. The integration of the control unit is not limited to an LSI, and a dedicated circuit or a general-purpose processor may be used.

For example, a hard disk drive (HDD), a solid state drive (SSD), a flash memory, or the like is used as the storage unit 15. The storage unit 15 stores an information processing program 15A for executing the image registration process according to this embodiment. Note that this information processing program 15A may be stored in the ROM 12.

The information processing program 15A may be pre-installed in the information processing device 10, for example. The information processing program 15A may be realized by storing it in a non-volatile storage medium or distributing it via a network and installing it appropriately in the information processing device 10. Note that examples of non-volatile storage media include a CD-ROM (Compact Disc Read Only Memory), an optical magnetic disk, a HDD, a DVD-ROM (Digital Versatile Disc Read Only Memory), a flash memory, a memory card, etc.

For example, a liquid crystal display (LCD) or an organic EL (Electro Luminescence) display is used for the display unit 16. The display unit 16 may have an integrated touch panel. The operation unit 17 is provided with devices for operation input, such as a keyboard and a mouse. The display unit 16 and the operation unit 17 receive various instructions from a user of the information processing device 10. The display unit 16 displays various information such as the results of processing executed in response to instructions received from the user and notifications regarding the processing.

The communication unit 18 is connected to a network such as the Internet, a LAN (Local Area Network), or a WAN (Wide Area Network), and is capable of communicating with external devices such as the imaging device 20, the imaging device 30, and an image forming device via the network.

However, as mentioned above, depending on the level of image quality, the number of registered images for an object, the level of image quality of other areas of an object, etc., it is desirable to register an image as a registered image even if the image quality is relatively poor.

When registering a registration image, which is an image showing a random pattern on the surface of an object to be uniquely identified, in a database, the information processing device 10 according to this embodiment extracts multiple regions from the image representing the object, and determines whether or not to register the image of each region as a registration image based on the image quality of the multiple extracted regions and a threshold determined by at least one of the characteristics of the object and the characteristics of the imaging device that captures the object.

Specifically, the CPU 11 of the information processing device 10 according to this embodiment writes the information processing program 15A stored in the storage unit 15 to the RAM 13 and executes it, thereby functioning as each unit shown in FIG. 3. The CPU 11 is an example of a processor.

Figure 3 is a block diagram showing an example of the functional configuration of the information processing device 10 according to this embodiment.

As shown in FIG. 3, the CPU 11 of the information processing device 10 according to this embodiment functions as an acquisition unit 11A, an extraction unit 11B, an image quality evaluation unit 11C, and a threshold determination unit 11D.

The storage unit 15 according to this embodiment stores a threshold database (hereinafter referred to as "threshold DB") 15B. The threshold DB 15B may be stored in an external storage device. The threshold DB 15B stores two or more thresholds for determining whether or not to register each of a plurality of regions extracted from an image representing the registered object 21 as a registered image. The two or more thresholds are thresholds that are determined in advance by the threshold determination unit 11D based on the characteristics of the registered object 21 and the characteristics of the photographing device 20 that photographs the registered object 21. The two or more thresholds are associated with an object ID (Identification) that is identification information of the registered object 21, a photographing device ID that is identification information of the photographing device 20, and a photographing location ID that is identification information of the photographing location of the registered object 21. The two or more thresholds are determined in advance for each combination of the registered object 21 and the photographing device 20. The two or more thresholds may be determined based on either the characteristics of the registered object 21 or the characteristics of the photographing device 20.

The acquisition unit 11A acquires an image obtained by photographing the surface of the registered object 21 with the photographing device 20. The image referred to here is an image representing the random pattern described above. The photographing device 20 may photograph multiple different locations of the registered object 21, or may photograph only one location. There may be one photographing device 20, or multiple photographing devices. The photographing device 20 may capture multiple images for one photographing location. When photographs are taken at multiple different locations, multiple images are acquired for each of the multiple photographing locations.

The extraction unit 11B extracts multiple regions from the image acquired by the acquisition unit 11A. Note that if multiple images are acquired for each of multiple shooting locations, multiple regions are extracted from each of the multiple images.

The image quality evaluation unit 11C calculates an index value (hereinafter referred to as "image quality index value") representing the image quality of each of the multiple regions extracted by the extraction unit 11B. The image quality evaluation unit 11C determines whether or not to register each of the multiple regions as a registered image based on the image quality index value for each of the multiple regions and two or more thresholds acquired from the threshold DB 15B. The image quality evaluation unit 11C further determines a notification mode representing a mode for notifying an administrator or the like of the status of the images included in the dataset.

