Damage status estimation system and damage status estimation program

Abstract

To quickly and easily grasp the damage state of a disaster area without entering the disaster area.SOLUTION: The damage status estimation system 10 includes an input unit 14 for inputting video data 12 of a disaster area, a video data acquisition unit 16 for acquiring the video data 12, a camera shake correction unit 18, a smoke region detection unit 20 configured to detect a smoke region in the video data 12, a smoke region information calculation unit 22 configured to calculate smoke region information of at least one of an area, a position, and the number of smoke regions, a damage status estimation unit 24 configured to estimate a damage status based on the smoke region information, and an output unit 26 for outputting the estimated damage status.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a system and a program for estimating the damage situation in a disaster area. [Background technology]

[0002] When a large-scale earthquake occurs, it is extremely important to quickly confirm the extent of damage in the affected area in order to establish an initial response system. However, it is often difficult to enter the affected area and confirm the extent of damage immediately after the earthquake occurs.

[0003] In response to this, a method is known in which disaster-stricken areas are photographed from the air after an earthquake occurs, and areas with damaged buildings are automatically extracted based on the aerial images (see, for example, Non-Patent Document 1). This method makes it possible to confirm the extent of damage in disaster-stricken areas after an earthquake occurs without having to enter the disaster-stricken areas. [Prior art documents] [Non-patent literature]

[0004] [Non-Patent Document 1] Hajime Mitomi, Masashi Matsuoka, Fumio Yamazaki, "An Attempt at Automatic Extraction of Building Damage Areas Using Aerial Images of Recent Earthquake Disasters," Journal of the Japan Society of Civil Engineers, April 2002, No. 703 / I-59, pp. 267-278 Summary of the Invention [Problem to be solved by the invention]

[0005] However, with the above method, it is necessary to arrange and prepare for aerial photography after the earthquake occurs, and then to dispatch a helicopter or airplane to the disaster area to take the aerial photography, making it difficult to quickly confirm the damage situation in the disaster area. The present invention has been made in consideration of the above problems, and aims to provide a system and program that can quickly and easily grasp the damage situation in a disaster area without entering the disaster area. [Means for solving the problem]

[0006] The present invention solves the above-mentioned problems by providing a damage situation estimation system that includes a video data acquisition unit for acquiring video data of the disaster area, a smoke area detection unit configured to detect smoke areas in the video data, a smoke area information calculation unit configured to calculate at least one piece of smoke area information regarding the area, position, and number of smoke areas, and a damage situation estimation unit configured to estimate the damage situation based on the smoke area information.

[0007] The inventors focused on smoke in video data as a characteristic element of a disaster. Smoke is not only generated by fires, but can also be generated by damaged structures. This damage estimation system can quantitatively estimate the damage situation in a disaster-stricken area based on at least one of smoke area information, such as the area, location, and number of smoke areas. For example, the extent of damage to structures can be estimated based on the area of ​​the smoke areas. Furthermore, the location of damaged structures can be estimated based on the location of the smoke areas. Furthermore, the number of damaged structures can be estimated based on the number of smoke areas. Furthermore, the tendency for the disaster to expand or converge can be estimated based on the change in the area of ​​the smoke areas over time. Furthermore, this damage estimation system can use, for example, news footage or footage from surveillance cameras installed in the disaster-stricken area as video data. Therefore, the damage situation in a disaster-stricken area can be estimated quickly and easily without entering the disaster-stricken area.

[0008] The smoke region detection unit may also be configured to detect flame regions in the video data, the smoke region information calculation unit may be configured to calculate at least one of flame region information, such as the area, position, and number of flame regions, and the damage situation estimation unit may be configured to estimate the damage situation based on the flame region information as well. In the case of a fire, the damage situation can be accurately grasped by estimating the damage situation based on the flame region information in addition to the smoke region information. For example, the extent of fire damage can be estimated based on the area of ​​the flame region. Furthermore, the location of the building where the fire is occurring can be estimated based on the position of the flame region. Furthermore, the number of buildings where the fire is occurring can be estimated based on the number of flame regions. Note that while it may be possible to distinguish between smoke caused by a fire and dust caused by damage to a building based on color information or duration of the smoke region, in some cases, this distinction is difficult. Even in such cases, the accuracy of the distinction can be improved by distinguishing whether the smoke is dust caused by damage to a building or smoke caused by a fire based on the positional relationship between the smoke region and the flame region.

