Disaster information processing apparatus, method of operating the disaster information processing apparatus, operating program of the disaster information processing apparatus, and disaster information processing system

Abstract

Provided are a disaster information processing device, an operation method for the disaster information processing device, an operation program for the disaster information processing device, and a disaster information processing system which can control an operation of a monitoring camera which is suitable for an environmental condition of a disaster occurrence area. According to the present invention, an effective visual field range derivation unit acquires, by deriving an effective visual field range from which a damage state of the area can be ascertained and which is changed according to a local environmental condition, the effective visual field range that is in a bird's eye image of the area and captured by the monitoring camera. A control signal generation unit generates a control signal of the monitoring camera according to the effective visual field range. The operation of the monitoring camera is controlled by the control signal.

Description

Technical Field

[0001] The technology of the present disclosure relates to a disaster information processing apparatus, an operation method of the disaster information processing apparatus, an operation program of the disaster information processing apparatus, and a disaster information processing system.

Background Art

[0002] Various disasters such as earthquakes, tsunamis, volcanic eruptions, floods and / or landslides caused by heavy rain, and large-scale fires occur in various places. Conventionally, various measures have been proposed to grasp the damage situation of such disasters. For example, Patent Document 1 describes a technique of acquiring an aerial view image of a disaster-stricken area photographed by a surveillance camera and grasping the damage situation of the disaster from the acquired aerial view image. In Patent Document 1, the installation position of the surveillance camera, the shooting range of the surveillance camera, etc. are stored, and a surveillance camera capable of shooting a desired aerial view image is identified based on such information.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] The shooting range of the surveillance camera described in Patent Document 1 is one range determined based on the performance of the surveillance camera assuming a specific condition of the environmental conditions in the disaster-stricken area. However, the effective visual field range in which the damage situation of the disaster can actually be grasped in the aerial view image changes according to the environmental conditions in the disaster-stricken area such as the weather (sunny, cloudy, rainy, snowy, etc.), sand and dust due to the collapse of buildings, or smoke from a fire. For this reason, there has been a risk of misidentifying a surveillance camera capable of shooting a desired aerial view image based only on the information of the shooting range of the surveillance camera.

[0005] One embodiment of the technology disclosed herein provides a disaster information processing device capable of controlling the operation of a surveillance camera adapted to the environmental conditions in a disaster area, a method of operating the disaster information processing device, an operation program for the disaster information processing device, and a disaster information processing system.

Means for Solving the Problems

[0006] The disaster information processing device of the present disclosure includes a processor and a memory connected to or incorporated in the processor. The processor acquires an effective field of view range in an aerial image of the disaster area taken by the surveillance camera, which is an effective field of view range that enables grasping of the damage situation in the disaster area and changes according to the environmental conditions in the disaster area, and controls the operation of the surveillance camera based on the acquired effective field of view range.

[0007] Preferably, based on the effective field of view range, the processor performs at least any one of setting the zoom magnification of the surveillance camera, setting the tilt angle of the surveillance camera, and setting whether to take an aerial image.

[0008] Preferably, the processor acquires the effective field of view range from an aerial image taken by the surveillance camera in real time.

[0009] A plurality of patterns of effective field of view ranges corresponding to environmental conditions are stored in advance in the storage unit, and preferably, the processor acquires from the storage unit the effective field of view range corresponding to the current environmental conditions in the disaster area.

[0010] There are a plurality of surveillance cameras, and preferably, the processor controls the operation of each of the plurality of surveillance cameras based on the effective field of view range of each of the plurality of surveillance cameras.

[0011] Preferably, the processor analyzes the damage situation of each building in the disaster area using the aerial image.

[0012] When the processor has photographed a building to be analyzed for damage status with a plurality of surveillance cameras, it is preferable to analyze the damage status of the building to be analyzed using the bird's-eye view images taken by the plurality of surveillance cameras respectively.

[0013] The processor preferably analyzes the damage status for each section including a plurality of adjacent buildings in the disaster area using the bird's-eye view image.

[0014] When the processor has photographed a section to be analyzed for damage status with a plurality of surveillance cameras, it is preferable to analyze the damage status of the section to be analyzed using the bird's-eye view images taken by the plurality of surveillance cameras respectively.

[0015] The operation method of the disaster information processing apparatus of the present disclosure includes: obtaining an effective field of view range in the bird's-eye view image of the disaster area photographed by the surveillance camera, which is an effective field of view range in which the damage status of the disaster area can be grasped and which changes according to the environmental conditions of the disaster area; and controlling the operation of the surveillance camera based on the obtained effective field of view range.

[0016] The operation program of the disaster information processing apparatus of the present disclosure causes a computer to execute a process including: obtaining an effective field of view range in the bird's-eye view image of the disaster area photographed by the surveillance camera, which is an effective field of view range in which the damage status of the disaster area can be grasped and which changes according to the environmental conditions of the disaster area; and controlling the operation of the surveillance camera based on the obtained effective field of view range.

[0017] The disaster information processing system of the present disclosure includes a surveillance camera that photographs a bird's-eye view image for grasping the damage status of a disaster area, a processor, and a memory connected to or built in the processor. The processor obtains an effective field of view range in which the damage status in the bird's-eye view image can be grasped and which changes according to the environmental conditions of the disaster area, and controls the operation of the surveillance camera based on the obtained effective field of view range.

Advantages of the Invention

[0018] According to the technology of the present disclosure, it is possible to provide a disaster information processing device capable of controlling the operation of a surveillance camera adapted to the environmental conditions of a disaster area, a method for operating the disaster information processing device, an operation program of the disaster information processing device, and a disaster information processing system.

Brief Description of the Drawings

[0019]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

Figure 21

Figure 22

Figure 23

Figure 24

Figure 25

Figure 26

Embodiments for Carrying Out the Invention

[0020] [First Embodiment] As shown in FIG. 1 as an example, a disaster information processing system 2 for grasping the damage situation of a disaster includes a surveillance camera 10 and a disaster information processing server 11. The surveillance camera 10 is installed, for example, on the rooftop of a high-rise building 12 about 50 m to 100 m above the ground. As shown by the arrow 13, the surveillance camera 10 can perform a tilting operation in the + direction (upward direction) and the - direction (downward direction), that is, a tilt motion. Further, the surveillance camera 10 has a zoom lens, and it is possible to set a zoom magnification in a range from an equal magnification (wide end) to, for example, 10 times (tele end). The disaster information processing server 11 is installed, for example, in a disaster countermeasure headquarters (office, government office, etc.) of a local government such as a prefecture, a city, a town, or a village. The disaster information processing server 11 is an example of a "disaster information processing device" according to the technology of the present disclosure. Note that the surveillance camera 10 may be able to perform a panning operation, that is, a swing motion in the left-right direction.

[0021] The surveillance camera 10 and the disaster information processing server 11 are communicably connected to each other via a network 14. The surveillance camera 10 and the disaster information processing server 11 are connected to the network 14 by a wired connection or a wireless connection. The network 14 is, for example, a WAN (Wide Area Network) such as the Internet or a public communication network. When using the WAN, it is preferable to construct a VPN (Virtual Private Network) or use a communication protocol with a high security level such as HTTPS (Hypertext Transfer Protocol Secure) in consideration of information security.

[0022] A client terminal 15 is also connected to the network 14 by a wired connection or a wireless connection. The client terminal 15 is, for example, a desktop personal computer assigned to the staff of the disaster countermeasure headquarters, and has a display 16 and an input device 17. Various screens are displayed on the display 16. The input device 17 is a keyboard, a mouse, a touch panel, a microphone, or the like. In FIG. 1, only one client terminal 15 is depicted, but of course, there may be a plurality of client terminals 15.