As described above, the threshold determination unit 11D determines two or more thresholds for each combination of the registration object 21 and the image capture device 20, and registers the determined two or more thresholds in the threshold DB 15B.

Next, the operation of the information processing device 10 according to this embodiment will be described with reference to FIG.

Figure 4 is a flowchart showing an example of the processing flow of information processing program 15A according to this embodiment.

First, when the information processing device 10 is instructed to execute an image registration process, the information processing program 15A is started by the CPU 11 and executes the following steps.

In step S101 of FIG. 4, the CPU 11 acquires an image representing the registration object 21, as shown in FIG. 5 as an example.

Figure 5 is a schematic diagram showing an example of a registration object 21 and an image according to this embodiment.

As shown in FIG. 5, multiple different locations of the registered object 21 are photographed multiple times by the photographing device 20. In the example of FIG. 5, multiple images (hereinafter referred to as "image A") are acquired for the photographing location 21A of the registered object 21, and multiple images (hereinafter referred to as "image B") are acquired for the photographing location 21B of the registered object 21.

In step S102, the CPU 11 extracts multiple regions from the images acquired in step S101. Specifically, in the example of FIG. 5 described above, multiple regions A-a, Ab, and A-c are extracted from each of the multiple images A, and multiple regions B-a, B-b, and B-c are extracted from each of the multiple images B.

In step S103, the CPU 11 evaluates the image quality of each of the multiple regions extracted in step S102. Specifically, in the example of FIG. 5 described above, the image quality is evaluated for each of the multiple regions A-a, Ab, and A-c extracted from each of the multiple images A. Similarly, the image quality is evaluated for each of the multiple regions B-a, B-b, and B-c extracted from each of the multiple images B. An index value that can estimate the cause of deterioration in the shooting conditions is applied as the image quality index value used to evaluate the image quality.

As an example, the image quality index value may be at least one of an index value indicating the degree of blur in an area, an index value indicating the degree of shaking in an area, an index value indicating the brightness in an area, and an index value indicating randomness in an area. The index value indicating the degree of blur is expressed, for example, as the standard deviation of the Laplacian filter value. The index value indicating the degree of shaking is expressed, for example, as the difference between the maximum and minimum standard deviations of the Sobel filter values in four directions. The index value indicating brightness is expressed, for example, as the average brightness value. The index value indicating randomness is expressed, for example, as a value (self-randomness evaluation value) obtained by cutting out a portion of the center 1/4 size of the image, finding correlation values with an image of the same size starting from each coordinate of the image, subtracting the average value from the maximum value of the correlation value group, and dividing the result by the standard deviation.

As an example, image quality is evaluated by comparing the image quality index value with two or more thresholds, as shown in FIG. 6. Here, as described above, the two or more thresholds are pre-registered in the threshold DB 15B for each combination of the registered object 21 and the photographing device 20. For example, the CPU 11 accepts the designation of the pre-associated object ID, photographing device ID, and photographing location ID according to an input operation by the user, and obtains the corresponding two or more thresholds from the threshold DB 15B.

Figure 6 is a diagram used to explain the image quality evaluation method according to this embodiment.

As shown in FIG. 6, the two or more thresholds are expressed as, for example, a first threshold and a second threshold. In the example of FIG. 6, a higher image quality index value indicates better image quality. The first threshold is a threshold for determining a normal value and a first abnormal value. The normal value indicates that the image quality is normal, and the first abnormal value indicates that the image quality is abnormal. The second threshold is a threshold for determining a first abnormal value and a second abnormal value. The second abnormal value indicates that the degree of abnormality is greater than that of the first abnormal value. In other words, the first range from the second threshold to the first threshold is a range in which the shooting conditions may deteriorate but the identification performance of the unique identification is not reduced. The second range less than the second threshold is a range in which the shooting conditions may deteriorate and the identification performance of the unique identification is reduced. Note that the third range from the first threshold to the first threshold is a normal range in which there is no possibility of deterioration of the shooting conditions and no deterioration of the identification performance of the unique identification.

