Camera placement optimization system, method, and program

Abstract

We provide a camera placement optimization system that can optimize the placement of cameras that capture the target object. [Solution] The virtual space construction unit constructs a 3D virtual space that mimics the behavior of the real world. The pseudo-light placement unit places pseudo-lights, which are light sources that illuminate a range equivalent to the field of view of the camera, at candidate camera placement locations in the 3D virtual space. The image acquisition unit places a pseudo-camera, which is a camera with the accuracy expected in the real world, at the location where the pseudo-lights are placed, and acquires an image of the target to be captured in the 3D virtual space with the pseudo-camera while the pseudo-lights are illuminating. The placement optimization unit optimizes the placement of the pseudo-lights based on the acquired image and identifies the optimized pseudo-light placement as the camera placement.

Description

Technical Field

[0001] The present disclosure relates to a camera placement optimization system, a camera placement optimization method, and a camera placement optimization program for optimizing the placement of cameras that image an imaging target.

Background Art

[0002] It is desired to more appropriately place cameras for imaging a desired target, such as surveillance cameras and security cameras.

[0003] For example, Patent Document 1 describes an apparatus for efficiently searching for optimal placement conditions for placement targets. The apparatus described in Patent Document 1 repeats an update process for updating the placement conditions of each of a plurality of placement targets. At that time, a contribution degree is calculated for each of the plurality of placement targets, and the update of the placement conditions for the placement target with the minimum contribution degree is changed more greatly than the update of the placement conditions for the placement targets other than the placement target with the minimum contribution degree.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] On the other hand, cameras are generally used for defect inspection of products. In order to perform defect inspection, it is preferable that cameras can be arranged so as to comprehensively image the products to be imaged.

[0006] Therefore, an object of the present disclosure is to provide a camera placement optimization system, a camera placement optimization method, and a camera placement optimization program that can optimize the placement of cameras that image an imaging target.

Means for Solving the Problems

[0007] The camera placement optimization system disclosed herein is characterized by comprising: a virtual space construction unit that constructs a 3D virtual space that mimics the behavior of the real world; a pseudo-light placement unit that places pseudo-lights, which are light sources that illuminate a range equivalent to the field of view of a camera, at candidate camera placement positions in the 3D virtual space; an image acquisition unit that places a pseudo-camera, which is a camera with the accuracy expected in the real world, at the position where the pseudo-lights are placed, and acquires an image of the target to be captured in the 3D virtual space with the pseudo-camera while the pseudo-lights are illuminating it; and a placement optimization unit that optimizes the placement of the pseudo-lights based on the acquired image and identifies the optimized pseudo-light placement as the camera placement.

[0008] The camera placement optimization method according to this disclosure is characterized by constructing a 3D virtual space that mimics the behavior of the real world, placing a pseudo-light, which is a light source that illuminates a range equivalent to the field of view of the camera, at the candidate camera placement location in the 3D virtual space, placing a pseudo-camera, which is a camera with the accuracy expected in the real world, at the location where the pseudo-light is placed, acquiring an image of the target to be captured in the 3D virtual space with the pseudo-camera while the pseudo-light is illuminating, optimizing the placement of the pseudo-light based on the acquired image, and identifying the optimized pseudo-light placement as the camera placement.

[0009] The camera placement optimization program disclosed herein is characterized by causing a computer to perform the following: a virtual space construction process to construct a 3D virtual space that mimics the behavior of the real world; a pseudo-light placement process in which pseudo-lights, which are light sources that illuminate a range equivalent to the field of view of a camera, are placed at candidate camera placement locations in the 3D virtual space; an image acquisition process in which a pseudo-camera, which is a camera with the accuracy expected in the real world, is placed at the location where the pseudo-lights are placed, and an image of the target to be captured in the 3D virtual space is acquired by the pseudo-camera while the pseudo-lights are illuminating the target; and a placement optimization process in which the placement of the pseudo-lights is optimized based on the acquired image, and the optimized pseudo-light placement is identified as the camera placement. [Effects of the Invention]

[0010] According to this disclosure, the arrangement of cameras that image the target can be optimized. [Brief explanation of the drawing]

[0011] [Figure 1] This is a block diagram showing an example configuration of a camera placement optimization system. [Figure 2] This is an explanatory diagram showing an example of illumination using a simulated light. [Figure 3] This is an explanatory diagram showing an example of illumination using multiple pseudo-lights. [Figure 4] This flowchart shows an example of how the camera placement optimization system works. [Figure 5] This is a block diagram showing an example configuration of an obstacle placement optimization system. [Figure 6] This is an explanatory diagram illustrating an example of the lighting conditions when an obstruction is placed. [Figure 7] This flowchart shows an example of how the obstacle placement optimization system works. [Figure 8] This is a block diagram showing a modified example of the occluder placement optimization system. [Figure 9] This is a block diagram showing an example configuration of a verification system. [Figure 10] This is an explanatory diagram showing an example of how to arrange the objects to be recognized. [Figure 11] This is an explanatory diagram showing an example of image acquisition using texture maps. [Figure 12] This flowchart shows an example of how the verification system works. [Figure 13] This is a block diagram showing an example configuration of a camera placement optimization system. [Figure 14] This is an explanatory diagram showing an example of image acquisition. [Figure 15] This flowchart shows an example of how the camera placement optimization system works. [Figure 16] Block diagram showing an overview of the camera placement optimization system disclosed herein. [Figure 17] It is a schematic block diagram showing the configuration of a computer according to at least one embodiment.

Embodiments of the Invention

[0012] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[0013] Embodiment 1. FIG. 1 is a block diagram showing a configuration example of a camera arrangement optimization system of the present disclosure. The camera arrangement optimization system 100 of the present embodiment includes a storage unit 10, a virtual space construction unit 20, a pseudo light arrangement unit 30, an imaging range specifying unit 40, an arrangement optimization unit 50, and an output unit 60.

[0014] The storage unit 10 stores various information used by the camera arrangement optimization system 100 for processing. Specifically, the storage unit 10 stores various information used when the virtual space construction unit 20 described later constructs a three-dimensional virtual space (hereinafter simply referred to as a three-dimensional virtual space) that simulates the behavior of the real world. The storage unit 10 is realized by, for example, a magnetic disk or the like.

[0015] In addition, in the present embodiment, the storage unit 10 stores information indicating the cameras to be arranged, as well as candidates for the arrangement positions (arrangement ranges), viewing angles, and changeable postures (angles) of the cameras (hereinafter, these pieces of information may be collectively referred to as camera information). That is, the camera information can be said to be information representing a camera arrangement pattern indicating an imaging range. The camera information is defined in advance by a user or the like and is stored in the storage unit 10 in the form of parameters or the like.