[0023] As an example, as shown in FIG. 2, the surveillance camera 10 captures the imaging range 21 including the area 20 according to a preset frame rate (for example, 30 fps (frames per second)), and outputs an aerial image 22 of the area 20. The imaging range 21 is a range determined based on the performance of the surveillance camera 10 when the zoom magnification is at the equal magnification (wide end) and the tilt angle is 0° (reference angle). The imaging range 21 is, for example, 4000 m when expressed by the distance from the surveillance camera 10. Here, the bottom side of the imaging range 21 (the lower side of the aerial image 22) is set to a distance of 0 m from the surveillance camera 10. The area 20 is an area where a disaster has occurred and a disaster countermeasure headquarters is located, and is an example of the "disaster-stricken area" according to the technology of the present disclosure.

[0024] As an example, as shown in FIG. 3, the computer constituting the disaster information processing server 11 includes a storage 30, a memory 31, a CPU (Central Processing Unit) 32, and a communication unit 33. These are interconnected via a bus line 34. The storage 30 is an example of the "storage unit" according to the technology of the present disclosure. Also, The CPU 32 is an example of the "processor" according to the technology of the present disclosure.

[0025] The storage 30 is a hard disk drive built into the computer constituting the disaster information processing server 11 or connected through a cable or network. Alternatively, the storage 30 is a disk array with multiple hard disk drives installed. The storage 30 stores control programs such as an operating system, various application programs, and various data associated with these programs. Note that a solid state drive may be used instead of the hard disk drive.

[0026] The memory 31 is a work memory for the CPU 32 to execute processing. The CPU 32 loads the program stored in the storage 30 into the memory 31 and executes the processing according to the program. Thereby, the CPU 32 comprehensively controls the operations of each part of the computer. The communication unit 33 controls the transmission of various information with external devices such as the monitoring camera 10. Note that the memory 31 may be built into the CPU 32.

[0027] As an example, as shown in FIG. 4, an operation program 40 is stored in the storage 30 of the disaster information processing server 11. The operation program 40 is an application program for causing the computer to function as the disaster information processing server 11. That is, the operation program 40 is an example of the "operation program of the disaster information processing device" according to the technology of the present disclosure.

[0028] When the operation program 40 is started, the CPU 32 of the computer constituting the disaster information processing server 11 functions as a read / write (hereinafter abbreviated as RW (Read Write)) control unit 45, a valid viewing range derivation unit 46, a control signal generation unit 47, a transmission control unit 48, a damage situation analysis unit 49, and a screen distribution control unit 50 in cooperation with the memory 31 and the like.

[0029] The RW control unit 45 controls the storage of various data in the storage 30 and the reading of various data in the storage 30. For example, the RW control unit 45 receives the bird's-eye view image 22 from the monitoring camera 10 and stores the received bird's-eye view image 22 in the storage 30. When the RW control unit 45 receives a processing request (not shown) from the client terminal 15, the RW control unit 45 reads out the bird's-eye view image 22 from the storage 30 and outputs the read bird's-eye view image 22 to the valid viewing range derivation unit 46. Also, when the RW control unit 45 receives a distribution request (not shown) from the client terminal 15, the RW control unit 45 reads out the bird's-eye view image 22 from the storage 30 and outputs the read bird's-eye view image 22 to the damage situation analysis unit 49. Note that the storage of the bird's-eye view image 22 in the storage 30 is performed in response to an instruction from the staff of the disaster countermeasure headquarters.

[0030] The effective viewing range derivation unit 46 derives the effective viewing range 55 in the bird's-eye view image 22 obtained with the default settings of the zoom magnification being equal magnification and the tilt angle being 0°. The effective viewing range 55 differs from the shooting range 21 shown in FIG. 2 and varies according to the environmental conditions of the area 20. The effective viewing range derivation unit 46 outputs the derived effective viewing range 55 to the control signal generation unit 47.

[0031] The control signal generation unit 47 generates a control signal 56 for the surveillance camera 10 according to the effective viewing range 55. The control signal generation unit 47 outputs the generated control signal 56 to the transmission control unit 48. The transmission control unit 48 performs control to transmit the control signal 56 to the surveillance camera 10.

[0032] The damage situation analysis unit 49 analyzes the damage situation 69 (see FIG. 11) of the disaster in the area 20 based on the bird's-eye view image 22. The damage situation analysis unit 49 outputs the analysis result 57 of the damage situation 69 to the screen distribution control unit 50.

[0033] The screen distribution control unit 50 generates a damage situation display screen 58 based on the analysis result 57. The screen distribution control unit 50 performs control to distribute the screen data of the generated damage situation display screen 58 to the client terminal 15 that is the source of the distribution request. The screen data is, for example, screen data for web distribution created by a markup language such as XML (Extensible Markup Language). The client terminal 15 reproduces and displays the damage situation display screen 58 on the web browser based on the screen data. Note that, instead of XML, other data description languages such as JSON (Javascript (registered trademark) Object Notation) may be used.

[0034] As an example, as shown in FIG. 5, the effective visual field range derivation unit 46 includes a building information adding unit 60, a building image extraction unit 61, a first processing unit 62, and an effective visual field range determination unit 63. The building information adding unit 60 refers to the map 64 with building information, and adds building information 65 to each building 78 (see FIG. 7) shown in the bird's-eye view image 22, and makes the bird's-eye view image 22 into a bird's-eye view image 22I with building information. The building information adding unit 60 outputs the bird's-eye view image 22I with building information to the building image extraction unit 61.

[0035] The map 64 with building information is stored in the storage 30, read out from the storage 30 by the RW control unit 45, and output to the building information adding unit 60. The map 64 with building information is a three-dimensional map of the area 20, in which feature points such as the corners of the rooftops and building information 65 are associated with each building 78. Specifically, the building information 65 is the name of the owner of the building (residence) 78 such as "Ichio Fuji", or the name of the building 78 such as "Fuji Building No. 1". In addition, the building information 65 also includes the distance from the monitoring camera 10 of the building 78 and the address of the building 78.

[0036] Based on the longitude and latitude information of the installation position of the monitoring camera 10, the tilt angle, etc., the building information adding unit 60 aligns the orientation of the buildings in the map 64 with building information with the orientation of the buildings 78 shown in the bird's-eye view image 22. In addition, the building information adding unit 60 extracts feature points such as the corners of the rooftops of the buildings 78 shown in the bird's-eye view image 22. The building information adding unit 60 matches the map 64 with building information aligned with the orientation of the buildings 78 shown in the bird's-eye view image 22 and the bird's-eye view image 22, and searches for the position where the correlation between the feature points of the map 64 with building information and the feature points of the bird's-eye view image 22 is the highest. Then, at the position where the correlation is the highest, the building information 65 in the map 64 with building information is added to each building 78 in the bird's-eye view image 22.

[0037] The building image extraction unit 61 extracts building images 66 of any five buildings 78 from the bird's-eye view image 22I with building information at intervals of 10 m from the surveillance camera 10, for example. The building image extraction unit 61 uses, for example, a machine learning model (not shown) that takes the bird's-eye view image 22 as an input image and outputs an image of each individual building 78 shown in the bird's-eye view image 22. The building image extraction unit 61 outputs a building image group 67 including a set of building images 66 of any five buildings 78 and building information 65 at intervals of 10 m from the surveillance camera 10 to the first processing unit 62.

[0038] The first processing unit 62 inputs the building image 66 to the damage situation analysis model 68. Then, it causes the damage situation analysis model 68 to output the damage situation 69. The damage situation 69 assumes disasters such as earthquakes and is one of "total collapse", "partial collapse", "safe", and "unknown". The first processing unit 62 causes the damage situation analysis model 68 to output the damage situation 69 for all of the building images 66 of the five buildings 78 at intervals of 10 m from the surveillance camera 10 included in the building image group 67. The first processing unit 62 outputs an analysis result 70 for determining the effective field of view range, which summarizes the damage situation 69 for each building 78 together with the distance from the surveillance camera 10, to the effective field of view range determination unit 63.