In this case, if the image quality index value of a certain area is within the first range, the image quality of the area is determined to be "not recommended", if the image quality index value of a certain area is within the second range, the image quality of the area is determined to be "not applicable", and if the image quality index value of a certain area is within the third range, the image quality of the area is determined to be "good". Note that in the example of Figure 6, "good" corresponds to a circle, "not recommended" corresponds to a triangle, and "not applicable" corresponds to an x.

Here, the first threshold and the second threshold are determined by evaluating and stratifying the matching scores of a sample image group including a plurality of images that satisfy predetermined shooting conditions. The sample images of the sample image group include, for example, images of poor image quality collected in advance, simulation images of poor image quality collected in advance, images of poor image quality collected during operation, and the like. For example, a circle indicates a sample image that satisfies the condition that the matching score is at a level equivalent to the still shooting conditions. This condition is, for example, within 3σ (σ: standard deviation) of the matching score for 100 times for the same individual under still shooting conditions. A cross indicates a sample image that satisfies the condition that the matching score is at an indistinguishable level. This condition is, for example, within 3σ of the matching score for 100 times for individuals under different still shooting conditions. A triangle indicates a sample image that is intermediate between the sample image with a circle and the sample image with a cross, and is suspected of having deteriorated shooting conditions. In this way, the first threshold and the second threshold are determined by stratifying the sample image group collected in advance.

Also, contrary to the example of FIG. 6, if a lower image quality index value indicates better image quality, the relationship between the normal value and the second abnormal value and the relationship between the first threshold and the second threshold can be reversed. That is, from the left, the normal value, the first threshold, the first abnormal value, the second threshold, and the second abnormal value are displayed. In this case, the first range equal to or greater than the first threshold and less than the second threshold is a range in which there is a possibility of deterioration in the shooting conditions, but which does not degrade the identification performance of the unique identification. The second range equal to or greater than the second threshold is a range in which there is a possibility of deterioration in the shooting conditions, which degrades the identification performance of the unique identification. The third range less than the first threshold is a normal range in which there is no possibility of deterioration in the shooting conditions and no degrade in the identification performance of the unique identification.

The image quality evaluation method using the above-mentioned four image quality index values will be specifically described. For example, the standard deviation of the Laplacian filter value is the first image quality index value, the difference between the maximum and minimum standard deviations of the Sobel filter values in four directions is the second image quality index value, the average brightness is the third image quality index value, and a portion of the image that is 1/4 the size of the center is cut out and correlation values with an image of the same size starting from each coordinate of the image are calculated, and the value obtained by subtracting the average value from the maximum value of the correlation value group and dividing the result by the standard deviation (self-randomness evaluation value) is set as the fourth image quality index value. The first image quality index value to the fourth image quality index value are used to determine the overall image quality of each area as "good", "not recommended", or "not applicable".

First, for each of the first to fourth image quality index values, the first and second thresholds shown in FIG. 6 are used to determine whether the image quality index is "good", "not recommended", or "not applicable". Next, an evaluation is obtained for combinations of two image quality index values. That is, all combinations of two image quality index values, such as the combination of the first image quality index value and the second image quality index value, the combination of the first image quality index value and the third image quality index value, ... are determined to be "good", "not recommended", or "not applicable". For example, a combination of two image quality index values that includes "not applicable" is determined to be "not applicable". A combination of two image quality index values that are both "good" is determined to be "good". A combination of two image quality index values where one is "not recommended" and the other is "good" is determined to be "not recommended". A combination of two image quality index values where both are "not recommended" is determined to be "not recommended" or "not applicable" using the image quality index values and a data table (not shown) prepared in advance.

Then, in the evaluation of the combinations obtained above, if there is one or more "not applicable", the overall evaluation is judged as "not applicable", if all are "good", it is judged as "good", and otherwise it is judged as "not recommended". This overall evaluation becomes the evaluation for one area.

In step S104, the CPU 11 selects the image that can be registered in each area and has the best image quality based on the image quality evaluation results in step S103. Images that can be registered in a certain area are considered to be "good" or "not recommended". If there are multiple "good" images in a certain area, the "good" image with the smallest input order is used as the image with the best image quality. If there are no "good" images and multiple "not recommended" images, the image quality index value is evaluated for its closeness to the "not recommended" threshold, and the image with the smallest total value of all image quality index values is selected. If the image quality index value is V, the "not recommended" threshold (first threshold) is TH1, and the "not applicable" threshold (second threshold) is TH2, the closeness (distance) D to the "not recommended" threshold (first threshold) can be calculated, for example, by the following formula (1):

D=(V-TH1)/(TH2-TH1)...(1)

In step S105, the CPU 11 determines a registration image for each image of the multiple regions selected in step S104 according to a predetermined setting. Specifically, for each of the multiple regions selected in step S104, if the image quality index value is within the first or third range (i.e., "not recommended" or "good") according to the "settings" included in the image quality evaluation result table shown in FIG. 7 described below, for example, the CPU 11 determines to register the region as a registration image, and if the image quality index value of a certain region is within the second range (i.e., "not applicable"), the CPU 11 determines not to register the region as a registration image.