[0016] Note that the information indicating the arrangement position of the camera may be information indicating a position in the real world or information indicating a position in the three-dimensional virtual space described later. When information indicating a position in the real world is stored in the storage unit 10, the position may be converted into a position in the three-dimensional virtual space and used.

[0017] The virtual space construction unit 20 constructs a 3D virtual space that mimics the behavior of the real world. An example of a 3D virtual space is the 3D (3-Dimensional) city model used in the Ministry of Land, Infrastructure, Transport and Tourism's urban digital twin realization project, PLATEAU, which is generated from existing data.

[0018] However, the method by which the virtual space construction unit 20 constructs the 3D virtual space is not particularly limited, nor is the content of the 3D virtual space limited to the 3D city model described above. Since various other construction methods and model contents are widely known, detailed explanations are omitted here.

[0019] The pseudo-light placement unit 30 places a light source that illuminates an area equivalent to the camera's field of view (i.e., a light source that illuminates the area the camera can capture) at the candidate camera placement location in the 3D virtual space. Hereinafter, such a light source will be referred to as a pseudo-light. In this embodiment, it is assumed that multiple pseudo-lights are placed in the 3D virtual space.

[0020] Since an object can be imaged when the light reflected from the object reaches the camera, the inventors of this disclosure conceived the idea that the range that can be monitored by a camera is the range that light emitted from the position where the camera is placed reaches. Therefore, by installing a pseudo-light placement unit 30 at the candidate camera placement position and illuminating it with the pseudo-light, it becomes possible to precisely determine the imaging range.

[0021] The pseudo-light placement unit 30 places pseudo-lights in a three-dimensional virtual space. In this embodiment, the pseudo-light placement unit 30 places pseudo-lights in the three-dimensional virtual space for each of the assumed placement patterns. Note that the placement patterns here include not only patterns of candidate camera placement positions, but also patterns in which the orientation of the camera (pseudo-light) is changed at the same placement position. This is because even at the same placement position, if the orientation of the camera (pseudo-light) changes, the imaging range will also change.

[0022] Furthermore, the arrangement patterns discussed here also include patterns in which multiple cameras are placed simultaneously. This is because imaging with multiple cameras allows for comprehensive imaging of the target area. In addition, the performance of the cameras to be arranged may be the same or different.

[0023] Furthermore, the pseudo-light arrangement section 30 may exclude patterns from the expected arrangement patterns in which it is clear that the desired object will not be imaged. By excluding such patterns, the cost of the imaging range identification and optimization processes described later can be reduced.

[0024] The imaging range identification unit 40 illuminates the area illuminated by the positioned pseudo-lights and identifies the area to be imaged. The imaging range identification unit 40 then generates information representing the identified area and the area to be imaged (hereinafter referred to as range identification information). In this embodiment, since it is assumed that multiple pseudo-lights are positioned, the imaging range identification unit 40 illuminates the area illuminated by the combined light from the multiple pseudo-lights and generates range identification information representing the area illuminated by the combined light and the area to be imaged.

[0025] In other words, the area illuminated by the light identified here can be considered as the camera's imaging range, and the object being imaged that is illuminated by the light can be considered as the object being imaged by the camera. The content of the range identified by the imaging range identification unit 40 is arbitrary. For example, the imaging range identification unit 40 may identify the area that the illuminated light reaches, or it may identify the number of objects being imaged that are included in the area that the illuminated light reaches. In addition, the imaging range identification unit 40 may also identify whether or not each object is illuminated by light (i.e., whether or not it is being imaged). Examples of objects being imaged include people and vehicles.

[0026] Furthermore, if there are objects or areas that should be avoided from imaging (hereinafter referred to as "areas outside of imaging"), the imaging range identification unit 40 may also identify whether or not imaging is performed on those areas outside of imaging. The imaging range identification unit 40 stores information representing the range and the objects to be imaged (i.e., range identification information) in the storage unit 10.

[0027] Figure 2 is an explanatory diagram illustrating an example of illumination using a pseudo-light. Note that, for the sake of simplicity, Figure 2 shows the illuminated light in two dimensions; however, in reality, the light is emitted in three dimensions. Specifically, Figure 2 shows an example where a light source (i.e., pseudo-light C1) that achieves a reach and spread equivalent to the camera's imaging range is placed in a three-dimensional virtual space, and the light is emitted downwards in the figure.

[0028] The light emitted from the pseudo-light C1 illuminates an area equivalent to the camera's imaging range. In this case, persons 102 located within the illuminated area 101 are identified as being imaged by the camera, while persons 103 not located within area 101 are identified as not being imaged by the camera.

[0029] Furthermore, if an obstruction such as building 104, as illustrated in Figure 2, is present, the light will be blocked by the obstruction, and light will not reach the area behind building 104. This is equivalent to the camera's imaging being blocked by building 104. In this case, person 105, who is located behind building 104, is identified as not being captured by the camera.

[0030] As described above, the pseudo-light placement processing by the pseudo-light placement unit 30 and the imaging range identification processing by the imaging range identification unit 40 are repeatedly executed for all conceivable placement patterns. Therefore, the storage unit 10 stores information representing the area illuminated by light and the target of imaging (i.e., range identification information) for each conceivable camera placement pattern.

[0031] The placement optimization unit 50 optimizes the camera placement based on the information identified by the imaging range identification unit 40 (i.e., range identification information). Specifically, the placement optimization unit 50 optimizes the placement of the pseudolights based on the range identification information and identifies the optimized pseudolight placement as the camera placement.

[0032] From an optimization perspective, factors such as the shooting range, the degree to which the target is included in the shooting range, the degree to which non-target objects are excluded from the shooting range, and the number of cameras installed (number of simulated lights) can be considered. In other words, a camera arrangement that has a wider shooting range, includes more target objects, excludes as little non-target objects as possible, and requires fewer cameras (number of simulated lights) can be considered desirable.

[0033] Therefore, the arrangement optimization unit 50 may optimize the arrangement of the pseudolights so that more imaging targets are illuminated. The arrangement optimization unit 50 may also optimize the arrangement of the pseudolights so that less light is illuminated to targets or areas that should be avoided from imaging (i.e., areas outside the imaging targets). Furthermore, the arrangement optimization unit 50 may optimize the arrangement of the pseudolights so that fewer pseudolights are used.

[0034] Such optimization problems can be described as combinatorial optimization problems that satisfy various perspectives. The placement optimization unit 50 may optimize the camera placement by defining a combinatorial optimization problem based on the above perspectives.