[0009] The damage situation estimation unit may also be configured to identify which of a plurality of area ranges, which are divided by a predetermined threshold, the areas of the smoke region and the flame region belong to, and determine the level of the damage degree according to the identified area range. By the damage situation estimation unit determining the level of the damage degree as the damage situation, the damage situation can be easily shown to the user, which contributes to a quick and easy understanding of the damage situation.

[0010] The damage situation estimation unit may be configured to estimate the damage situation by assigning a weight to the area of ​​the smoke region and a weight to the area of ​​the flame region differently. In the case of a fire, the damage situation can be estimated more accurately by assigning a weight to the area of ​​the flame region that is heavier than the weight to the area of ​​the smoke region.

[0011] The smoke region information calculation unit may also be configured to calculate movement information of the smoke region and the flame region over time, and the damage situation estimation unit may be configured to estimate the damage situation based on the movement information of the smoke region and the flame region. The direction and speed of the fire spread can be estimated based on the movement information such as the amount and direction of movement of the smoke region and the flame region.

[0012] The system may further include an edge detection unit configured to detect edges in the video data, a variance calculation unit that calculates a variance value of the edges for each block that constitutes the video data, a tile crack discrimination unit that distinguishes between block units in which tile cracks have not occurred and block units in which tile cracks have occurred based on the variance value for each block, and a tile crack information calculation unit configured to calculate tile crack information based on the discrimination result of the tile crack discrimination unit, and the damage situation estimation unit may be configured to estimate the damage situation based on the tile crack information as well. The inventors have also focused on tile cracks in the video data as a characteristic element when a disaster occurs. By having the damage situation estimation unit estimate the damage situation based on the tile crack information as well, the damage situation can be estimated in more detail.

[0013] The system may further include an edge detection unit configured to detect edges in the video data and a structural damage detection unit configured to detect structural damage areas based on changes in the edges before and after a predetermined time interval, and the damage situation estimation unit may be configured to determine whether the smoke is caused by structural damage based on the positional relationship between the smoke area and the structural damage area. While it may be possible to distinguish between smoke caused by a fire and smoke caused by structural damage based on the color information and duration of the smoke area, and the presence or absence of flame areas, in some cases, this may be difficult. Even in such cases, the accuracy of the distinction can be improved by distinguishing whether the smoke is smoke caused by structural damage or smoke caused by a fire based on the positional relationship between the smoke area and the structural damage area. This allows for a more accurate understanding of the damage situation.

[0014] Furthermore, the present invention solves the above-mentioned problems by providing a damage situation estimation program that causes a computer to function as a video data acquisition unit for acquiring video data of the disaster area, a smoke area detection unit configured to detect smoke areas in the video data, a smoke area information calculation unit configured to calculate at least one piece of smoke area information regarding the area, position, and number of smoke areas, and a damage situation estimation unit configured to estimate the damage situation based on the smoke area information. [Effects of the Invention]

[0015] According to the present invention, the damage situation in a disaster-stricken area can be grasped quickly and easily without entering the disaster-stricken area. [Brief explanation of the drawings]

[0016] [Figure 1] FIG. 1 is a block diagram showing a schematic configuration of a damage situation estimation system according to a first embodiment of the present invention. [Figure 2] A flowchart showing how to use the damage estimation system [Figure 3] A diagram showing the steps by which the damage estimation system detects smoke and flame areas in video. [Figure 4] A graph showing an example of calculations of smoke area and flame area using the damage estimation system. [Figure 5] A graph showing an example of the amount of movement of the smoke area calculated by the damage estimation system. [Figure 6] FIG. 10 is a block diagram showing a schematic configuration of a damage situation estimation system according to a second embodiment of the present invention. [Figure 7] A flowchart showing how to use the damage estimation system [Figure 8] A diagram showing the steps by which the damage estimation system detects areas of broken roof tiles in an image. [Figure 9] A graph showing an example of calculation of the value equivalent to the area of ​​the cracked roof tile by the damage estimation system. [Figure 10]FIG. 10 is a block diagram showing a schematic configuration of a damage situation estimation system according to a third embodiment of the present invention. [Figure 11] A flowchart showing how to use the damage estimation system [Figure 12] A diagram showing the steps by which the damage estimation system detects damaged areas of buildings in video. [Figure 13] FIG. 10 is a block diagram showing a schematic configuration of a damage situation estimation system according to a fourth embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0017] 1 shows a damage situation estimation system 10 according to a first embodiment of the present invention. The damage situation estimation system 10 includes an input unit 14 for inputting video data 12 of a disaster area, a video data acquisition unit 16 for acquiring the video data 12, an image stabilization unit 18, a smoke region detection unit 20 configured to detect smoke regions in the video data 12, a smoke region information calculation unit 22 configured to calculate the area, position, and number of smoke regions as smoke region information, a damage situation estimation unit 24 configured to estimate the damage situation based on the smoke region information, and an output unit 26 for outputting the estimated damage situation.