[0039] The effective field of view range determination unit 63 determines the effective field of view range 55 based on the analysis result 70 for determining the effective field of view range. In FIG. 5, cases are illustrated where the damage situation 69 of buildings 78 such as "Fujii Ichio" and "Fujii Jiro" at a distance of 0 m from the surveillance camera 10 is "safe", and the damage situation 69 of "Fuji No. 1 Building" at a distance of 1300 m from the surveillance camera 10 is "unknown". Also illustrated is a case where the effective field of view range 55 is narrowed compared to the shooting range 21 due to the influence of clouds, fog, rain, snow, smog, etc., and the effective field of view range 55 is determined to be "1200 m".

[0040] The damage situation analysis model 68 is a machine learning model constructed by methods such as neural networks, support vector machines, and boosting. The damage situation analysis model 68 is stored in the storage 30 and is read from the storage 30 by the RW control unit 45 and output to the first processing unit 62.

[0041] As an example, as shown in FIG. 6, in the learning phase, the damage situation analysis model 68 is given learning data 75 and learned. The learning data 75 is a pair of a learning building image 66L and a correct damage situation 69CA corresponding to the learning building image 66L. The learning building image 66L is obtained by inputting an aerial image of a certain area into the building image cutting unit 61. The correct damage situation 69CA is the result of an actual determination by a qualified person such as a residential damage appraiser of the damage situation 69 of the building 78 shown in the learning building image 66L.

[0042] In the learning phase, the learning building image 66L is input to the damage situation analysis model 68. The damage situation analysis model 68 outputs a learning damage situation 69L for the learning building image 66L. Based on this learning damage situation 69L and the correct damage situation 69CA, a loss calculation of the damage situation analysis model 68 using a loss function is performed. Then, various coefficients of the damage situation analysis model 68 are updated and set according to the result of the loss calculation, and the damage situation analysis model 68 is updated according to the updated settings.

[0043] In the learning phase of the damage situation analysis model 68, the above series of processes of inputting the learning building image 66L into the damage situation analysis model 68, outputting the learning damage situation 69L from the damage situation analysis model 68, loss calculation, update setting, and updating of the damage situation analysis model 68 are repeatedly performed while the learning data 75 is exchanged. The repetition of the above series of processes ends when the discrimination accuracy of the learning damage situation 69L with respect to the correct damage situation 69CA reaches a predetermined setting level. In this way, the damage situation analysis model 68 whose discrimination accuracy has reached the setting level is stored in the storage 30 and used in the first processing unit 62.

[0044] As an example, as shown in FIG. 7, the effective field of view determination unit 63 counts the number of stories of the building 78 where the damage situation 69 is "unknown" for each distance from the surveillance camera 10. The effective field of view determination unit 63 determines, among the distances from the surveillance camera 10, that the number of stories of the building 78 where the damage situation 69 is "unknown" is less than 5, and thereafter, the distance at which the number of stories of the building 78 where the damage situation 69 is "unknown" is 5 for two or more consecutive times is determined as the effective field of view 55. In FIG. 7, as in the case of FIG. 5, an example where the effective field of view 55 is determined to be "1200 m" is illustrated.

[0045] As an example, as shown in FIG. 8, when the effective field of view 55 is "1200 m" illustrated in FIG. 7, the control signal generation unit 47 generates a control signal 56 with an equal zoom ratio and a tilt angle of 0° for photographing within the effective field of view 55 at a distance of 1200 m or less from the surveillance camera 10. Further, the control signal generation unit 47 generates a control signal 56 with a zoom ratio of 10 times and a tilt angle of -5° for photographing in the range exceeding 1200 m from the surveillance camera 10.

[0046] FIG. 9 shows a state where the surveillance camera 10 photographs the aerial image 22 of the area 20 based on the control signal 56 with an equal zoom ratio and a tilt angle of 0°. In this case, all the buildings 78 in the photographing range 21 are shown in the aerial image 22. However, since the effective field of view 55 is 1200 m, the buildings 78 outside the effective field of view 55 are shown in a state where the damage situation cannot be grasped.

[0047] FIG. 10 shows a state in which the surveillance camera 10 captures an aerial view image 22 of the area 20 based on a control signal 56 with a zoom magnification of 10 times and a tilt angle of -5°. In this case, the shooting range 80 is a narrower range than in the case of FIG. 9 and exceeds the effective field of view range 55 because the zoom magnification and the tilt angle are different. Therefore, in the aerial view image 22 captured in this case, there is a high possibility that buildings 78 outside the effective field of view range 55 are shown in a state where the damage situation can be grasped. Note that the storage 30 stores both the aerial view image 22 obtained in the case of FIG. 9 and the aerial view image 22 obtained in the case of FIG. 10, and both aerial view images 22 are output to the damage situation analysis unit 49. Then, analysis results 57 based on each of the two aerial view images 22 are output.

[0048] As shown in FIG. 11 as an example, the damage situation analysis unit 49 includes a building information adding unit 85, a building image cutting unit 86, and a second processing unit 87. Similar to the building information adding unit 60 of the effective field of view range deriving unit 46, the building information adding unit 85 refers to the map 64 with building information and adds building information 65 to each building 78 shown in the aerial view image 22, making the aerial view image 22 an aerial view image 22I with building information. The building information adding unit 85 outputs the aerial view image 22I with building information to the building image cutting unit 86.

[0049] The building image cutting unit 86 cuts out building images 66 from the aerial view image 22I with building information, similar to the building image cutting unit 61 of the effective field of view range deriving unit 46. However, the building image cutting unit 86 cuts out the building images 66 of all the buildings 78 shown in the aerial view image 22I with building information. The building image cutting unit 86 outputs a building image group 88 including a set of the building images 66 of all the buildings 78 and the building information 65 to the second processing unit 87.

[0050] Similar to the first processing unit 62 of the effective field of view range derivation unit 46, the second processing unit 87 inputs the building image 66 into the damage situation analysis model 68. Then, it causes the damage situation analysis model 68 to output the damage situation 69. The second processing unit 87 causes the damage situation analysis model 68 to output the damage situation 69 for the building images 66 of all the buildings 78 included in the building image group 88. The second processing unit 87 outputs the analysis result 57 that summarizes the damage situation 69 for each building 78. In FIG. 11, the case where the damage situation 69 of buildings 78 such as "Fujii Ichio" and "Fujii Jiro" is "safe" and the damage situation 69 of buildings such as "Fuji No. 1 Building" is "semi-damaged" is illustrated.

[0051] In the analysis result 70 for determining the effective field of view range shown in FIG. 5, the damage situation 69 of the "Fuji No. 1 Building" was "unknown", but in FIG. 11, it is "semi-damaged". This indicates that based on the bird's-eye view image 22 obtained in the case of FIG. 10, the damage situation 69 of the "Fuji No. 1 Building" in the range at a distance of 1300 m beyond the effective field of view range 55 could be determined as "semi-damaged".

[0052] As an example, as shown in FIG. 12, the damage situation display screen 58 displayed on the display 16 of the client terminal 15 has a building-by-building damage situation display area 90 and a statistical damage situation display area 91. In the building-by-building damage situation display area 90, the building information 65, the building image 66, and the damage situation 69 of each building are displayed. In the statistical damage situation display area 91, the total number of buildings in the region 20 that are completely damaged, semi-damaged, safe, and unknown is displayed. When the confirmation button 92 is selected, the display of the damage situation display screen 58 is cleared.