[0035] For example, constraints are defined for the placement of pseudolights that must be satisfied (e.g., all targets are captured, areas outside the target are not captured, the number of cameras is kept below a predetermined number). Furthermore, an objective function is defined that includes the degree of desirability resulting from the placement of pseudolights as a feature. Examples of these features include the number of targets captured, the degree to which areas outside the target are suppressed from being captured, the number of cameras, and the imaging range covered.

[0036] The placement optimization unit 50 may optimize the placement of the pseudolights to maximize the objective function under the constraints defined in this manner. Note that the viewpoints and features listed here are just examples, and other viewpoints and features may also be included.

[0037] Figure 3 is an explanatory diagram illustrating an example of illumination using multiple pseudolights. The example shown in Figure 3 shows how much of the target (or area outside the target) is imaged when light is emitted from two pseudolights C2 and C3, and is a type of range identification information. Compared to the pseudolight arrangement in pattern 111, the pseudolight arrangement in pattern 112 images two areas outside the monitoring target, and fewer monitored areas are imaged. Therefore, for example, the value of the objective function shown above is expected to be larger for pattern 111 than for pattern 112. Also, if the constraint is defined as not to image areas outside the monitoring target, the pseudolight arrangement in pattern 112 can be said to be a constraint violation.

[0038] The output unit 60 outputs the optimized camera arrangement. The output unit 60 may also output information such as the placement position and orientation of each of the multiple cameras.

[0039] The virtual space construction unit 20, the pseudo-light placement unit 30, the imaging range identification unit 40, the placement optimization unit 50, and the output unit 60 are implemented by a computer processor (e.g., a CPU (Central Processing Unit) or GPU (Graphics Processing Unit)) that operates according to a program (camera placement optimization program). For example, the program may be stored in the memory unit 10 of the camera placement optimization program 100, and the processor may read the program and operate as the virtual space construction unit 20, the pseudo-light placement unit 30, the imaging range identification unit 40, the placement optimization unit 50, and the output unit 60 according to the program.

[0040] Furthermore, each function of the camera placement optimization program 100 may be provided in SaaS (Software as a Service) format. Also, the virtual space construction unit 20, the pseudo-light placement unit 30, the imaging range identification unit 40, the placement optimization unit 50, and the output unit 60 may each be implemented with dedicated hardware.

[0041] Furthermore, some or all of the components of each device may be implemented by general-purpose or dedicated circuits, processors, etc., or combinations thereof. These may be comprised of a single chip or multiple chips connected via a bus. Some or all of the components of each device may be implemented by a combination of the aforementioned circuits, etc., and programs.

[0042] Furthermore, if some or all of the components of the camera placement optimization program 100 are implemented by multiple information processing devices or circuits, these multiple information processing devices or circuits may be centrally located or distributed. For example, the information processing devices or circuits may be implemented in a form in which each is connected via a communication network, such as a client-server system or a cloud computing system.

[0043] Next, an example of the operation of the camera placement optimization system 100 of this embodiment will be described. Figure 4 is a flowchart of an example of the operation of the camera placement optimization system 100 of this embodiment.

[0044] The virtual space construction unit 20 constructs a three-dimensional virtual space (step S11). The pseudo-light placement unit 30 places pseudo-lights at candidate camera placement locations in the three-dimensional virtual space (step S12). The imaging range identification unit 40 emits light from the placed pseudo-lights and generates range identification information representing the range illuminated by the light and the target to be imaged (step S13). Then, the placement optimization unit 50 optimizes the placement of the pseudo-lights based on the range identification information and identifies the optimized pseudo-light placement as the camera placement (step S14).

[0045] As described above, according to this embodiment, the virtual space construction unit 20 constructs a three-dimensional virtual space, and the pseudo-light placement unit 30 places pseudo-lights at candidate camera placement locations. The imaging range identification unit 40 then emits light from the placed pseudo-lights and generates range identification information representing the range illuminated by the light and the target to be imaged. The placement optimization unit 50 optimizes the placement of the pseudo-lights based on the range identification information and identifies the optimized pseudo-light placement as the camera placement. Therefore, the camera placement can be optimized considering the actual environment.

[0046] For example, in complex spaces, it is difficult to determine whether a surveillance camera can monitor the target area, or whether it is avoiding capturing images of areas that should not be included. Furthermore, when installing multiple cameras, determining the optimal placement and number of cameras can be costly.

[0047] However, in this embodiment, since the imaging range can be determined by irradiating light in a three-dimensional virtual space, it becomes easier to decide whether or not to place a camera.

[0048] Furthermore, the camera placement used when verifying pedestrian and traffic flow can also be optimized using the camera placement optimization system 100 according to this embodiment. Cameras are also installed when measuring pedestrian and traffic flow, but the measurement results can vary greatly depending on the placement and orientation of these cameras. As shown in this embodiment, by constructing the environment to be measured in a three-dimensional virtual space and illuminating it with light from the camera's position, it becomes possible to determine whether or not necessary information (faces, license plates, etc.) can be acquired.

[0049] Embodiment 2. Next, a second embodiment of the present disclosure will be described. In the second embodiment, a method for arranging occlusion (screen) taking into account the actual environment will be described so that objects or areas that should be avoided from being imaged (i.e., areas outside the image target) are not imaged. By arranging such occlusion, it is possible to suppress the image capture of objects that are not to be imaged.

[0050] Figure 5 is a block diagram showing an example configuration of the occluder placement optimization system of the present disclosure. The occluder placement optimization system 200 of this embodiment includes a storage unit 110, a virtual space construction unit 20, a pseudo-light placement unit 32, an occluder placement unit 130, an imaging range identification unit 40, a placement optimization unit 150, and an output unit 160.

[0051] In other words, the occluder placement optimization system 200 of this embodiment differs from the camera placement optimization system 100 of the first embodiment in that, instead of the memory unit 10, pseudo-light placement unit 30, placement optimization unit 50, and output unit 60, it comprises a memory unit 110, pseudo-light placement unit 32, placement optimization unit 150, and output unit 160, and further comprises an occluder placement unit 130. The other configurations are the same as those of the camera placement optimization system 100.

[0052] First, we will explain a configuration in which multiple patterns of obstacles are placed around a single predetermined pseudo-light. Configurations in which multiple patterns of obstacles are placed around multiple pseudo-lights will be described later.