[0018] The damage situation estimation system 10 is a computer system. A single computer may include the input unit 14, video data acquisition unit 16, image stabilization unit 18, smoke region detection unit 20, smoke region information calculation unit 22, damage situation estimation unit 24, and output unit 26. Alternatively, some or all of the input unit 14, video data acquisition unit 16, image stabilization unit 18, smoke region detection unit 20, smoke region information calculation unit 22, damage situation estimation unit 24, and output unit 26 may be included in different computers connected via a network. The damage situation estimation system 10 includes a computer program that causes a computer or computer system to function as the video data acquisition unit 16, image stabilization unit 18, smoke region detection unit 20, smoke region information calculation unit 22, and damage situation estimation unit 24 in order to estimate the damage situation.

[0019] The video data 12 may be, for example, footage captured by a surveillance camera installed in the disaster area or news footage. The video data 12 may also be footage captured by a drone, smartphone, or the like. Furthermore, the video data 12 is not limited to footage captured by a visible light camera, but may also be footage captured by an infrared camera such as a thermographic camera. The input unit 14 is a part of the damage situation estimation system 10 to which the video data 12 is input. The video data acquisition unit 16 is a part that acquires the video data 12 input to the input unit 14 for processing by the computer program. The video data acquisition unit 16 may also function as the input unit. The image stabilization unit 18 is configured to detect image movement in the video data 12 due to earthquake shaking or camera shake, calculate a correction value based on the detected image movement, and input the correction value into the video data 12 to perform image stabilization.

[0020] The smoke region detection unit 20 is configured to detect not only smoke regions in the video data 12 but also fire regions in the video data 12. The smoke region detection unit 20 is configured to detect smoke and fire regions in the video data 12 using a trained model obtained by machine learning such as deep learning. Note that the smoke region detection unit 20 may be configured to detect smoke and fire regions in the video data 12 based on color information, contour shape, etc. of the region without using a trained model. Furthermore, if the video data 12 is video captured by a thermographic camera, the smoke region detection unit 20 may be configured to detect smoke and fire regions in the video data 12 based on the temperature level and contour shape of the region.

[0021] In addition to calculating the smoke region information, the smoke region information calculation unit 22 is configured to also calculate the area of ​​a flame region in the video data 12 as flame region information. The smoke region information calculation unit 22 is configured to divide the smoke region or the flame region into a plurality of rectangular regions, identify the positions of the four corners of the rectangular regions, calculate the area of ​​each rectangle by multiplying the lengths of the two orthogonal sides of each rectangular region, and further calculate the area of ​​the smoke region or the flame region by accumulating the areas of the rectangles. The area of ​​the smoke region or the flame region in the video data 12 changes over time. The smoke region information calculation unit 22 repeatedly calculates the area of ​​the smoke region or the flame region in the video data 12. The values ​​of the area of ​​the smoke region or the flame region relative to the time axis may fluctuate. The smoke region information calculation unit 22 may be configured to calculate an approximation curve of the curve of the area of ​​the smoke region or the flame region relative to the time axis. The smoke region information calculation unit 22 may also be configured to provide the values ​​of the area of ​​the smoke region or the flame region that fluctuate over time in the form of a table.