[0053] Next, the operation of the above configuration will be described with reference to the flowcharts of FIGS. 13 and 14. First, when the operation program 40 is started in the disaster information processing server 11, as shown in FIG. 4, the CPU 32 of the disaster information processing server 11 functions as an RW control unit 45, a valid visual field range derivation unit 46, a control signal generation unit 47, a transmission control unit 48, a damage situation analysis unit 49, and a screen distribution control unit 50. As shown in FIG. 5, the valid visual field range derivation unit 46 includes a building information adding unit 60, a building image cutting unit 61, a first processing unit 62, and a valid visual field range determination unit 63. Also, as shown in FIG. 11, the damage situation analysis unit 49 includes a building information adding unit 85, a building image cutting unit 86, and a second processing unit 87.

[0054] An aerial image 22 of the area 20 where a disaster has occurred is transmitted from the surveillance camera 10 to the disaster information processing server 11. As shown in FIG. 13 as an example, in the disaster information processing server 11, the aerial image 22 is received by the RW control unit 45 (step ST100). The aerial image 22 is stored in the storage 30 by the RW control unit 45 in response to an instruction from the staff of the disaster countermeasure headquarters.

[0055] When a processing request (not shown) from the client terminal 15 is received, the RW control unit 45 reads the aerial image 22 from the storage 30, and the read aerial image 22 is output from the RW control unit 45 to the valid visual field range derivation unit 46. Then, as shown in FIGS. 5 and 7, the valid visual field range 55 in the aerial image 22 is derived by the valid visual field range derivation unit 46 (step ST110). The valid visual field range 55 is output from the valid visual field range derivation unit 46 to the control signal generation unit 47.

[0056] Based on the valid visual field range 55, a control signal 56 as shown in FIG. 8 is generated by the control signal generation unit 47 (step ST120). The control signal 56 is output from the control signal generation unit 47 to the transmission control unit 48. Then, the control signal 56 is transmitted to the surveillance camera 10 by the transmission control unit 48 (step ST130).

[0057] In the surveillance camera 10, as shown in FIGS. 9 and 10, the operation is controlled according to the control signal 56. Then, the bird's-eye view image 22 in the state shown in FIGS. 9 and 10 is stored in the storage 30 by the RW control unit 45 in response to an instruction from the staff of the disaster countermeasure headquarters.

[0058] As an example, as shown in FIG. 14, when a distribution request from the client terminal 15 is received, the RW control unit 45 reads out the bird's-eye view image 22 from the storage 30, and the read bird's-eye view image 22 is output from the RW control unit 45 to the damage situation analysis unit 49. Then, as shown in FIG. 11, in the damage situation analysis unit 49, the damage situation 69 of each building 78 in the area 20 is analyzed using the bird's-eye view image 22 (step ST200). The analysis result 57 of the damage situation 69 is output from the damage situation analysis unit 49 to the screen distribution control unit 50.

[0059] Based on the analysis result 57, the screen distribution control unit 50 generates the damage situation display screen 58 shown in FIG. 12. The screen data of the damage situation display screen 58 is distributed by the screen distribution control unit 50 to the client terminal 15 that is the source of the distribution request (step ST210). The damage situation display screen 58 is displayed on the display 16 of the client terminal 15 that is the source of the distribution request and is provided for the staff of the disaster countermeasure headquarters to view.

[0060] As described above, the CPU 32 of the disaster information processing server 11 includes an effective field of view derivation unit 46 and a control signal generation unit 47. The effective field of view derivation unit 46 obtains the effective field of view 55 in the bird's-eye view image 22 of the area 20 photographed by the surveillance camera 10, which can grasp the damage situation 69 of the area 20 and changes according to the environmental conditions of the area 20, by deriving the effective field of view 55. The control signal generation unit 47 generates a control signal 56 for the surveillance camera 10 according to the effective field of view 55. The operation of the surveillance camera 10 is controlled by this control signal 56. Therefore, it is possible to control the operation of the surveillance camera 10 to conform to the environmental conditions of the area 20.

[0061] As shown in FIGS. 8 to 10, the control signal generation unit 47 generates a control signal 56 for setting the zoom magnification of the surveillance camera 10 and the tilt angle of the surveillance camera 10 based on the effective viewing range 55. For this reason, in the default setting where the zoom magnification is equal magnification and the tilt angle is 0°, the possibility that the building 78 with the damage situation 69 being "unknown" can be included in the bird's-eye view image 22 with relatively high image quality increases, and the number of buildings 78 with the damage situation 69 being "unknown" can be reduced.

[0062] The effective viewing range derivation unit 46 derives the effective viewing range 55 from the bird's-eye view image 22 captured by the surveillance camera 10 in real time. For this reason, an effective viewing range 55 that better conforms to the current environmental conditions of the area 20 can be obtained, and thus the operation of the surveillance camera 10 can be controlled to better conform to the current environmental conditions of the area 20.

[0063] The damage situation analysis unit 49 analyzes the damage situation 69 of each building 78 in the area 20 using the bird's-eye view image 22. For this reason, the damage situation 69 of the building 78 can be easily grasped without the trouble of actually walking around the area 20.

[0064] The control signal 56 is not limited to the content exemplified in FIG. 8. As an example, as shown in FIG. 15, when the effective viewing range 55 is, for example, 1200 m, a control signal 56 may be generated with the zoom magnification set to equal magnification and the tilt angle set to -20° for photographing the effective viewing range 55 within a distance of 1200 m or less from the surveillance camera 10.

[0065] FIG. 16 shows a state in which the surveillance camera 10 photographs the bird's-eye view image 22 of the area 20 based on the control signal 56 of FIG. 15 with the zoom magnification set to equal magnification and the tilt angle set to -20°. The photographing range 102 in this case is almost the same range as the effective viewing range 55. For this reason, in the bird's-eye view image 22 photographed in this case, buildings 78 outside the effective viewing range 55 are not shown, and only buildings 78 existing almost within the effective viewing range 55 are shown.

[0066] Thus, according to the control signal 56 that sets the shooting range 102 of the surveillance camera 10 to be approximately the same as the effective field of view range 55, buildings 78 in a range exceeding the effective field of view range 55, where the probability of the damage situation being "unknown" is very high, do not appear in the bird's-eye view image 22. Therefore, the processing burden on the damage situation analysis unit 49 for buildings 78 in a range exceeding the effective field of view range 55 can be reduced.

[0067] The control signal 56 may be as shown in FIG. 17. The control signal 56 shown in FIG. 17 is configured to stop the surveillance camera 10 from shooting the bird's-eye view image 22 when the effective field of view range 55 is less than a preset threshold value, for example, less than 500 m. As the threshold value, a value is set that is considered to have no significant merit compared to the case of analyzing the damage situation 69 based on the bird's-eye view image 22 in the effective field of view range 55 or when analyzing the damage situation manually.

[0068] According to the control signal 56 that stops the surveillance camera 10 from shooting the bird's-eye view image 22 when the effective field of view range 55 is less than the threshold value, the surveillance camera 10 can be prevented from performing unnecessary shooting.

[0069] The machine learning model used in the building image extraction unit 61 may be a model that outputs an evaluation value of the image quality of the building 78 shown in the building image 66 in addition to the building image 66. Then, the effective field of view range 55 may be determined based on whether or not a building image 66 with an evaluation value of the image quality of the building 78 equal to or higher than a preset threshold value can be extracted. Specifically, the distance immediately before the distance at which only building images 66 with an evaluation value of the image quality of the building 78 less than the threshold value can be extracted is determined as the effective field of view range 55.

[0070] The landmark building and its distance may be registered in advance, and the effective field of view range 55 may be determined based on the damage situation 69 for the building image 66 in which the landmark building is extracted.

[0071] [Second Embodiment] In the first embodiment described above, the effective visual field range derivation unit 46 derives the effective visual field range 55 from the bird's-eye view image 22 captured by the surveillance camera 10 in real time, but it is not limited to this. As in the second embodiment shown in FIG. 18, the effective visual field range 55 corresponding to a plurality of patterns of environmental conditions may be stored in advance.