In step S106, the CPU 11 determines the notification mode of the data set at each shooting location according to, for example, the "result" included in the image quality evaluation result table shown in FIG. 7 described later. As an example, the notification mode is determined to be one of "normal", "warning", and "error". If all images in the data set are "good", the notification mode is determined to be "normal", if the data set contains one or more "not recommended" images, the notification mode is determined to be "warning", and if there are no registered images in the data set, the notification mode is determined to be "error". In the example of FIG. 5 described above, the data set is one data set consisting of areas A-a, A-b, and A-c for image A, and one data set consisting of areas B-a, B-b, and B-c for image B. A flag indicating a setting number described later is assigned to each of these data sets in advance.

In step S107, the CPU 11 determines the overall notification mode for the registration object 21 according to, for example, a notification mode table shown in FIG. 9 described below, and ends the series of processes by the information processing program 15A. Once the notification mode is determined, one of "Normal", "Warning", or "Error" is notified to the terminal device of the administrator, etc., by email, etc.

Next, we will explain in detail how to determine the registration image and notification mode with reference to Figures 7 to 9.

Figure 7 shows an example of an image quality evaluation result table 15C according to this embodiment.

The image quality evaluation result table 15C shown in FIG. 7 is stored in the storage unit 15, as an example. The image quality evaluation result table 15C may be stored in an external storage device. The image quality evaluation result table 15C includes an item called "image quality evaluation result of each area." As an example, the "image quality evaluation result of each area" registers the image quality evaluation results of areas A-a, A-b, and A-c in each of multiple images A obtained by photographing the shooting location 21A shown in FIG. 5 multiple times. As an example of the image quality legend, a circle indicates "good," a triangle indicates "not recommended," and an x indicates "not applicable."

The image quality evaluation result table 15C further includes an item called "settings." As an example, the setting numbers included in the setting table 15D shown in FIG. 8 are registered in this "settings."

Figure 8 shows an example of a setting table 15D according to this embodiment.

The setting table 15D shown in FIG. 8 is stored in the storage unit 15 as an example. The setting table 15D may be stored in an external storage device. The setting table 15D includes items such as "Can non-recommended items be saved?" and "Can a dataset including non-target items be saved?" for each setting number. Setting numbers 1 to 4 in the example of FIG. 8 indicate flags assigned to datasets by the user, for example. Setting number 1 is set to "Can non-recommended items be saved" and "Can a dataset including non-target items be saved." Note that "save" here is synonymous with "register." Setting number 2 is set to "Cannot non-recommended items be saved" and "Can a dataset including non-target items be saved." Setting number 3 is set to "Can non-recommended items be saved" and "Can a dataset including non-target items be saved" cannot be saved. Setting number 4 is set to "Cannot non-recommended items be saved" and "Can a dataset including non-target items be saved" cannot be saved.

The image quality evaluation result table 15C shown in FIG. 7 further includes an item called "Result". This "Result" includes items called "Saved image" and "Notification mode of image A". In other words, the "Saved image" and "Notification mode of image A" are determined by the "image quality evaluation result of each area" and "Setting". In the example of FIG. 7, if the evaluation results of areas A-a, A-b, and A-c are all "Good" and setting numbers 1, 2, 3, and 4 (see FIG. 8) are set, all of areas A-a, A-b, and A-c are determined as saved images, and the notification mode of image A is determined as "Normal". Also, if the evaluation result of area A-a is "Good", the evaluation results of areas A-b and A-c are "Not recommended", and setting numbers 1 and 3 (see FIG. 8) are set, all of areas A-a, A-b, and A-c are determined as saved images, and the notification mode of image A is determined as "Warning".