[0022] The smoke region information calculation unit 22 is also configured to calculate the positions of smoke regions and flame regions. For example, the movement information of each region is calculated by assuming that the center of gravity of the smoke region or flame region is the position of each region. Alternatively, the amount of movement of each region may be calculated by assuming that the leading edge of the direction in which the smoke region or flame region moves or expands is the position of each region. The smoke region information calculation unit 22 is also configured to calculate the number of smoke regions and flame regions. The smoke region information calculation unit 22 repeatedly calculates the positions and number of smoke regions and flame regions in the video data 12. The smoke region information calculation unit 22 is also configured to calculate movement information of the smoke regions and flame regions over time. The movement information is, for example, the direction and amount of movement of the smoke regions and flame regions. The movement information is calculated based on changes in the positions of the smoke regions and flame regions over time. For example, XY coordinates along the ground surface are set, and the amount of movement of the region in the X-axis and Y-axis directions is calculated separately, and the movement direction and the amount of movement along the movement direction can be calculated based on this. The smoke region information calculation unit 22 may be configured to calculate color information and density of the smoke region as the smoke region information. The smoke region information calculation unit 22 may be configured to provide the smoke region information and flame region information in a table.

[0023] The damage situation estimation unit 24 is configured to estimate the damage situation based on not only the smoke region information but also the flame region information. Specifically, the damage situation estimation unit 24 is configured to identify which of a plurality of area ranges, defined by a predetermined threshold, the total area value of the smoke region and the flame region, belongs to, and determine the level of damage according to the identified area range. For example, the damage situation estimation unit 24 identifies which of three area ranges, defined by a predetermined threshold, the total area value of the smoke region and the flame region, belongs to, and determines the level of damage as small, medium, or large according to the identified area range. If the smoke region information calculation unit 22 is configured to calculate an approximation curve of the curve of the total area value of the smoke region or the flame region against the time axis, the value of the approximation curve may be identified as one of three area ranges, defined by a predetermined threshold, and determine the level of damage as small, medium, or large according to the identified area range. The damage situation estimation unit 24 may also be configured to estimate the damage situation by assigning a different weight to the area of ​​the smoke region and the area of ​​the flame region. In the case of a fire, the damage situation of the fire can be accurately grasped by weighting the area of ​​the flame region more heavily than the area of ​​the smoke region. The damage situation estimation unit 24 may also be configured to determine that the disaster is tending to expand when the areas of the smoke region and the flame region increase over time, and to determine that the disaster is tending to converge when the areas of the smoke region and / or the flame region decrease over time. Note that the damage level may change over time. The damage situation estimation unit 24 may determine the damage level multiple times at predetermined time intervals, or may determine the damage level only once, for example, at the end of the video.

[0024] The damage situation estimation unit 24 is also configured to estimate the damage situation based on movement information of the smoke and flame regions. Specifically, the damage situation estimation unit 24 estimates the direction and speed of fire spread based on the movement direction and amount of movement of the smoke and flame regions. The damage situation estimation unit 24 is also configured to estimate the location of collapsed or burned buildings based on location information of the smoke and flame regions. The damage situation estimation unit 24 is also configured to estimate the number of collapsed or burned buildings based on information on the number of smoke and flame regions. Note that smoke may not only be caused by fire, but may also be caused by damage to buildings. The damage situation estimation unit 24 may be configured to distinguish between smoke caused by fire and dust cloud caused by damaged buildings based on color information and duration of the smoke regions and the positional relationship between the smoke and flame regions. The output unit 26 is a part of the damage situation estimation system 10 that outputs the estimation results of the damage situation estimation unit 24.

[0025] Next, a method for using the damage situation estimation system 10 will be described with reference to the flowchart in Figure 2. First, video data 12 of the disaster area is input to the input unit 14 (S102), and the video data acquisition unit 16 acquires the video data 12 (S104). Next, the image stabilization unit 18 performs image stabilization processing for the video data 12. Specifically, the image stabilization unit 18 first detects video movement in the video data 12 due to earthquake shaking or camera shake (S106). Next, a correction value is calculated based on the video movement (S108). The correction value is then input into the video data 12 (S110). This corrects camera shake in the video data 12.

[0026] Next, as shown in Fig. 3, the smoke region detection unit 20 detects smoke regions and fire regions in the video data 12 (S112). The region surrounded by a rectangular frame and labeled "smoke" is the smoke region. The region surrounded by a rectangular frame and labeled "fire" is the fire region.