[0072] As an example, as shown in FIG. 18, in the second embodiment, the effective visual field range table 110 is stored in advance in the storage 30. The effective visual field range table 110 is a table in which the effective visual field range 55 for each environmental condition is registered. The environmental conditions include "sunny (spring, summer)", "sunny (autumn, winter)", "cloudy", "rain", "snow", "smog", etc. Spring and summer are, for example, from March to September, and autumn and winter are, for example, from October to February. In the effective visual field range 55, for example, "2500 m" is registered for "sunny (spring, summer)", "3000 m" is registered for "sunny (autumn, winter)", and "2000 m" is registered for "cloudy". The effective visual field range 55 can be obtained, for example, by deriving the effective visual field range 55 from the bird's-eye view images 22 actually captured multiple times before the occurrence of a disaster in each environmental condition, like the effective visual field range derivation unit 46 of the first embodiment, and obtaining the average value.

[0073] The RW control unit 45 receives the current environmental condition 111 of the area 20. The RW control unit 45 acquires the effective visual field range 55 corresponding to the received environmental condition 111 by reading it from the effective visual field range table 110 in the storage 30. The RW control unit 45 outputs the effective visual field range 55 to the control signal generation unit 47. Since the subsequent processing is the same as that of the first embodiment, the description is omitted. In FIG. 18, the case where the environmental condition 111 is "sunny (autumn, winter)" and "3000 m" is read as the effective visual field range 55 is illustrated. Note that the environmental condition 111 may be input by the staff of the disaster countermeasure headquarters, or may be received from a public institution such as the Meteorological Agency via the network 14.

[0074] Thus, in the second embodiment, the effective visual field range 55 corresponding to a plurality of patterns of environmental conditions is pre-stored in the storage 30, and the RW control unit 45 acquires the effective visual field range 55 corresponding to the current environmental condition 111 of the area 20 by reading it from the storage 30. Therefore, unlike the first embodiment, the labor of deriving the effective visual field range 55 from the bird's-eye view image 22 can be saved.

[0075] Environmental conditions for storing the effective visual field range 55 in the effective visual field range table 110 may include "dust storm", "pollen scattering", "midsummer day", and "sweltering day". In the case of "dust storm" and "pollen scattering", similar to "smog", the effective visual field range 55 is narrowed due to the influence of fine particles scattered in the air. Also, in the case of "midsummer day" and "sweltering day", the distance appears hazy due to the influence of heat haze.

[0076] [Third Embodiment] In each of the above embodiments, the case where there is one monitoring camera 10 has been illustrated, but it is not limited to this. As in the third embodiment shown in FIGS. 19 and 20, there may be a plurality of monitoring cameras 10.

[0077] As an example, as shown in FIG. 19, a plurality of monitoring cameras 10A, 10B, 10C,... are connected to the disaster information processing server 120 of this embodiment. The disaster information processing server 120 receives the bird's-eye view images 22A from the monitoring camera 10A, the bird's-eye view images 22B from the monitoring camera 10B, the bird's-eye view images 22C from the monitoring camera 10C,.... The effective visual field range derivation unit 121 derives the effective visual field range 55A in the bird's-eye view image 22A, the effective visual field range 55B in the bird's-eye view image 22B, the effective visual field range 55C in the bird's-eye view image 22C,.... The control signal generation unit 122 generates control signals 56A corresponding to the effective visual field range 55A, control signals 56B corresponding to the effective visual field range 55B, control signals 56C corresponding to the effective visual field range 55C,.... The control signal 56A is transmitted to the monitoring camera 10A, the control signal 56B is transmitted to the monitoring camera 10B, the control signal 56C is transmitted to the monitoring camera 10C,....

[0078] FIG. 20 shows a specific example of controlling the operations of each of a plurality of surveillance cameras 10. In FIG. 20, for simplicity, the mode of controlling the operations of each of two surveillance cameras 10A and 10B is shown. Incidentally, surveillance camera 10A is installed on the roof of high-rise building 12A, and surveillance camera 10B is installed on the roof of high-rise building 12B facing high-rise building 12A.

[0079] As shown above the arrow, first, surveillance camera 10A captures an aerial view image 22A of the default shooting range 21A based on a control signal 56A that sets the zoom magnification to the same magnification and the tilt angle to 0°. Similarly, surveillance camera 10B also captures an aerial view image 22B of the default shooting range 21B based on a control signal 56B that sets the zoom magnification to the same magnification and the tilt angle to 0°.

[0080] Here, consider the case where the effective field of view range 55A of the aerial view image 22A by surveillance camera 10A becomes less than 500 m due to the smoke 125 of a fire. In this case, as shown below the arrow, the control signal generation unit 122 generates a control signal 56A that causes surveillance camera 10A to stop capturing the aerial view image 22A. Further, the control signal generation unit 122 generates a control signal 56B that sets the zoom magnification of surveillance camera 10B to 10 times and the tilt angle to -5° in order to include building 78, which is outside the effective field of view range 55B of the aerial view image 22B by surveillance camera 10B and is directly below the smoke 125 of the fire, in the aerial view image 22B.

[0081] As described above, in the third embodiment, there are a plurality of surveillance cameras 10, and the control signal generation unit 122 generates a control signal 56 for controlling the operations of each of the plurality of surveillance cameras 10 based on the effective field of view range 55 of each of the plurality of surveillance cameras 10. Therefore, as shown in FIG. 20, it is possible to capture the aerial view image 22 in cooperation with a plurality of surveillance cameras 10, such as covering a range that cannot be covered by one surveillance camera 10 with the other surveillance camera 10.

[0082] As an example of controlling the operation of each of the plurality of surveillance cameras 10, in addition to the example shown in FIG. 20, the following can be considered. For example, in bad weather such as thick fog or heavy rain, when the effective field of view range 55 of the overhead images 22 by all the surveillance cameras 10 is less than the threshold value, a control signal 56 for stopping the shooting of the overhead image 22 is generated for all the surveillance cameras 10, and the shooting of the overhead image 22 is stopped for all the surveillance cameras 10.

[0083] [Fourth Embodiment] In the first embodiment described above, the damage situation 69 is analyzed based only on the overhead image 22 captured by one surveillance camera 10, but it is not limited to this. As in the fourth embodiment shown in FIGS. 21 and 22, the damage situation may be analyzed based on the overhead images 22 respectively captured by a plurality of surveillance cameras 10. Hereinafter, for simplicity, the case of using two surveillance cameras 10A and 10B as in the case of FIG. 20 will be exemplified. And the shooting ranges of the surveillance cameras 10A and 10B overlap at least partially, and it is assumed that the same building 78 appears in different manners in the overhead image 22A and the overhead image 22B.

[0084] As shown in FIG. 21 as an example, the building information adding unit 131 of the damage situation analysis unit 130 in the fourth embodiment, similar to the building information adding unit 85 in the first embodiment, refers to the map 64 with building information, and for each building 78 shown in the overhead image 22A by the surveillance camera 10A, building information 65 is added, and the overhead image 22A is made into an overhead image 22AI with building information. Also, the building information adding unit 131 adds building information 65 to each building 78 shown in the overhead image 22B by the surveillance camera 10B, and makes the overhead image 22B into an overhead image 22BI with building information. The building image cutting unit 132 cuts out the first building image 66A from the overhead image 22AI with building information, and cuts out the second building image 66B from the overhead image 22BI with building information. The building image cutting unit 132 outputs a first building image group 88A including a plurality of sets of the first building image 66A and the building information 65, and a second building image group 88B including a plurality of sets of the second building image 66B and the building information 65 to the second processing unit 133.