[0053] The storage unit 110 stores various types of information used by the occluder placement optimization system 200 for processing. Specifically, in addition to the information stored by the storage unit 10 in the first embodiment, the storage unit 110 stores information indicating the characteristics of the occluders to be placed and candidate placement locations (placement ranges) of those occluders (hereinafter, this information may be collectively referred to as occluder information). The characteristics of the occluders are represented, for example, by their size, shape, the transmittance of their materials, etc. In other words, occluder information can be said to be information that represents the placement pattern of occluders. The occluder information is defined in advance by the user, etc., and stored in the storage unit 110 in the form of parameters, etc.

[0054] The information indicating the placement of the obstacle may be information indicating its location in the real world, or information indicating its location in a 3D virtual space. If information indicating its location in the real world is stored in the memory unit 110, that location can be converted to a location in the 3D virtual space and used.

[0055] Specifically, a three-dimensional virtual space with pseudolights is constructed, similar to the method shown in the first embodiment. Once the placement of the pseudolights in the three-dimensional virtual space is determined, the placement of occluders is then determined relative to the placement of the pseudolights. That is, first, the virtual space construction unit 20 constructs the three-dimensional virtual space. Then, the pseudolight placement unit 32 places pseudolights at the camera placement positions. The method by which the pseudolight placement unit 32 places pseudolights is the same as the method by which the pseudolight placement unit 30 in the first embodiment places pseudolights. Note that while the pseudolight placement unit 30 in the first embodiment placed pseudolights in the three-dimensional virtual space for all conceivable placement patterns, the pseudolight placement unit 32 in this embodiment places one predetermined pseudolight in the three-dimensional virtual space.

[0056] The shield placement unit 130 places shields in a three-dimensional virtual space to block the light emitted by the pseudo-light. In this embodiment, the shield placement unit 130 places shields in the three-dimensional virtual space for each of the possible placement patterns. The placement patterns here include not only patterns of candidate placement positions for the shields, but also patterns in which the orientation of the shields is changed at the same placement position. This is because even at the same placement position, if the orientation of the shields changes, the area that is shielded will also change.

[0057] Furthermore, the placement patterns discussed here also include patterns where multiple occluders are placed simultaneously in a 3D virtual space. By using multiple occluders simultaneously for occlusion, it is possible to verify placements that take multiple occluders into account.

[0058] Similar to the pseudo-light, the occluding object placement section 130 may exclude patterns from the expected placement patterns that clearly do not suppress imaging of objects other than the target of imaging. Excluding such patterns can reduce the cost of the occluding object identification and optimization processes described later.

[0059] Figure 6 is an explanatory diagram illustrating an example of the lighting situation when an obstruction is placed. The example shown in Figure 6 shows to what extent the imaging target (or area outside the imaging target) is imaged when light is shone from a pseudo-light source with an obstruction in place, and is a type of range identification information. When the obstruction S1 is placed as shown in pattern 121, three areas outside the monitoring target are obstructed, while the four monitoring targets in area 131 are also obstructed. On the other hand, when the obstruction S2 is placed as shown in pattern 122, only the three areas outside the monitoring target can be obstructed without obstructing the monitoring targets. Therefore, pattern 122 is a more preferable obstruction method than pattern 121.

[0060] The imaging range identification unit 40 illuminates the area illuminated by the light from a simulated light source with an obstruction in place, and identifies the area to be imaged and the target area. The imaging range identification unit 40 then generates information representing the identified area and target area (i.e., range identification information). The method by which the imaging range identification unit 40 generates the range identification information is the same as the method shown in the first embodiment.

[0061] The imaging range identification unit 40 then stores the generated range identification information in the storage unit 110. In other words, the storage unit 110 stores range identification information for each pattern of placed occluding objects.

[0062] The placement optimization unit 150 optimizes the placement of occluders based on the range identification information. From the perspective of optimization, the degree to which areas outside the target of imaging are not included in the imaging range, the degree to which the target of imaging is included in the imaging range, and the number of occluders to be installed can be considered. In other words, an occluder placement that includes as few areas outside the target of imaging as possible, includes as many targets of imaging as possible, and has fewer occluders to be installed can be considered a desirable occluder placement.

[0063] Therefore, the placement optimization unit 150 may optimize the placement of shims so that less light is irradiated onto objects or areas that should be avoided from imaging (i.e., areas outside the imaging target). The placement optimization unit 150 may also optimize the placement of shims so that less light is obstructed from irradiating onto the imaging target. Furthermore, the placement optimization unit 150 may optimize the placement of shims so that fewer shims are placed.

[0064] Such optimization problems, like those in the first embodiment, can be described as combinatorial optimization problems that satisfy various perspectives. The placement optimization unit 150 may optimize the camera placement by defining a combinatorial optimization problem based on the above perspectives.

[0065] For example, constraints are defined for the placement of occluders that must always be satisfied (e.g., do not image areas outside the target area, do not obstruct the target area, keep the number of occluders below a predetermined number, etc.). Furthermore, an objective function is defined that includes the degree of desirability caused by the placement of occluders as a feature. Examples of these features include the degree to which imaging of areas outside the target area is suppressed, the number of targets to be imaged, the number of occluders, and the range of imaging that is obstructed.

[0066] The placement optimization unit 150 may optimize the placement of occluders to maximize the objective function under the constraints defined in this manner. Note that the viewpoints and features listed here are just examples, and other viewpoints and features may also be included.

[0067] The output unit 160 outputs the optimized arrangement of occluders. The output unit 160 may also output information such as the placement position and orientation of each of the multiple occluders.

[0068] The virtual space construction unit 20, the pseudo-light placement unit 32, the occluder placement unit 130, the imaging range identification unit 40, the placement optimization unit 150, and the output unit 160 are all implemented by a computer processor that operates according to a program (occluder placement optimization program).

[0069] Next, an example of the operation of the occluder placement optimization system 200 of this embodiment will be described. Figure 7 is a flowchart showing an example of the operation of the occluder placement optimization system 200 of this embodiment.

[0070] The virtual space construction unit 20 constructs a three-dimensional virtual space (step S21). The pseudo-light placement unit 32 places pseudo-lights at the camera placement locations in the three-dimensional virtual space (step S22). The occluding object placement unit 130 places occluding objects in the three-dimensional virtual space (step S23). The imaging range identification unit 40 emits light from the pseudo-lights with the occluding objects in place and generates range identification information representing the range illuminated by the light and the object to be imaged (step S24). Then, the placement optimization unit 150 optimizes the placement of the occluding objects based on the range identification information (step S25).

[0071] As described above, according to this embodiment, the virtual space construction unit 20 constructs a three-dimensional virtual space, and the pseudo-light placement unit 32 places pseudo-lights at the camera placement locations. The occluding object placement unit 130 places occluding objects in the three-dimensional virtual space, and the imaging range identification unit 40 emits light from the pseudo-lights with the occluding objects in place, generating range identification information that represents the area illuminated by the light and the object to be imaged. The placement optimization unit 150 then optimizes the placement of occluding objects based on the range identification information. Therefore, the placement of occluding objects to block imaging by the camera can be optimized considering the actual environment.