Similarly, if the evaluation result of area A-a is "good", the evaluation results of areas A-b and A-c are "not recommended", and setting numbers 2 and 4 (Figure 8) are set, area A-a is determined as the saved image, and the notification mode of image A is determined as "warning". If the evaluation results of areas A-a, A-b, and A-c are all "not recommended", and setting numbers 2 and 4 (Figure 8) are set, areas A-a, A-b, and A-c are determined as not having saved images, and the notification mode of image A is determined as "error". If the evaluation result of area A-a is "good", the evaluation result of area A-b is "not recommended", the evaluation result of area A-c is "not applicable", and setting number 1 (Figure 8) is set, areas A-a and A-b are determined as the saved images, and the notification mode of image A is determined as "warning". Also, if the evaluation result of area A-a is "good", the evaluation result of area A-b is "not recommended", the evaluation result of area A-c is "not applicable", and setting number 2 (Figure 8) is set, area A-a is determined as the saved image, and the notification mode of image A is determined as "warning". Also, if the evaluation result of area A-a is "good", the evaluation result of area A-b is "not recommended", the evaluation result of area A-c is "not applicable", and setting numbers 3 and 4 (Figure 8) are set, areas A-a, A-b, and A-c are determined as no saved images, and the notification mode of image A is determined as "error". Also, if the evaluation results of areas A-a, A-b, and A-c are all "not applicable", and setting numbers 1, 2, 3, and 4 (Figure 8) are set, areas A-a, A-b, and A-c are determined as no saved images, and the notification mode of image A is determined as "error".

Figure 9 shows an example of a notification mode table 15E according to this embodiment.

The notification mode table 15E shown in FIG. 9 is stored in the storage unit 15 as an example. The notification mode table 15E may be stored in an external storage device. In the notification mode table 15E, the notification mode of the entire registered object 21 is registered for the combination of the notification mode of the image A and the notification mode of the image B. In the example of FIG. 9, when the notification mode of the image A is "normal" and the notification mode of the image B is "normal", the notification mode of the entire is determined to be "normal". When the notification mode of the image A is "including not recommended or not applicable" and the notification mode of the image B is "normal", the notification mode of the entire is determined to be "warning". When the notification mode of the image A is "including not recommended or not applicable" and the notification mode of the image B is "including not recommended or not applicable", the notification mode of the entire is determined to be "warning". When the notification mode of the image A is "including not recommended or not applicable" and the notification mode of the image B is "including not recommended or not applicable", the notification mode of the entire is determined to be "warning". When the notification mode of the image A is "including not recommended or not applicable" and the notification mode of the image B is "error", the notification mode of the entire is determined to be "error".

As a result, one of the notification modes, "Normal", "Warning", or "Error", is displayed. A "Warning" notifies an administrator that poor quality images are included, so that, for example, maintenance of the shooting environment can be carried out in a planned manner without suddenly stopping a production line, etc. Images to be saved (registered images) are determined, for example, based on the settings in the above-mentioned setting table 15D, that is, "Can non-recommended images be saved?" and "Can a data set including non-targets be saved?". For objects that have a large area to be cut out, the impact on the identification performance of unique identification is suppressed even if "non-recommended images" are included.

In addition, a location with poor image quality during registration can be identified, and if that location has good image quality and identification fails (or is determined to be a fake) during matching, it can be determined to be an error due to poor photography rather than a failed identification (fake). Note that photography can be repeated during matching until other locations are determined to have good image quality. In the case of authenticity determination (one-to-one matching), locations with poor image quality are identified from the history data of the registered individual in question, and in the case of one-to-many matching, locations with a high rate of poor image quality are identified from the history data of multiple individuals in the range to be matched.

The image quality evaluation method according to this embodiment can also be applied to cases where the photographing conditions at the time of registration are different from those at the time of comparison, even for the same object. For example, if continuous photographs are taken mechanically with a photographing device installed on the production line during registration, and photographs are taken by an operator with a smartphone or the like during comparison, sampling is performed at the time of registration and comparison, and different thresholds are used for registration and comparison. In addition, the storage settings may be strict for the photographing device used for registration, and relaxed for the photographing device used for comparison (storing even if the image quality is not good).

The image quality evaluation method according to this embodiment can be applied to the same imaging device when different product numbers are manufactured on a production line, even though the characteristics of the random pattern of the object are different. For example, when photographing the random pattern of a mark applied to glass, different threshold values are used for sandblasting and black ceramic printing.