[0027] Next, the smoke region information calculation unit 22 calculates information on the smoke region and the flame region in the video data 12 (S114). Specifically, the smoke region information calculation unit 22 calculates the area of ​​the smoke region and the area of ​​the flame region in the video data 12. FIG. 4 schematically shows an example of calculation of the area of ​​the smoke region and the area of ​​the flame. As shown in FIG. 4, the area of ​​the smoke region and the area of ​​the flame region in the video data 12 change over time. The smoke region information calculation unit 22 repeatedly calculates the area of ​​the smoke region and the area of ​​the flame region in the video data 12. Furthermore, the smoke region information calculation unit 22 calculates an approximation curve TL of the curve of the sum of the area of ​​the smoke region and the area of ​​the flame region against the time axis, as shown schematically in FIG. 4. Note that FIG. 4 shows an example in which a flame region is temporarily detected, and the area of ​​the flame is reflected only in the part marked "Flame Detected." The other parts show the area of ​​the smoke region.

[0028] The smoke region information calculation unit 22 also calculates the positions and number of smoke regions and flame regions. The smoke region information calculation unit 22 also calculates movement information of the smoke regions and flame regions over time. Figure 5 shows a schematic example of how movement information of the smoke regions is calculated. The amount of movement of the smoke regions in the X direction and the Y direction over time are calculated. In the example of Figure 5, the X axis is an axis along the east-west direction (east is positive), and the Y axis is an axis along the north-south direction (north is positive). Movement information is calculated in a similar manner for the flame regions.

[0029] Next, the damage situation estimation unit 24 estimates the damage situation (S116). Specifically, the damage situation estimation unit 24 estimates the damage situation based on the sum of the areas of the smoke region and the flame region. For example, the damage situation estimation unit 24 identifies which of three area ranges, defined by a predetermined threshold, the sum of the areas of the smoke region and the flame region shown by the approximate curve TL in FIG. 4, belongs to, and determines the damage level as low, medium, or high depending on the identified area range. Note that the damage situation estimation unit 24 may estimate the damage situation by assigning a different weight to the area of ​​the smoke region and the area of ​​the flame region. For example, the area of ​​the flame region may be multiplied by a predetermined magnification to increment the area of ​​the flame region and increment the area of ​​the flame region. In the example of FIG. 4, the smoke region is continuously detected, while the flame region is temporarily detected. By incrementing the area of ​​the temporarily detected flame, the estimated damage level temporarily increases. In addition, the damage situation estimation unit 24 determines that the disaster is tending to expand if the area of ​​the smoke and / or flame region increases over time, and determines that the disaster is tending to converge if the area of ​​the smoke and / or flame region decreases.

[0030] The damage situation estimation unit 24 also estimates the damage situation based on movement information of the smoke area and the flame area. Specifically, the damage situation estimation unit 24 estimates the direction and speed of the fire spread based on movement information such as the amount and direction of movement of the smoke area and the flame area. In the example of FIG. 5, it is estimated that the smoke area will tend to spread in a northeasterly direction. The damage situation estimation unit 24 also estimates the positions of buildings that have collapsed or caught fire based on the positions of the smoke area and the flame area. The damage situation estimation unit 24 also estimates the number of buildings that have collapsed or caught fire based on the number of smoke area and the flame area. The damage situation estimation unit 24 outputs the above estimation results to the output unit 26 (S118). This completes the damage situation estimation process by the damage situation estimation system 10.

[0031] In this way, the damage situation estimation system 10 can estimate the damage situation in a disaster-stricken area based on smoke area information. It can also estimate the trends of disaster expansion and convergence based on the change in the area of ​​the smoke area over time. Furthermore, video data such as footage from surveillance cameras installed in the disaster-stricken area and news footage can be used. Therefore, the damage situation in a disaster-stricken area can be estimated quickly and easily without entering the disaster-stricken area. In the case of a fire, the damage situation can be accurately estimated by estimating the damage situation based on both smoke area information and flame area information. Furthermore, by weighting the area of ​​the flame area more heavily than the area of ​​the smoke area, the damage situation of a fire can be estimated more accurately. Furthermore, the direction and speed of fire spread can be estimated based on the movement information of the smoke area and the flame area. While it may be possible to distinguish between smoke caused by a fire and smoke caused by structural damage based on the color information and duration of the smoke, discrimination may be difficult in other cases. Even in such cases, discrimination accuracy can be improved by distinguishing whether the smoke is smoke caused by structural damage or smoke caused by a fire based on the positional relationship between the smoke area and the flame area. Furthermore, the damage situation estimation unit 24 quantitatively determines the level of damage as the damage situation, so that the damage situation can be easily shown to the user, which contributes to quick and easy estimation of the damage situation.