[0072] Next, a modified example of the occluder placement optimization system 200 of this embodiment will be described. In this modified example, a configuration in which occluders are placed in multiple patterns for multiple pseudolights will be described.

[0073] Figure 8 is a block diagram showing a modified example of the occluder placement optimization system. The occluder placement optimization system 210 of this modified example includes a storage unit 112, a virtual space construction unit 20, a pseudo-light placement unit 30, an occluder placement unit 130, an imaging range identification unit 40, a placement optimization unit 152, and an output unit 160.

[0074] The storage unit 112 stores both the information stored by the storage unit 10 in the first embodiment and the information stored by the storage unit 110 in the second embodiment. In other words, the storage unit 112 stores both camera information and obstruction information.

[0075] The contents of the virtual space construction unit 20, the pseudo-light placement unit 30, and the imaging range identification unit 40 are the same as in the first embodiment. Also, the contents of the occluding object placement unit 130 are the same as in the second embodiment. That is, in this embodiment, the pseudo-light placement unit 30 places pseudo-lights in the 3D virtual space for all possible placement patterns, and the occluding object placement unit 130 places occluding objects in the 3D virtual space for all possible placement patterns. For example, suppose there are 100 candidate pseudo-camera placements and 10 candidate occluding object placements. In this case, all 10 types of occluding objects are placed for each of the 100 pseudo-camera placements.

[0076] The imaging range identification unit 40 then, with the occluders arranged according to the assumed patterns, irradiates light according to the assumed arrangement pattern of pseudo-lights and identifies the range illuminated by the light and the target to be imaged. The imaging range identification unit 40 then generates information representing the identified range and the target to be imaged (i.e., range identification information). The imaging range identification unit 40 then stores the generated range identification information in the storage unit 112. In other words, the storage unit 112 stores range identification information for each arrangement pattern of pseudo-lights and each pattern of occluders.

[0077] The placement optimization unit 152 optimizes the placement of pseudolights and occluders based on the range identification information. Specifically, the placement optimization unit 152 optimizes the placement of pseudolights and occluders based on the range identification information and identifies the optimized pseudolight placement as the camera placement.

[0078] The placement optimization unit 152, as in the first and second embodiments, can optimize the placement of pseudolights (cameras) and occluders by defining a combinatorial optimization problem that satisfies various viewpoints.

[0079] Furthermore, in this modified example, the optimization conditions of the first embodiment may be modified by not setting any conditions for areas not to be imaged. In this case, it is conceivable that areas not to be imaged may be captured even with optimization of the camera placement alone, but such cases can be appropriately handled by using occlusions.

[0080] Embodiment 3. Next, a third embodiment of the present disclosure will be described. In the third embodiment, a method for verifying the accuracy of object recognition by a positioned camera will be described. Figure 9 is a block diagram showing an example configuration of the verification system of the present disclosure. The verification system 300 of this embodiment includes a storage unit 120, a virtual space construction unit 20, a pseudo-light placement unit 32, a recognition target placement unit 250, an image acquisition unit 260, an accuracy verification unit 270, and an output unit 280.

[0081] The contents of the virtual space construction unit 20 and the pseudo-light placement unit 32 are the same as in the second embodiment. The verification system 300 may also include the functions of the placement optimization unit 50 in the first embodiment and the functions of the obstacle placement unit 130 in the second embodiment.

[0082] Similar to the first embodiment, the memory unit 120 stores various information used by the virtual space construction unit 20 when constructing a three-dimensional virtual space. Furthermore, the memory unit 120 stores information indicating the camera's placement position, field of view, and orientation. The memory unit 120 may also store various information for controlling the camera. The memory unit 120 is implemented, for example, by a magnetic disk or the like.

[0083] In this embodiment, similar to the second embodiment, pseudo-lights are pre-placed in a three-dimensional virtual space, and light is emitted from these placed pseudo-lights to generate range identification information. Specifically, first, the virtual space construction unit 20 constructs a three-dimensional virtual space, and the pseudo-light placement unit 32 places pseudo-lights at the camera placement positions.

[0084] The recognition target placement unit 250 places the object whose recognition accuracy is to be verified (hereinafter sometimes simply referred to as the recognition target) in a three-dimensional virtual space. The position in which the recognition target placement unit 250 places the recognition target is arbitrary and may be within an area illuminated by light or within an area not illuminated by light. The recognition target placement unit 250 specifies, for example, the information of the recognition target to be placed in the three-dimensional virtual space and the placement position of the recognition target. Examples of the recognition target information include 3D information indicating the form of the recognition target, or simply an object displaying text.

[0085] Furthermore, if a recognition target already exists in the 3D virtual space, the recognition target placement unit 250 may specify information that identifies the recognition target and the location where the recognition target exists. In other words, in this embodiment, the placement of a recognition target also includes specifying a recognition target that already exists in the 3D virtual space.

[0086] Figure 10 is an explanatory diagram showing an example in which the recognition targets are placed within the range that can be captured by the camera. In Figure 10, the range 141 that can be captured by the camera C4 is identified, and the recognition targets 142 and 143 are placed within that range 141.

[0087] The image acquisition unit 260 places a camera with the accuracy expected in the real world (hereinafter referred to as the pseudo-camera) at the location where the pseudo-light is positioned. Then, while the pseudo-light is illuminating, the image acquisition unit 260 acquires an image of the 3D virtual space obtained by the pseudo-camera (more specifically, an image of the object to be recognized in the 3D virtual space; also called a rendering image). As mentioned above, the range illuminated by the pseudo-light is equivalent to the range that the pseudo-camera can capture.

[0088] Furthermore, if there are multiple pseudo-lights that illuminate the object to be recognized, the image acquisition unit 260 may place a pseudo-camera at each of the locations where the multiple pseudo-lights are positioned. The image acquisition unit 260 may then acquire an image of the object to be recognized using each of the placed pseudo-cameras.

[0089] Furthermore, in order to clearly show the area of ​​the object being imaged that has been illuminated by light, the image acquisition unit 260 may acquire an image of the object being imaged that has been illuminated by the pseudo-light while the three-dimensional virtual space is dimmed.

[0090] Furthermore, a texture map may be used to determine the state of light illumination on a three-dimensional object to be imaged. Specifically, the image acquisition unit 260 may place a texture map representing the object to be imaged in a three-dimensional virtual space, unfold the surface of the texture map illuminated by light, and acquire an image of the surface of the texture map illuminated by light.