In addition, when images taken at multiple locations are registered as the same individual, the image quality evaluation method according to this embodiment can be applied with the same program (one request) even though the characteristics of the random pattern differ depending on the location. For example, when a sand mold for a casting and a finished product are registered as the same individual, different thresholds are used.

The image quality evaluation method according to this embodiment can be applied to cut-out images in which the entire surface is a random pattern. In this case, the image quality that affects the identification performance is evaluated, improving the accuracy of the image quality evaluation. For example, an area of 115 mm x 72 mm is photographed due to the angle of view of a camera, which is an example of an imaging device, but the area used for identification is, for example, approximately 10 mm x 10 mm. When targeting a casting with irregularities, it is possible to ignore blurring other than the area used for identification. This allows randomness to be evaluated.

In addition, the notification based on the notification mode according to this embodiment may include, for example, the location where the poor image quality occurred, the cause of the poor image quality estimated from the image quality evaluation results, etc. This reduces the time required for maintenance. Note that possible causes of the poor image quality include, for example, blurring, shaking, poor lighting, and the presence of dust. The likelihood that the cause of the poor image quality is due to the image quality index value used in the image quality evaluation may be displayed based on the relationship between the threshold value and the image quality index value.

In addition, one shooting location may be photographed multiple times, and the image quality may be evaluated each time an image is photographed. When a predetermined number of the cropped images (i.e., images of the cropped area) are determined to have good image quality, shooting at that shooting location may be terminated. This reduces the time required for unique identification compared to selecting cropped images after a predetermined number of photographs. Also, this reduces the time required for unique identification compared to adjusting the shooting system until all cropped images have good image quality.

In addition, when performing unique identification, priority may be given to the results of matching images in areas of good image quality in determining the individual ID. This makes it less likely that misidentification will occur. For example, first, calculation (determination) of unique identification may be performed only on registered images with good image quality, and if an individual ID cannot be determined, calculation (determination) of unique identification may be performed on registered images with good image quality and those that are not recommended.

Thus, according to this embodiment, when a registration image that is an image showing a random pattern on the surface of an object to be uniquely identified is registered in a database, multiple regions are extracted from the image representing the object, and whether or not to register the image of each region as a registration image is determined based on the image quality of the multiple extracted regions and a threshold determined by at least one of the characteristics of the object and the characteristics of the imaging device that captures the object. Therefore, even if the image quality of the registration image that is an image showing a random pattern on the surface of an object to be uniquely identified is poor, it will be registered as a registration image if unique identification can be performed.

In the above embodiments, the term "processor" refers to a processor in a broad sense, including general-purpose processors (e.g., CPU: Central Processing Unit, etc.) and dedicated processors (e.g., GPU: Graphics processing Unit, ASIC: Application Specific Integrated Circuit, FPGA: Field Programmable Gate Array, programmable logic device, etc.).

In addition, the processor operations in each of the above embodiments may not only be performed by a single processor, but may also be performed by multiple processors located at physically separate locations working together. Furthermore, the order of each processor operation is not limited to the order described in each of the above embodiments, and may be changed as appropriate.

The above describes an information processing device according to an embodiment. The embodiment may be in the form of a program for causing a computer to execute the functions of each unit of the information processing device. The embodiment may be in the form of a non-transitory storage medium that stores these programs and is readable by a computer.

In addition, the configuration of the information processing device described in the above embodiment is merely an example, and may be modified according to circumstances without departing from the spirit of the invention.

The processing flow of the program described in the above embodiment is also an example, and unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the program.

In the above embodiment, a case has been described in which the processing according to the embodiment is realized by a software configuration using a computer by executing a program, but this is not limited to this. The embodiment may be realized, for example, by a hardware configuration or a combination of a hardware configuration and a software configuration.

10 Information processing device 11 CPU
11A Acquisition unit 11B Extraction unit 11C Image quality evaluation unit 11D Threshold value determination unit 12 ROM
13 RAM
14 I/O
15 Storage unit 15A Information processing program 15B Threshold DB
15C: image quality evaluation result table 15D: setting table 15E: notification mode table 16: display unit 17: operation unit 18: communication unit 20, 30: image capture device 21: registered object 31: matching object 100: unique identification system

Claims (9)