[0085] The second processing unit 133 inputs the first building image 66A and the second building image 66B associated with the same building information 65 into the damage situation analysis model 134. Then, it causes the damage situation analysis model 134 to output the damage situation 135. Similar to the damage situation 69 in the first embodiment, the damage situation 135 is any one of "total destruction", "semi-destruction", "safe", and "unknown". The second processing unit 133 causes the damage situation analysis model 134 to output the damage situation 135 for all the first building images 66A and the second building images 66B included in the first building image group 88A and the second building image group 88B and associated with the same building information 65. For the first building images 66A and the second building images 66B not associated with the same building information 65, they are input into the damage situation analysis model 68 in the first embodiment to output the damage situation 69.

[0086] As an example, as shown in FIG. 22, in the learning phase, the damage situation analysis model 134 is given the learning data 140 and learns. The learning data 140 is a set of the first learning building image 66AL and the second learning building image 66BL, and the correct damage situation 135CA corresponding to the first learning building image 66AL and the second learning building image 66BL. The first learning building image 66AL is obtained by inputting an aerial image taken of a certain area by a certain surveillance camera into the building image extraction unit 132. The second learning building image 66BL is obtained by inputting an aerial image taken of a certain area by another surveillance camera into the building image extraction unit 132. The correct damage situation 135CA is the result of a qualified person such as a residential damage appraiser actually discriminating the damage situation 135 of the building 78 shown in the first learning building image 66AL and the second learning building image 66BL.

[0087] In the learning phase, for the damage situation analysis model 134, the first building image 66AL for learning and the second building image 66BL for learning are input. The damage situation analysis model 134 outputs the learning damage situation 135L for the first building image 66AL for learning and the second building image 66BL for learning. Based on this learning damage situation 135L and the correct damage situation 135CA, a loss calculation of the damage situation analysis model 134 using a loss function is performed. Then, various coefficients of the damage situation analysis model 134 are updated and set according to the result of the loss calculation, and the damage situation analysis model 134 is updated according to the update setting.

[0088] In the learning phase of the damage situation analysis model 134, the above series of processes including the input of the first building image 66AL for learning and the second building image 66BL for learning into the damage situation analysis model 134, the output of the learning damage situation 135L from the damage situation analysis model 134, the loss calculation, the update setting, and the update of the damage situation analysis model 134 are repeatedly performed while the learning data 140 is being exchanged. The repetition of the above series of processes ends when the discrimination accuracy of the learning damage situation 135L with respect to the correct damage situation 135CA reaches a predetermined setting level. In this way, the damage situation analysis model 134 whose discrimination accuracy has reached the setting level is stored in the storage 30 and used by the second processing unit 133.

[0089] As described above, in the fourth embodiment, when the building 78 to be analyzed for the damage situation 135 by the damage situation analysis unit 130 is photographed by a plurality of monitoring cameras 10, the damage situation 135 of the building 78 to be analyzed is analyzed using the bird's-eye view images 22 respectively photographed by the plurality of monitoring cameras 10. Therefore, the possibility of grasping the damage situation 135 of the building 78 that is not clear only from the bird's-eye view image 22 photographed by one monitoring camera 10 is increased, and as a result, the reliability of the analysis result 57 can be improved.

[0090] Note that the number of monitoring cameras 10 is not limited to two, and therefore, the building image 66 input to the damage situation analysis model 134 may also be a building image 66 cut out from the bird's-eye view images 22 respectively photographed by three or more monitoring cameras 10.

[0091] [Fifth Embodiment] In the fifth embodiment shown in FIGS. 23 and 24, the damage situation is analyzed not for each building 78 but for each section.

[0092] As shown in FIG. 23 as an example, the damage situation analysis unit 145 of the fifth embodiment includes a section image extraction unit 146 and a second processing unit 147. The section image extraction unit 146 refers to the landmark building information 148 and extracts a section image 149 for each section from the bird's-eye view image 22. A section is a plurality of regions obtained by dividing the area 20, and is a region including a plurality of adjacent buildings 78 in the area 20. In this example, the sections are chome such as "Fuji 1-chome" and "Watabuki Fox Hole 2-chome". The section image extraction unit 146 outputs a section image group 151 including a plurality of sets of a section image 149 and section information 150 representing the section of the section image 149 to the second processing unit 147.

[0093] The landmark building information 148 is stored in the storage 30, read out from the storage 30 by the RW control unit 45, and output to the section image extraction unit 146. The landmark building information 148 includes an image of a landmark building that is a building 78 located at the corner of each section and section information 150 of the section to which the landmark building belongs. The section image extraction unit 146 uses a well-known image recognition technique to find the landmark building from the bird's-eye view image 22 relying on the image of the landmark building. Then, the area surrounded by the line connecting the found landmark buildings is cut out from the bird's-eye view image 22 as the section image 149.

[0094] The second processing unit 147 inputs the block image 149 into the damage situation analysis model 152, and causes the damage situation analysis model 152 to output the damage situation 153. The damage situation 153 is one of "severe damage", "minor damage", and "unknown". The second processing unit 147 causes the damage situation analysis model 152 to output the damage situation 153 for all the block images 149 included in the block image group 151. The second processing unit 147 outputs an analysis result 154 obtained by summarizing the damage situation 153 for each block. In FIG. 23, a case where the damage situation 153 of blocks such as "Fuji 1-chome" and "Mianbu Foxhole 2-chome" is "severe damage" is illustrated.

[0095] Similar to the damage situation analysis models 68 and 134, the damage situation analysis model 152 is a machine learning model constructed by methods such as neural networks, support vector machines, and boosting. The damage situation analysis model 152 is stored in the storage 30, read from the storage 30 by the RW control unit 45, and output to the second processing unit 147.

[0096] As shown in FIG. 24 as an example, in the learning phase, the damage situation analysis model 152 is given the learning data 160 and learned. The learning data 160 is a pair of a learning block image 149L and a correct damage situation 153CA corresponding to the learning block image 149L. The learning block image 149L is obtained by inputting an aerial view image of a certain area into the block image cutting unit 146. The correct damage situation 153CA is the result of an actual discrimination by a qualified person such as a residential damage appraiser of the damage situation 153 of the block shown in the learning block image 149L.

[0097] In the learning phase, the damage situation analysis model 152 is input with the learning partition image 149L. The damage situation analysis model 152 outputs the learning damage situation 153L for the learning partition image 149L. Based on this learning damage situation 153L and the correct damage situation 153CA, a loss calculation of the damage situation analysis model 152 using a loss function is performed. Then, various coefficients of the damage situation analysis model 152 are updated according to the result of the loss calculation, and the damage situation analysis model 152 is updated according to the update settings.

[0098] In the learning phase of the damage situation analysis model 152, the above series of processes of inputting the learning partition image 149L into the damage situation analysis model 152, outputting the learning damage situation 153L from the damage situation analysis model 152, loss calculation, update setting, and updating the damage situation analysis model 152 are repeatedly performed while the learning data 160 is being exchanged. The repetition of the above series of processes ends when the discrimination accuracy of the learning damage situation 153L with respect to the correct damage situation 153CA reaches a predetermined setting level. In this way, the damage situation analysis model 152 whose discrimination accuracy has reached the setting level is stored in the storage 30 and used in the second processing unit 147.

[0099] As described above, in the fifth embodiment, the damage situation analysis unit 145 analyzes the damage situation 153 for each partition including a plurality of adjacent buildings 78 in the area 20. Therefore, the analysis of the damage situation 153 can be completed in a shorter time than when analyzing the damage situation of each individual building 78. As a result, the approximate damage situation 153 can be grasped quickly, albeit somewhat roughly.

[0100] [Sixth Embodiment] In the sixth embodiment shown in FIGS. 25 and 26, the damage situation for each section is analyzed based on the bird's-eye view images 22 respectively captured by a plurality of surveillance cameras 10. Hereinafter, for simplicity, the case of using two surveillance cameras 10A and 10B, similar to the case of FIG. 20 etc., will be exemplified. And it is assumed that the shooting ranges of the surveillance cameras 10A and 10B overlap at least partially, and the same section is shown in different manners in the bird's-eye view image 22A and the bird's-eye view image 22B.