[0091] Figure 11 is an explanatory diagram illustrating an example of image acquisition using a texture map. First, a texture map is generated in advance using an arbitrary tool 161, and the image acquisition unit 260 places the generated texture map in a three-dimensional virtual space 162. The image acquisition unit 260 unfolds the surface of the texture map illuminated with light and acquires an image 163 of the surface of the illuminated texture map.

[0092] In the example shown in Figure 11, it is assumed that light is shone onto the generated texture map from a placed pseudo-light 164. In this case, the position of the light shone on the texture map can be confirmed in the 3D virtual space 162. Furthermore, by unfolding the surface of the illuminated texture map, the position of the light shone on each part can be confirmed in more detail. For example, in the example shown in Figure 11, it can be confirmed that the front of the body can be photographed because light is shone on the front of the body, and that the back of the body cannot be photographed because light is not shone on the back of the body.

[0093] The accuracy verification unit 270 verifies the recognition accuracy of the recognition target captured in the acquired image. Specifically, the accuracy verification unit 270 determines whether or not light is illuminating the recognition target captured in the acquired image. If light is not illuminating the recognition target, the recognition target will not be captured by the camera in the first place, making it possible to determine that the recognition target cannot be recognized or that the camera's position is inappropriate without verifying the accuracy of the recognition target. On the other hand, if light is illuminating the recognition target, the accuracy verification unit 270 verifies whether the recognition accuracy of the recognition function (hereinafter sometimes referred to as the recognition function) used on the acquired image of the recognition target meets the predetermined recognition criteria.

[0094] Examples of recognition functions include face recognition applications, object recognition applications, and character recognition applications. These applications may be dedicated applications or general-purpose applications (tools). For example, if a face recognition application is used as the recognition function, the accuracy verification unit 270 may verify whether the face recognition of the acquired image meets predetermined recognition criteria.

[0095] However, the recognition function is not limited to these applications; any function capable of determining whether or not recognition is possible is acceptable, and the content and accuracy of that function are arbitrary. Furthermore, multiple recognition functions may be used for a single image.

[0096] Furthermore, the system may allow the user to verify whether or not they can recognize the image. In this case, the accuracy verification unit 270 may display the image captured by the pseudo-camera on a display device (not shown) via the output unit 280 (described later) to receive the user's verification result regarding recognition feasibility. The accuracy verification unit 270 may then determine the final recognition result based on the received verification result and recognition accuracy, according to predetermined rules.

[0097] The output unit 280 outputs the recognition result of the acquired image. As described above, the output unit 280 may also output the image captured by the camera itself.

[0098] Furthermore, suppose that in a situation where a camera corresponding to the simulated camera is actually deployed in the real world, it is determined that the recognition accuracy of the image acquired by that simulated camera does not meet the predetermined recognition criteria. In this case, the output unit 280 may output a control signal to that camera to control the camera's imaging method (e.g., field of view).

[0099] The virtual space construction unit 20, the pseudo-light placement unit 32, the recognition target placement unit 250, the image acquisition unit 260, the accuracy verification unit 270, and the output unit 280 are all implemented by a computer processor that operates according to a program (verification program).

[0100] Next, an example of the operation of the verification system 300 of this embodiment will be described. Figure 12 is a flowchart showing an example of the operation of the verification system 300 of this embodiment.

[0101] The virtual space construction unit 20 constructs a 3D virtual space (step S31). The pseudo-light placement unit 32 places pseudo-lights at the camera placement locations in the 3D virtual space (step S32). The recognition target placement unit 250 places the recognition target in the 3D virtual space (step S33).

[0102] The image acquisition unit 260 places a pseudo-camera at the location where the pseudo-light is positioned (step S34). The image acquisition unit 260 acquires an image of the recognition target in the 3D virtual space using the pseudo-camera while the pseudo-light is illuminating it (step S35). The accuracy verification unit 270 determines whether or not the recognition target captured in the acquired image is illuminated by light (step S36). If the recognition target is illuminated by light (Yes in step S36), the accuracy verification unit 270 verifies whether the recognition accuracy meets the recognition criteria when the recognition function is used on the acquired image (step S37). On the other hand, if the recognition target is not illuminated by light (No in step S36), the accuracy verification unit 270 determines that the recognition criteria are not met without performing verification (step S38).

[0103] As described above, according to this embodiment, the virtual space construction unit 20 constructs a three-dimensional virtual space, and the pseudo-light placement unit 32 places pseudo-lights at the camera placement positions. The recognition target placement unit 250 places the recognition target in the three-dimensional virtual space, and the image acquisition unit 260 places a pseudo-camera at the position where the pseudo-lights are placed, and acquires an image of the recognition target while the pseudo-lights are illuminating it. Then, the accuracy verification unit 270, if it determines that the recognition target captured in the acquired image is illuminated by light, verifies whether the recognition accuracy meets the recognition criteria when the recognition function is used on the acquired image. Thus, it is possible to verify whether the target is captured with the desired recognition accuracy.

[0104] Embodiment 4. Next, a fourth embodiment of the present disclosure will be described. In the fourth embodiment, a method for optimizing the arrangement of cameras that image an object to be imaged will be described. Figure 13 is a block diagram showing an example configuration of the camera arrangement optimization system of the present disclosure. The camera arrangement optimization system 400 of this embodiment includes a storage unit 10, a virtual space construction unit 20, a pseudo-light arrangement unit 30, an image acquisition unit 360, an arrangement optimization unit 370, and an output unit 380.

[0105] The contents of the memory unit 10, the virtual space construction unit 20, and the pseudo-light placement unit 30 are the same as in the first embodiment. That is, the memory unit 10 stores various information used when constructing the 3D virtual space, as well as camera information. The virtual space construction unit 20 constructs the 3D virtual space, and the pseudo-light placement unit 30 places pseudo-lights at candidate camera placement locations in the 3D virtual space.

[0106] The image acquisition unit 360 places a pseudo-camera at the location where the pseudo-lights are positioned, and acquires an image of the target to be captured in the 3D virtual space using the pseudo-camera while the pseudo-lights are emitting light. In this embodiment, it is assumed that multiple pseudo-lights are positioned, so the image acquisition unit 360 acquires an image of the target to be captured while the multiple positioned pseudo-lights are emitting light.