A processor is provided.
The processor,
extracting a plurality of regions to be used when performing unique identification from an image representing a random pattern on the surface of an object to be uniquely identified, and determining whether or not to register each of the plurality of regions as a registered image representing the random pattern on the surface of the object based on an index value representing the image quality of each of the extracted plurality of regions and two or more threshold values determined by at least one of the characteristics of the object and the characteristics of an imaging device that images the object ;
the two or more threshold values include a first threshold value for determining whether the index value representing the image quality is a normal value indicating that the image quality is normal or a first abnormal value indicating that the image quality is abnormal, and a second threshold value for determining whether the index value representing the image quality is the first abnormal value or a second abnormal value indicating a greater degree of abnormality than the first abnormal value,
a first range of the first threshold value or more and less than the second threshold value or the second threshold value or more and less than the first threshold value is a range that does not degrade the identification performance of unique identification,
A second range that is equal to or greater than the second threshold value or less than the second threshold value is a range that reduces the identification performance of unique identification.
Information processing device. The processor determines, when the index value representing the image quality is within the first range, to register an area corresponding to the index value representing the image quality as a registered image;
The information processing apparatus according to claim 1 , further comprising: a step of: determining, when the index value representing image quality is included in the second range, not to register an area corresponding to the index value representing image quality as a registered image. The information processing apparatus according to claim 1 , wherein the two or more threshold values are determined by evaluating and stratifying a matching score of a sample image group including a plurality of images that satisfy a predetermined shooting condition. The information processing device of any one of claims 1 to 3, wherein the processor registers the two or more thresholds for each combination of the object and the photographing device, and acquires the two or more corresponding thresholds by specifying identification information of the object, identification information of the photographing device , and identification information of the photographing location of the object , which are associated in advance. A processor is provided.
The processor,
extracting a plurality of regions to be used when performing unique identification from an image representing a random pattern on the surface of an object to be uniquely identified, and determining whether or not to register each of the plurality of regions as a registered image representing the random pattern on the surface of the object based on an index value representing the image quality of each of the extracted plurality of regions and two or more threshold values determined by at least one of the characteristics of the object and the characteristics of an imaging device that images the object;
The index value representing the image quality is at least one of an index value indicating a degree of blur in the region, an index value indicating a degree of shaking in the region, an index value indicating a brightness in the region, and an index value indicating randomness in the region.
Information processing device. The index value indicating the degree of blur is expressed as a standard deviation of Laplacian filter values,
the index value indicating the degree of blur is expressed as a difference between a maximum value and a minimum value of a standard deviation of the Sobel filter values in four directions,
The brightness index value is expressed as an average value of luminance,
The information processing device according to claim 5, wherein the index value indicating the randomness is expressed as a value obtained by cutting out a central 1/4 size portion of the image, calculating correlation values with an image of the same size starting from each coordinate of the image, subtracting the average value from the maximum value of the correlation values, and dividing the result by the standard deviation. The information processing device according to any one of claims 1 to 6 , wherein the plurality of regions are extracted for each image obtained by photographing each of a plurality of locations of the object. extracting a plurality of regions to be used when performing unique identification from an image representing a random pattern on the surface of an object to be uniquely identified, and determining whether or not to register each of the plurality of regions as a registered image representing the random pattern on the surface of the object based on an index value representing the image quality of each of the extracted plurality of regions and two or more threshold values determined by at least one of the characteristics of the object and the characteristics of an imaging device that images the object ;
the two or more threshold values include a first threshold value for determining whether the index value representing the image quality is a normal value indicating that the image quality is normal or a first abnormal value indicating that the image quality is abnormal, and a second threshold value for determining whether the index value representing the image quality is the first abnormal value or a second abnormal value indicating a greater degree of abnormality than the first abnormal value,
a first range of the first threshold value or more and less than the second threshold value or the second threshold value or more and less than the first threshold value is a range that does not degrade the identification performance of unique identification,
A second range that is equal to or greater than the second threshold value or less than the second threshold value is a range that reduces the identification performance of unique identification ,
An information processing program for execution by a computer. extracting a plurality of regions to be used when performing unique identification from an image representing a random pattern on the surface of an object to be uniquely identified, and determining whether or not to register each of the plurality of regions as a registered image representing the random pattern on the surface of the object based on an index value representing the image quality of each of the extracted plurality of regions and two or more threshold values determined by at least one of the characteristics of the object and the characteristics of an imaging device that images the object;
the index value representing the image quality is at least one of an index value indicating a degree of blur in the region, an index value indicating a degree of shaking in the region, an index value indicating a brightness in the region, and an index value indicating a randomness in the region;
An information processing program for execution by a computer.