[0101] As shown in FIG. 25 as an example, the section image cutting-out unit 166 of the damage situation analysis unit 165 in the sixth embodiment, similar to the section image cutting-out unit 146 in the fifth embodiment, refers to the landmark building information 148 and cuts out the first section image 149A for each section from the bird's-eye view image 22A captured by the surveillance camera 10A. Also, the section image cutting-out unit 166 cuts out the second section image 149B for each section from the bird's-eye view image 22B captured by the surveillance camera 10B. The section image cutting-out unit 166 outputs a first section image group 151A including a plurality of pairs of the first section image 149A and the section information 150, and a second section image group 151B including a plurality of pairs of the second section image 149B and the section information 150 to the second processing unit 167.

[0102] The second processing unit 167 inputs the first section image 149A and the second section image 149B associated with the same section information 150 into the damage situation analysis model 168. And it causes the damage situation analysis model 168 to output the damage situation 169. The damage situation 169 is, similar to the damage situation 153 in the fifth embodiment, any one of "severe damage", "minor damage", and "unknown". The second processing unit 167 causes the damage situation analysis model 168 to output the damage situation 169 for all the first section images 149A and the second section images 149B included in the first section image group 151A and the second section image group 151B and associated with the same building information 65. For the first section images 149A and the second section images 149B not associated with the same section information 150, they are input into the damage situation analysis model 152 in the fifth embodiment to output the damage situation 153.

[0103] As an example, as shown in FIG. 26, in the learning phase, the damage situation analysis model 168 is trained by being given learning data 170. The learning data 170 is a set of a first learning partition image 149AL and a second learning partition image 149BL, and a correct damage situation 169CA corresponding to the first learning partition image 149AL and the second learning partition image 149BL. The first learning partition image 149AL is obtained by inputting an aerial image taken by a certain surveillance camera of a certain area into the partition image extraction unit 166. The second learning partition image 149BL is obtained by inputting an aerial image taken by another surveillance camera of a certain area into the partition image extraction unit 166. The correct damage situation 169CA is the result of an actual determination by a qualified person such as a residential damage appraiser of the damage situation 169 of the partition shown in the first learning partition image 149AL and the second learning partition image 149BL.

[0104] In the learning phase, the first learning partition image 149AL and the second learning partition image 149BL are input into the damage situation analysis model 168. The damage situation analysis model 168 outputs a learning damage situation 169L for the first learning partition image 149AL and the second learning partition image 149BL. Based on this learning damage situation 169L and the correct damage situation 169CA, a loss calculation of the damage situation analysis model 168 using a loss function is performed. Then, various coefficients of the damage situation analysis model 168 are updated and set according to the result of the loss calculation, and the damage situation analysis model 168 is updated according to the update setting.

[0105] In the learning phase of the damage situation analysis model 168, a series of processes including the input of the first learning section image 149AL and the second learning section image 149BL for learning into the damage situation analysis model 168, the output of the learning damage situation 169L from the damage situation analysis model 168, loss calculation, update setting, and update of the damage situation analysis model 168 are repeatedly performed while the learning data 170 is being exchanged. The repetition of the series of processes ends when the discrimination accuracy of the learning damage situation 169L with respect to the correct damage situation 169CA reaches a predetermined setting level. In this way, the damage situation analysis model 168 whose discrimination accuracy has reached the setting level is stored in the storage 30 and used by the second processing unit 167.

[0106] As described above, in the sixth embodiment, when the damage situation analysis unit 165 photographs the section to be analyzed for the damage situation 169 with a plurality of monitoring cameras 10, the damage situation 169 of the section to be analyzed is analyzed using the bird's-eye view images 22 respectively photographed by the plurality of monitoring cameras 10. Therefore, the possibility of grasping the damage situation 169 of a section that is not clear only from the bird's-eye view image 22 photographed by one monitoring camera 10 is increased, and as a result, the reliability of the analysis result 154 can be improved.

[0107] Similar to the fourth embodiment described above, the number of monitoring cameras 10 is not limited to two, and therefore, the section image 149 input to the damage situation analysis model 168 may also be a section image 149 cut out from the bird's-eye view images 22 respectively photographed by three or more monitoring cameras 10.

[0108] The section including a plurality of adjacent buildings is not limited to the exemplary cho-me. A rectangular area of a predetermined size bounded by a road may be used as the section.

[0109] As the damage situations 153 and 169, any of "severe damage", "minor damage", and "unknown" is exemplified, but it is not limited to this.As with the damage situation 69 etc., any of "total destruction", "partial destruction", "safe", and "unknown" may also be used.

[0110] In each of the above embodiments, the monitoring camera 10 is assumed to be a visible light camera, but it is not limited to this. As the monitoring camera 10, an infrared camera may be prepared for photographing in the evening or at night.

[0111] In each of the above embodiments, as an example of the damage situation, assuming mainly an earthquake as a disaster, one of "total destruction", "semi-destruction", "safe", and "unknown" was cited, but it is not limited to this. Assuming a flood as a disaster, one of "floor flooding", "under-floor flooding", "safe", and "unknown" may be output as the damage situation. Also, assuming a large-scale fire as a disaster, one of "total burn", "semi-burn", "safe", and "unknown" may be output as the damage situation. A damage situation analysis model corresponding to the type of disaster may be prepared and the damage situation analysis model may be selectively used depending on the type of disaster.

[0112] The damage situation analysis model used in the second processing unit may also be a model that outputs the reliability of the damage situation for each building 78. In this case, in the case of the first embodiment using the bird's-eye view image 22 photographed by one monitoring camera 10, only the damage situation with a reliability equal to or higher than a preset first threshold value is adopted. On the other hand, in the case of the fourth embodiment using the bird's-eye view image 22 photographed by a plurality of monitoring cameras 10, only the damage situation with a reliability equal to or higher than a preset second threshold value is adopted. The second threshold value is set to a value lower than the first threshold value. The reason for setting the second threshold value lower than the first threshold value is that more reliability can be placed on the damage situation when using the bird's-eye view image 22 photographed by a plurality of monitoring cameras 10.

[0113] Similarly, the damage situation analysis model used in the second processing unit of the fifth and sixth embodiments may also be a model that outputs the reliability of the damage situation for each section. In this case, in the case of the fifth embodiment using the bird's-eye view image 22 photographed by one monitoring camera 10, only the damage situation with a reliability equal to or higher than a preset first threshold value is adopted. On the other hand, in the case of the sixth embodiment using the bird's-eye view image 22 photographed by a plurality of monitoring cameras 10, only the damage situation with a reliability equal to or higher than a preset second threshold value (< the first threshold value) is adopted.

[0114] When the effective viewing range 55 is less than a preset threshold value, super-resolution technology using a machine learning model may be applied to the aerial image 22 to make the aerial image 22 a super-resolved aerial image 22, and the damage situation may be analyzed using the super-resolved aerial image 22. However, since the super-resolved aerial image 22 is a so-called fake image, it is preferable to clearly indicate that it is only for reference on the damage situation display screen 58.

[0115] In each of the above embodiments, for example, as the hardware structure of a processing unit (Processing Unit) that executes various processes such as the RW control unit 45, the effective viewing range derivation unit 46 and 121, the control signal generation unit 47 and 122, the transmission control unit 48, the damage situation analysis unit 49, 130, 145, and 165, the screen distribution control unit 50, the building information adding units 60, 85, and 131, the building image cutting units 61, 86, and 132, the first processing unit 62, the second processing unit 87, 133, 147, and 167, and the section image cutting units 146 and 166, the following various processors (Processor) can be used. As described above, in addition to the CPU 32 which is a general-purpose processor that executes software (operation program 40) and functions as various processing units, the various processors include a programmable logic device (PLD) which is a processor whose circuit configuration can be changed after manufacturing such as an FPGA (Field Programmable Gate Array), and a dedicated electric circuit which is a processor having a circuit configuration designed specifically to execute specific processes such as an ASIC (Application Specific Integrated Circuit).