[0032] Next, a second embodiment of the present invention will be described. FIG. 6 shows a damage situation estimation system 30 according to the second embodiment of the present invention. Compared to the damage situation estimation system 10 according to the first embodiment, the damage situation estimation system 30 further comprises an edge detection unit 32, a variance value calculation unit 34, a roof tile cracking determination unit 36, and a roof tile cracking information calculation unit 38. The damage situation estimation unit 24 is configured to estimate the damage situation based on the roof tile cracking information as well. The damage situation estimation system 30 also includes a computer program for causing a computer or computer system to function as the edge detection unit 32, the variance value calculation unit 34, the roof tile cracking determination unit 36, and the roof tile cracking information calculation unit 38 in order to estimate the damage situation. The other components are the same as those of the damage situation estimation system 10 according to the first embodiment, and therefore, the same components are designated by the same reference numerals as those in the first embodiment, and their description will be omitted.

[0033] The edge detection unit 32 is configured to detect edges such as lines and contours in the video data 12, calculate edge strength, and narrow down the edges based on the edge strength to detect edges. The variance calculation unit 34 is configured to calculate a variance value of edges for each block constituting the image of the video data 12. The tile crack discrimination unit 36 ​​is configured to distinguish between block units in which tile cracks are not present and block units in which tile cracks are present based on the variance value for each block. The tile crack information calculation unit 38 is configured to calculate tile crack information based on the discrimination result of the tile crack discrimination unit 36. The tile crack information calculation unit 38 calculates, for example, the total number of block units in which tile cracks are present as the tile crack information. Alternatively, the tile crack information calculation unit 38 may calculate, as the tile crack information, the total area of ​​block units in which tile cracks are present. Alternatively, the tile crack information calculation unit 38 may calculate, as the tile crack information, the position of block units in which tile cracks are present.

[0034] Next, a method of using the damage situation estimation system 30 will be described with reference to the flowchart in Figure 7. Note that steps common to the method of using the damage situation estimation system 10 of the first embodiment will be assigned the same reference numerals as in the first embodiment, and descriptions thereof will be omitted. First, the edge detection unit 32 detects edges such as lines and contours in the video data 12 (S202). Next, the edge detection unit 32 calculates edge strength (S204). Furthermore, the edge detection unit 32 narrows down the edges based on the edge strength (S206). As a result, edges are detected.

[0035] Next, the variance calculation unit 34 calculates the variance of edges for each block unit constituting the image of the video data 12 (S208). Calculating the variance for each block unit allows the complexity of the image to be determined. This allows the roof tile pattern to be identified. Next, the roof tile crack determination unit 36 ​​distinguishes between block units in which roof tile cracks have occurred and block units in which roof tile cracks have occurred based on the variance for each block unit (S210). For example, if the difference in variance for each block unit before and after a predetermined time is greater than a predetermined threshold, it is determined that roof tile cracks have occurred. In FIG. 8, the area surrounded by a rectangular frame and labeled "cracked tiles" is the area in which roof tile cracks have occurred. Note that the variances of multiple block units at the same time may be compared, and a block unit with a relatively large variance value may be determined to be a block unit in which roof tile cracks have occurred. For example, a block unit in which the variance value is greater than a predetermined threshold value relative to the average value of the variances of multiple block units may be determined to be a block unit in which roof tile cracks have occurred. Next, the roof tile crack information calculation unit 38 calculates roof tile crack information based on the determination result of the roof tile crack determination unit 36 ​​(S212). For example, the roof tile crack information calculation unit 38 calculates the total number of block units in which roof tile cracks have occurred as the roof tile crack information. As shown in FIG. 9, the number of block units in which roof tile cracks have occurred in the video data 12 changes over time. The roof tile crack information calculation unit 38 repeatedly calculates the total number of block units in which roof tile cracks have occurred in the video data 12.