[0107] The image acquisition unit 360 performs the image acquisition process for all possible pseudo-camera arrangement patterns. The image acquisition unit 360 then stores each acquired image of the target in the storage unit 10. As a result, the storage unit 10 stores images of the target illuminated by light for each possible camera arrangement pattern. In addition, the image acquisition unit 360 may, similar to the image acquisition unit 260 in the third embodiment, place a texture map representing the target in a three-dimensional virtual space, unfold the surface of the illuminated texture map, and acquire an image of the illuminated surface of the texture map.

[0108] Figure 14 is an explanatory diagram showing an example of image acquisition. In the example shown in Figure 14, a product 151 to be inspected, including parts 1 to 8, is displayed in a three-dimensional virtual space, and the image acquisition unit 360 has captured an image 152 showing the product illuminated by light. The image acquisition unit 360 may acquire the image as is, captured from each direction, or, as shown in Figure 14, may acquire separate images for each object (for example, objects 1 to 8). Acquiring separate images for each object makes it easier to understand the state of light illumination for each object.

[0109] The placement optimization unit 370 optimizes the camera placement based on the image captured by the image acquisition unit 360. Specifically, the placement optimization unit 370 optimizes the placement of pseudolights based on the captured image and identifies the optimized pseudolight placement as the camera placement. At this time, the placement optimization unit 370 may optimize the camera placement using the texture map generated by the image acquisition unit 360.

[0110] From an optimization perspective, factors such as the range of light illuminating the target being imaged and the number of cameras installed (number of simulated lights) are considered desirable. In other words, a camera arrangement where the range of light illuminating the target being imaged is wider and the number of cameras installed (number of simulated lights) is smaller is considered preferable.

[0111] Therefore, the arrangement optimization unit 370 may optimize the arrangement of the pseudolights so that a larger area of ​​the imaging target is illuminated by light. Alternatively, the arrangement optimization unit 370 may optimize the arrangement of the pseudolights so that a smaller number of pseudolights are used.

[0112] Such optimization problems can also be described as combinatorial optimization problems that satisfy various perspectives. The placement optimization unit 370 may optimize the camera placement by defining a combinatorial optimization problem based on the above perspectives.

[0113] For example, constraints are defined regarding the placement of pseudolights that must always be met (e.g., the number of cameras must be less than or equal to a predetermined number). Furthermore, an objective function is defined that includes the degree of desirability resulting from the placement of pseudolights as a feature. Examples of these features include the range of light illuminating the target being imaged and the number of cameras.

[0114] The placement optimization unit 370 may optimize the placement of the pseudolights to maximize the objective function under the constraints defined in this way. For example, the placement optimization unit 370 may optimize the placement of the pseudolights so that the light illuminating the imaging target from the pseudolights covers the widest possible area, under constraints on the number of pseudolights to be placed (e.g., an upper limit). Note that the viewpoints and features listed here are just examples, and other viewpoints and features may also be included.

[0115] The output unit 380 outputs the optimized camera arrangement. The output unit 380 may also output information such as the placement position and orientation of each of the multiple cameras.

[0116] The virtual space construction unit 20, the pseudo-light placement unit 30, the image acquisition unit 360, the placement optimization unit 370, and the output unit 380 are all implemented by a computer processor that operates according to a program (camera placement optimization program).

[0117] Next, an example of the operation of the camera placement optimization system 400 of this embodiment will be described. Figure 15 is a flowchart of an example of the operation of the camera placement optimization system 400 of this embodiment. The virtual space construction unit 20 constructs a three-dimensional virtual space (step S41). The pseudo-light placement unit 30 places pseudo-lights at candidate camera placement positions in the three-dimensional virtual space (step S42). The image acquisition unit 360 places a pseudo-camera at the position where the pseudo-lights are placed, and with the pseudo-lights illuminating, acquires an image of the target to be captured in the three-dimensional virtual space with the pseudo-camera (step S43). Then, the placement optimization unit 370 optimizes the placement of the pseudo-lights based on the acquired image and identifies the optimized pseudo-light placement as the camera placement (step S44).

[0118] As described above, according to this embodiment, the virtual space construction unit 20 constructs a three-dimensional virtual space, and the pseudo-light placement unit 30 places pseudo-lights at candidate camera placement locations. The image acquisition unit 360 places a pseudo-camera at the location where the pseudo-lights are placed, and acquires an image of the target object with the pseudo-camera while the pseudo-lights are illuminating it. Then, the placement optimization unit 370 optimizes the placement of the pseudo-lights based on the acquired image and identifies the optimized pseudo-light placement as the camera placement. Thus, the placement of cameras that capture the target object can be optimized.

[0119] For example, cameras are commonly used for product inspection, but in products such as circuit boards with multiple components, the component to be inspected may be hidden by other components. Furthermore, because each product has a different form factor, determining the optimal camera placement for each product's form factor is costly.

[0120] One way to solve such problems is to use X-rays to inspect products for defects. However, since X-ray inspection is costly, it is preferable to perform defect inspection using so-called visible light.

[0121] On the other hand, in this embodiment, the virtual space construction unit 20 constructs a three-dimensional virtual space on which the product is placed, and the image acquisition unit 360 acquires an image of the product with a pseudo-camera while the pseudo-light is illuminating it. Therefore, the imaging range can be confirmed based on the light illumination conditions, making it easier to consider the placement of the camera.

[0122] Next, an overview of the present disclosure will be described. Figure 16 is a block diagram showing an overview of the camera placement optimization system of the present disclosure. The optimization system 380 (for example, the camera placement optimization system 400) according to the present disclosure comprises: a virtual space construction unit 381 (for example, a virtual space construction unit 20) that constructs a three-dimensional virtual space that mimics the behavior of the real world; a pseudo-light placement unit 382 (for example, a pseudo-light placement unit 30) that places pseudo-lights, which are light sources that illuminate a range equivalent to the field of view of the camera, at candidate camera placement positions in the three-dimensional virtual space; an image acquisition unit 383 (for example, an image acquisition unit 360) that places a pseudo-camera, which is a camera with the accuracy expected in the real world, at the position where the pseudo-lights are placed, and acquires an image of the target to be captured in the three-dimensional virtual space with the pseudo-camera while the pseudo-lights are illuminating the area; and a placement optimization unit 384 (for example, a placement optimization unit 370) that optimizes the placement of the pseudo-lights based on the acquired image and identifies the optimized pseudo-light placement as the camera placement.

[0123] Such a configuration allows for the optimization of the placement of cameras that image the target object.

[0124] Alternatively, the pseudo-light placement unit 382 may place multiple pseudo-lights in a three-dimensional virtual space, and the image acquisition unit 383 may acquire an image of the target being captured while the placed multiple pseudo-lights are illuminating it.

[0125] Alternatively, the image acquisition unit 383 may acquire an image of the object being captured, illuminated by the pseudo-light, while the three-dimensional virtual space is dimmed.