[0116] One processing unit may be composed of one of these various processors, or may be composed of a combination of two or more processors of the same type or different types (for example, a combination of multiple FPGAs, and / or a combination of a CPU and an FPGA). Also, multiple processing units may be composed of one processor.

[0117] As an example of configuring a plurality of processing units with one processor, first, as represented by computers such as clients and servers, one processor is configured by a combination of one or more CPUs and software, and this processor functions as a plurality of processing units. Second, as represented by a System On Chip (SoC) or the like, there is a form in which a processor that realizes the functions of an entire system including a plurality of processing units with one integrated circuit (IC) chip is used. Thus, various processing units are configured using one or more of the above various processors as a hardware structure.

[0118] Furthermore, as a hardware structure of these various processors, more specifically, an electric circuit (circuitry) combining circuit elements such as semiconductor elements can be used.

[0119] The technology of the present disclosure can also appropriately combine the above-described various embodiments and / or various modifications. Also, of course, various configurations can be adopted without departing from the gist, not limited to the above embodiments.

[0120] The description content and illustration content shown above are detailed descriptions of the part related to the technology of the present disclosure and are only examples of the technology of the present disclosure. For example, the descriptions regarding the above configuration, function, action, and effect are descriptions regarding an example of the configuration, function, action, and effect of the part related to the technology of the present disclosure. Therefore, it goes without saying that within the scope not departing from the gist of the technology of the present disclosure, unnecessary parts can be deleted, new elements can be added, or replacements can be made to the description content and illustration content shown above. Also, in order to avoid complication and facilitate the understanding of the part related to the technology of the present disclosure, descriptions regarding common technical knowledge that do not particularly require explanation for implementing the technology of the present disclosure are omitted from the description content and illustration content shown above.

[0121] In this specification, "A and / or B" is synonymous with "at least one of A and B". That is, "A and / or B" means that it may be only A, only B, or a combination of A and B. Also, in this specification, when expressing three or more matters connected by "and / or", the same concept as "A and / or B" is applied.

[0122] All documents, patent applications, and technical standards described in this specification are incorporated herein by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually stated to be incorporated by reference.

Explanation of Reference Signs

[0123] 2 Disaster Information Processing System 10, 10A, 10B, 10C Surveillance Camera 11, 120 Disaster Information Processing Server 12, 12A, 12B High-rise Building 13 Arrow 14 Network 15 Client Terminal 16 Display 17 Input Device 20 Region 21, 21A, 21B, 80, 102, 126 Shooting Range 22, 22A, 22B, 22C Overhead Image 22I, 22AI, 22BI Overhead Image with Building Information 30 Storage 31 Memory 32 CPU 33 Communication Unit 34 Bus Line 40 Operating Program 45 Read / Write Control Unit (RW Control Unit) 46, 121 Effective Visual Field Range Derivation Unit 47, 122 Control Signal Generation Unit 48 Transmission Control Unit 49, 130, 145, 165 Damage Situation Analysis Unit 50 Screen distribution control unit 55, 55A, 55B, 55C Effective viewing range 56, 56A, 56B, 56C Control signal 57, 154 Analysis result 58 Damage situation display screen 60, 85, 131 Building information adding unit 61, 86, 132 Building image cutting unit 62 First processing unit 63 Effective viewing range determination unit 64 Map with building information 65 Building information 66 Building image 66A First building image 66AL Learning first building image 66B Second building image 66BL Learning second building image 66L Learning building image 67, 88 Building image group 68, 134, 152, 168 Damage situation analysis model 69, 135, 153, 169 Damage situation 69CA, 135CA, 153CA, 169CA Correct damage situation 69L, 135L, 153L, 169L Learning damage situation 70 Analysis result for effective viewing range determination 75, 140, 160, 170 Learning data 78 Building 87, 133, 147, 167 Second processing unit 88A First building image group 88B Second building image group 90 Damage situation display area by building 91 Statistical damage situation display area 92 Confirmation button 110 Effective viewing range table 111 Environmental conditions 125 Smoke 146, 166 Section image cutting unit 148 Landmark building information 149 Section image 149A First Area Drawing Image 149AL Learning First Area Drawing Image 149B Second Area Drawing Image 149BL Learning Second Area Drawing Image 149L Learning Area Drawing Image 150 Area Information 151 Area Drawing Image Group 151A First Area Drawing Image Group 151B Second Area Drawing Image Group ST100, ST110, ST120, ST130, ST200, ST210 Steps

Claims

1. A processor, and a memory connected to or incorporated in the processor, and the processor acquires a valid visual field range in an overhead image of a disaster area photographed by a surveillance camera, which is a range within which the damage situation of the disaster area can be grasped and which changes according to the environmental conditions of the disaster area, and controls the operation of the surveillance camera based on the acquired valid visual field range. A disaster information processing device.

2. The processor performs at least one of setting the zoom magnification of the surveillance camera, setting the tilt angle of the surveillance camera, and setting whether to take the overhead image based on the valid visual field range. The disaster information processing device according to claim 1.

3. The processor acquires the valid visual field range from the overhead image photographed by the surveillance camera in real time. The disaster information processing device according to claim 1 or claim 2.

4. The valid visual field ranges corresponding to a plurality of patterns of the environmental conditions are stored in advance in a storage unit, and the processor acquires the valid visual field range corresponding to the current environmental conditions of the disaster area from the storage unit. The disaster information processing device according to claim 1 or claim 2.

5. There are a plurality of the surveillance cameras, and the processor controls the operations of each of the plurality of surveillance cameras based on the valid visual field range of each of the plurality of surveillance cameras. The disaster information processing device according to any one of claims 1 to 4.

6. The processor analyzes the damage situation of each building in the disaster area using the overhead image. The disaster information processing device according to any one of claims 1 to 5.

7. The processor When the building to be analyzed for the damage situation has been photographed by a plurality of the surveillance cameras, analyzes the damage situation of the building to be analyzed using the overhead images respectively photographed by the plurality of surveillance cameras. The disaster information processing device according to claim 6.

8. The processor analyzes the damage situation of each section including a plurality of adjacent buildings in the disaster area using the overhead image. The disaster information processing device according to any one of claims 1 to 5.

9. The processor The disaster information processing apparatus according to claim 8, wherein when a section to be analyzed for the damage situation is photographed by a plurality of the monitoring cameras, the damage situation of the section to be analyzed is analyzed using the bird's-eye view images respectively photographed by the plurality of the monitoring cameras.

10. Obtaining an effective visual field range in a bird's-eye view image of a disaster area photographed by a monitoring camera, the effective visual field range being capable of grasping the damage situation of the disaster area and changing according to the environmental conditions of the disaster area, and Controlling the operation of the monitoring camera based on the obtained effective visual field range. An operating method of a disaster information processing apparatus including the above.

11. Obtaining an effective visual field range in a bird's-eye view image of a disaster area photographed by a monitoring camera, the effective visual field range being capable of grasping the damage situation of the disaster area and changing according to the environmental conditions of the disaster area, and Controlling the operation of the monitoring camera based on the obtained effective visual field range. An operating program of a disaster information processing apparatus for causing a computer to execute a process including the above.

12. A monitoring camera that photographs a bird's-eye view image for grasping the damage situation of a disaster area, A processor, A memory connected to or built in the processor, and comprising: The processor: Obtains an effective visual field range that can grasp the damage situation in the bird's-eye view image and changes according to the environmental conditions of the disaster area, Controls the operation of the monitoring camera based on the obtained effective visual field range. A disaster information processing system.

Images