[0036] The damage situation estimation unit 24 estimates the damage situation based on the roof tile crack information as well (S116). For example, the damage situation estimation unit 24 quantitatively estimates the damage situation related to the damage to the building using a numerical value corresponding to the total number of roof tile cracks per block. The damage situation estimation unit 24 may also identify to which of a plurality of number ranges, which are divided by a predetermined threshold, the total number of roof tile cracks per block belongs, and determine the level of damage level of the roof tile cracks according to the identified number range. In this way, the damage situation estimation unit 24 quantitatively estimates the damage situation based on the roof tile crack information as well, thereby enabling a more detailed estimation of the damage situation related to the damage to the building.

[0037] Next, a third embodiment of the present invention will be described. FIG. 10 shows a damage situation estimation system 40 according to the third embodiment of the present invention. The damage situation estimation system 40 is configured to include a structure damage detection unit 42 instead of the variance value calculation unit 34, the roof tile crack determination unit 36, and the roof tile crack information calculation unit 38 of the damage situation estimation system 30 according to the second embodiment. The damage situation estimation system 40 includes a computer program for causing a computer or computer system to function as the structure damage detection unit 42 to estimate the damage situation. The structure damage detection unit 42 is configured to detect areas where structure damage has occurred based on changes in edges before and after a predetermined time interval. The damage situation estimation unit 24 is configured to determine whether the smoke is caused by structure damage based on the positional relationship between the smoke area and the area where structure damage has occurred. The other components are the same as those of the damage situation estimation system 30 according to the second embodiment, and therefore, the same components are designated by the same reference numerals as those of the second embodiment and will not be described again.

[0038] A method for using the damage situation estimation system 40 will be described with reference to the flowchart in FIG. 11. Steps common to the method for using the damage situation estimation system 30 of the second embodiment are assigned the same reference numerals as in the second embodiment and will not be described again. The structure damage detection unit 42 detects areas where structure damage has occurred based on changes in the edges detected by the edge detection unit 32 before and after a predetermined time interval (S302). Structure damage is determined to have occurred when the difference between the edges at the same position before and after a predetermined time interval is greater than a predetermined threshold. In FIG. 12, the area surrounded by a rectangular frame and labeled "collapsed" is the area where structure damage has occurred. While it may be possible to distinguish between smoke caused by a fire and dust caused by structure damage based on the color information of the smoke and the presence or absence of flames, this may also be difficult. Even in such cases, the accuracy of the identification can be improved by determining whether the smoke is dust caused by structure damage or smoke caused by a fire based on the positional relationship between the smoke area and the area where structure damage has occurred. This allows for more accurate estimation of the damage situation.

[0039] Next, a fourth embodiment of the present invention will be described. FIG. 13 shows a damage situation estimation system 50 according to the fourth embodiment of the present invention. The damage situation estimation system 50 is configured by omitting the smoke region detection unit 20 and the smoke region information calculation unit 22 from the damage situation estimation system 30 according to the second embodiment. The damage situation estimation unit 24 estimates the damage situation based solely on roof tile crack information, and does not estimate the damage situation based on smoke region information or flame region information. The other components are the same as those of the damage situation estimation system 30 according to the second embodiment, so the same components are denoted by the same reference numerals as in the second embodiment and will not be described again. In this way, even when the damage situation estimation unit 24 estimates the damage situation based solely on roof tile crack information, the damage situation in a disaster-stricken area can be estimated quickly and easily without entering the disaster-stricken area.

[0040] In the first to third embodiments, the damage situation estimation unit 24 identifies which of three area ranges, divided by a predetermined threshold, the total area of ​​the smoke area and the flame area shown by the approximate curve TL in Figure 4 belongs to, and determines the level of damage as small, medium, or large depending on the identified area range.However, the level of damage may also be determined based on the total area of ​​the smoke area and the flame area that is not an approximate curve.

[0041] The damage situation estimation unit 24 may also identify which of two area ranges, defined by a predetermined threshold, the areas of the smoke area and the flame area belong to, and determine the damage level as either low damage or high damage depending on the identified area range. The damage situation estimation unit 24 may also identify which of four or more area ranges, defined by a predetermined threshold, the areas of the smoke area and the flame area belong to, and determine the damage level as either low damage or high damage depending on the identified area range. The damage situation estimation unit 24 may also quantitatively indicate the damage level using a numerical value corresponding to the total area of ​​the smoke area and the flame area, rather than using a graded scale.