[0126] Alternatively, the image acquisition unit 383 may generate a texture map representing the object to be imaged and place it in a three-dimensional virtual space, unfold the surface of the texture map illuminated with light, and acquire an image of the surface of the texture map illuminated with light.

[0127] Furthermore, the arrangement optimization unit 384 may optimize the arrangement of the pseudolights so that the light illuminating the imaging target from the pseudolights covers the widest possible area, given the constraints on the number of pseudolights to be arranged.

[0128] Furthermore, the placement optimization unit 384 may optimize the placement of pseudolights to maximize an objective function that includes the degree of desirability resulting from the placement of pseudolights as a feature, while using the placement conditions of pseudolights that must always be satisfied as constraints.

[0129] Figure 17 is a schematic block diagram showing the configuration of a computer according to at least one embodiment. Computer 1000 comprises a processor 1001, main memory 1002, auxiliary memory 1003, and interface 1004. Computer 1000 may also be connected to a computer for running a mathematical programming solver, an annealing machine, a simulator, etc.

[0130] The camera placement optimization system 100 described above is implemented in the computer 1000. The operation of each processing unit described above is stored in the auxiliary storage device 1003 in the form of a program (camera placement optimization program). The processor 1001 reads the program from the auxiliary storage device 1003, loads it into the main memory 1002, and executes the above processing according to the program.

[0131] In at least one embodiment, the auxiliary storage device 1003 is an example of a non-temporary tangible medium. Other examples of non-temporary tangible media include magnetic disks, magneto-optical disks, CD-ROMs (Compact Disc Read-only memory), DVD-ROMs (Read-only memory), and semiconductor memory connected via the interface 1004. Furthermore, if this program is distributed to the computer 1000 via a communication line, the computer 1000 that receives the program may expand it into the main memory 1002 and execute the above processing.

[0132] Furthermore, the program may be intended to implement some of the functions described above. In addition, the program may be a so-called differential file (differential program) that implements the functions described above in combination with other programs already stored in the auxiliary storage device 1003.

[0133] Although the present invention has been described above with reference to the embodiments and examples, the present invention is not limited to the above embodiments and examples. Various modifications to the structure and details of the present invention can be made, as can be understood by those skilled in the art within the scope of the present invention. [Explanation of Symbols]

[0134] 10,110,112,120 Storage section 20 Virtual Space Construction Department 30,32 Pseudo-light placement section 40 Imaging range identification unit 50,152,370 Placement Optimization Unit 60, 160, 280, 380 Output section 100,400 Camera Placement Optimization System 130 Shield placement part 150 Layout Optimization Unit 200,210 Obstacle Placement Optimization System 250 Recognition target placement section 260,360 Image Acquisition Unit 270 Accuracy Verification Department 300 Verification Systems

Claims

1. The virtual space construction unit constructs a 3D virtual space that mimics the behavior of the real world, In the aforementioned three-dimensional virtual space, a pseudo-light placement unit is provided that places a pseudo-light, which is a light source that illuminates an area equivalent to the field of view of the camera, at the candidate camera placement position. An image acquisition unit is provided which, at the location where the aforementioned pseudo-light is positioned, a pseudo-camera having the accuracy expected in the real world is positioned, and while the pseudo-light is illuminating the object, the pseudo-camera acquires an image of the object to be captured in the 3D virtual space. The system includes a configuration optimization unit that optimizes the arrangement of the pseudo-lights based on the acquired image and identifies the optimized arrangement of the pseudo-lights as the arrangement of the camera. A camera placement optimization system characterized by the following features.

2. The pseudo-light placement section places multiple pseudo-lights in a three-dimensional virtual space. The image acquisition unit acquires an image of the target object while multiple positioned pseudo-lights are illuminating it. The camera placement optimization system according to claim 1.

3. The image acquisition unit acquires an image of the target being captured, illuminated by a simulated light source, while the three-dimensional virtual space is dimmed. A camera placement optimization system according to claim 1 or claim 2.

4. The image acquisition unit generates a texture map representing the object to be imaged and places it in a three-dimensional virtual space, unfolds the surface of the texture map when light is shone on it, and acquires an image of the surface of the texture map when light is shone on it. A camera placement optimization system according to claim 1 or claim 2.

5. The placement optimization unit optimizes the placement of the pseudolights so that the light illuminating the target being imaged covers the widest possible area, while being constrained by the number of pseudolights to be placed. A camera placement optimization system according to claim 1 or claim 2.

6. The placement optimization unit optimizes the placement of pseudolights to maximize an objective function that includes the degree of desirability resulting from the placement of pseudolights as a feature, using the placement conditions of pseudolights that must always be satisfied as constraints. A camera placement optimization system according to claim 1 or claim 2.

7. By constructing a 3D virtual space that mimics the behavior of the real world, In the aforementioned three-dimensional virtual space, a pseudo-light, which is a light source that illuminates an area equivalent to the field of view of the camera, is placed at the candidate camera placement location. A pseudo-camera, which is a camera with the accuracy expected in the real world, is placed at the location where the pseudo-light is positioned, and while the pseudo-light is emitting light, an image of the object to be captured in the 3D virtual space is acquired with the pseudo-camera. Based on the acquired image, the arrangement of the pseudolights is optimized, and the optimized arrangement of the pseudolights is identified as the arrangement of the camera. A method for optimizing camera placement, characterized by the following features.

8. By placing multiple pseudo-lights in a 3D virtual space, The image of the target object is acquired while multiple positioned pseudo-lights are illuminating it. The camera placement optimization method according to claim 7.

9. On the computer, A virtual space construction process that creates a 3D virtual space that mimics the behavior of the real world. In the aforementioned three-dimensional virtual space, a pseudo-light placement process is performed to place a pseudo-light, which is a light source that illuminates an area equivalent to the field of view of the camera, at the candidate camera placement position. An image acquisition process is performed in which a pseudo-camera, which is a camera with the accuracy expected in the real world, is placed at the location where the pseudo-light is positioned, and an image of the target to be captured in the 3D virtual space is acquired by the pseudo-camera while the pseudo-light is illuminating it, and An arrangement optimization process that optimizes the arrangement of the pseudolights based on the acquired image and identifies the optimized arrangement of the pseudolights as the arrangement of the camera. A camera placement optimization program to enable the execution of the program.

10. On the computer, In the pseudo-light placement process, multiple pseudo-lights are placed in a 3D virtual space. The image acquisition process involves capturing an image of the target object while multiple placed pseudo-lights are illuminating it. The camera placement optimization method according to claim 9.

Images