[0042] Furthermore, in the first to third embodiments, the damage situation estimation unit 24 estimates the damage situation based on the areas of the smoke and flame areas as well as the positions, numbers, and movement information of the smoke and flame areas. However, the damage situation may be estimated based only on the areas of the smoke and flame areas, regardless of the positions, numbers, and movement information of the smoke and flame areas. The damage situation estimation unit 24 may estimate the damage situation based only on the positions and / or numbers of the smoke and flame areas, regardless of the areas of the smoke and flame areas. Furthermore, in the first to third embodiments, the smoke area information calculation unit 22 calculates three pieces of smoke area information, namely the area, positions, and numbers of the smoke areas. However, the smoke area information calculation unit may be configured to calculate only one or two pieces of smoke area information, namely the area, positions, and numbers of the smoke areas. Furthermore, in the first to third embodiments, the smoke region information calculation unit 22 calculates three pieces of flame region information: area, position, and number of flame regions, but the smoke region information calculation unit may be configured to calculate only one or two of the flame region information: area, position, and number of flame regions. Furthermore, the damage situation estimation unit 24 may estimate the damage situation based only on the smoke region information, regardless of the flame region information. In this case, the smoke region information calculation unit may be configured to calculate only the smoke region information, without calculating the flame region information. [Industrial Applicability]

[0043] The present invention can be used to estimate the damage situation in a disaster-stricken area. [Explanation of symbols]

[0044] 10, 30, 40, 50 Damage Situation Estimation System 12 Video data 14 Input section 16 Video data acquisition unit 18 Image stabilization unit 20 Smoke area detection unit 22 Smoke area information calculation unit 24 Damage Estimation Department 26 Output section 32 Edge detection unit 34 Variance calculation unit 36 Roof tile crack detection unit 38 Roof tile crack information calculation unit 42 Building damage detection unit

Claims

1. a video data acquisition unit for acquiring video data of the disaster area; a smoke region detector configured to detect smoke regions in the video data; a smoke region information calculation unit configured to calculate at least one of smoke region information of the area, position, and number of the smoke regions; a damage situation estimation unit configured to estimate a damage situation based on the smoke region information; A damage situation estimation system equipped with

2. In claim 1, the smoke region detection unit is configured to also detect flame regions in the video data; The smoke region information calculation unit is configured to also calculate at least one of flame region information of an area, a position, and a number of the flame regions; The damage situation estimation system is configured so that the damage situation estimation unit estimates the damage situation based on the flame region information as well.

3. In claim 2, The damage situation estimation unit is a damage situation estimation system configured to identify which of a plurality of area ranges, which are divided by a predetermined threshold, the areas of the smoke area and the flame area belong to, and to determine the level of damage according to the identified area range.

4. In claim 2 or 3, The damage situation estimation system is configured so that the damage situation estimation unit estimates the damage situation by assigning a different weight to the area of ​​the smoke region and the area of ​​the flame region.

5. In claim 2, the smoke region information calculation unit is configured to also calculate movement information of the smoke region and the flame region over time; A damage situation estimation system configured such that the damage situation estimation unit estimates the damage situation based on movement information of the smoke area and the flame area.

6. In claim 1 or 2, an edge detector configured to detect edges in the video data; a variance value calculation unit that calculates a variance value of the edge for each block constituting an image of the image data; a tile crack discrimination unit that discriminates between block units in which tile cracks have occurred and block units in which tile cracks have not occurred based on the variance value for each block unit; A roof tile crack information calculation unit configured to calculate roof tile crack information based on the determination result of the roof tile crack determination unit, The damage situation estimation system is configured so that the damage situation estimation unit estimates the damage situation based on the roof tile crack information as well.

7. In claim 1, an edge detector configured to detect edges in the video data; a structural damage detection unit configured to detect a structural damage area based on a change in the edge before and after a predetermined time interval; The damage situation estimation system is configured so that the damage situation estimation unit determines whether the smoke is caused by the structural damage based on the positional relationship between the area of ​​smoke and the area where the structural damage has occurred.

8. Computers are used to estimate the extent of damage. a video data acquisition unit for acquiring video data of the disaster area; a smoke region detector configured to detect smoke regions in the video data; a smoke region information calculation unit configured to calculate at least one of smoke region information of the area, position, and number of the smoke regions; a damage situation estimation unit configured to estimate a damage situation based on the smoke region information; A damage estimation program to make it function as a system.

Images