Ground Object Detection System, Ground Object Detection Method, and Program

Abstract

To perform a preparation work of a report having a size of a ground object and the number of ground objects written with simplicity and in a short time, and as a result, to allow for an inexpensive preparation work of the report.SOLUTION: A ground object detection device 1 has: an image coupling unit 19 that generates an overhead-view image in a state overhead-viewed from a sky; a machine learning unit 23 that detects a location of a stone gravel shot in the overhead-view image; a map / image display unit 31 that generates a display control signal for displaying a screen displaying the location of the stone gravel detected by the machine learning unit 23 overlapped on the overhead-view image; a processing request reception unit 11 that receives a setting input of an interest area with respect to the overhead-view image having the location of the stone gravel overlapped; and a spatial filter unit 25 that performs spatial filtering processing to the overhead-view image having the stone gravel overlapped on the basis of the setting input of the interest area.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a ground object detection system, a ground object detection method, and a program.

Background Art

[0002] For disaster prevention purposes, for example, a report has been created that detects ground objects such as riverbed gravel and records their size and number. In particular, the importance of detecting meter-order rocks that may cause damage downstream when washed away by floods is extremely high.

[0003] Generally, for gravel detection, first, a photo of the riverbed is taken by aerial photography, and the gravel captured in this photo is visually extracted and identified by an inspector. For the gravel determined to be dangerous by the inspector, its size is specified by eye measurement, and its number is counted to create a report. Further, when it is necessary to specify the location of the gravel, an inspector goes to the riverbed to survey the location of the gravel.

[0004] Patent Document 1 discloses a technique capable of accurately and efficiently collecting aerial survey images that can be used for generating high-precision three-dimensional images using an unmanned aircraft. Patent Document 2 also discloses a technique that accurately measures the subject distance and accurately projects ranging data and scale data based on the ranging data onto a photo screen even when the subject is far away and the distance to the subject is large.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0006] However, the method of creating a report that records the size and number of the above-mentioned gravel is manual, which requires time and effort, and the cost until the report is created is also high.

[0007] In addition, in the case of the work of creating a report manually, there is a possibility that the judgment criteria such as the determination of the size of the gravel may vary depending on the operator.

[0008] In addition, it also takes time to train the operator who creates the report.

[0009] The present invention has been made in view of the above problems, and provides a ground object detection system, a ground object detection method, and a program that can simplify and shorten the work of creating a report that records the size and number of ground objects, and as a result, can perform the report creation work at low cost.

Means for Solving the Problems

[0010] To solve the above problems, a ground object detection system according to one aspect of the present invention includes an image combining unit that combines a plurality of captured images of the ground including the ground object to generate an aerial view image in a state of overlooking from above, a machine learning unit that detects the position of the ground object captured in the aerial view image, a map / image display unit that generates a display control signal for displaying a screen on which the position of the ground object detected by the machine learning unit is superimposed on the aerial view image, a processing request receiving unit that receives a setting input of a region of interest for the aerial view image on which the position of the ground object is superimposed, a spatial filter unit that performs a spatial filtering process on the aerial view image on which the position of the ground object is superimposed based on the setting input of the region of interest, a land cover classification unit that estimates the land cover of the aerial view image and classifies the aerial view image into regions having the same land cover, a land cover analysis unit that classifies and analyzes the land cover and regions within the region of interest, and a spatial filter inference unit that estimates a region of interest for the aerial view image on which the position of the ground object for which the setting input has not yet been received is superimposed based on the analysis result of the land cover analysis unit, and performs a spatial filtering process on the aerial view image.

Effects of the Invention

[0011] According to the present invention, it is possible to create a report describing the size and number of ground objects simply and in a short time, and as a result, it is possible to create the report at low cost.

Brief Description of the Drawings

[0012]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Modes for Carrying Out the Invention

[0013] Hereinafter, embodiments of the present invention will be described with reference to the drawings. Note that the embodiments described below do not limit the invention according to the claims, and not all of the elements and combinations thereof described in the embodiments are essential for the solution means of the invention.

[0014] The ground object detection system of this embodiment is a system that detects ground objects such as gravel on a riverbed using images, measures the size and number of these ground objects, and creates a report. The ground object detection system of this embodiment is applied to a system for detecting gravel on a riverbed, but the ground objects detected by the ground object detection system are not limited to gravel.

[0015] <Overall Configuration of the System> FIG. 1 is a diagram showing the overall configuration of a ground object detection system according to an embodiment of the present invention.

[0016] As shown in FIG. 1, the ground object detection system 1 has a data processing server 3, and the data processing server 3 can be composed of one or more server computer machines. The data processing server 3 provides a service to assist each of one or more users in inspecting an object specified by each user using an image and in summarizing the inspection results in a report.

[0017] The data processing server 3 can perform data communication with a web browser of an information processing terminal (hereinafter referred to as a user terminal) 7 such as a personal computer or a smartphone used by each user via a web server 5. In FIG. 1, for the sake of convenience, only one user terminal 7 is shown. Hereinafter, the case where one user uses the data processing server 3 will be taken as an example to explain the details of this system.

[0018] The data processing server 3 has processing components such as a processing request reception unit 11, a data storage / search unit 13, and an analysis / calculation unit 15.

[0019] The processing request reception unit 11 receives various processing requests from the user (for example, various processing requests for performing inspections, or various processing requests for creating reports, etc.), and passes the requests to the processing components that perform the requested processing.

[0020] The data storage and retrieval unit 13 stores data such as maps, images, inspection results, and reports of various inspection targets, and reads out the requested map, image, or inspection result data in response to data requests from other processing components and passes it to that processing component. Also, programs such as the firmware executed when the data processing server 3 is started and the operating system that controls the basic operations of the data processing server 3 are stored in the data storage and retrieval unit 13.

[0021] The analysis and calculation unit 15 controls the entire data processing server 3. Also, when the firmware stored in the data storage and retrieval unit 13 is executed by the analysis and calculation unit 15 when the data processing server 3 is started, the analysis and calculation unit 15 executes functions as the following-described processing components.

[0022] The analysis and calculation unit 15 includes processing components such as an image registration unit 17, an image combining unit 19, a three-dimensional data generation unit 21, a machine learning unit 23, a spatial filter unit 25, a major axis / minor axis calculation unit 27, a histogram generation unit 29, a map / image display unit 31, a report generation unit 33, a report display unit 35, a report output unit 37, a land cover classification unit 39, a land cover analysis unit 41, a spatial filter inference unit 43, a report format registration unit 45, and a report format analysis unit 47.

[0023] The image registration unit 17 accepts captured images (which may include not only general visible light photographic images, but also photographic images using non-visible light such as infrared rays and other types of images) of the ground including ground objects (in this embodiment, gravel) that are the detection targets from the user terminal 7, and registers those images in the data storage and retrieval unit 13 in association with the geographical information (latitude, longitude, etc.) of the images and the user.

[0024] Normally, there are many ground objects (stones and gravel) that are objects to be detected on the ground. Therefore, a large number of captured images are registered for the detection target space such as a riverbed. For example, when a camera mounted on an unmanned aerial vehicle takes a photo of a vast detection target space such as a riverbed, a large number of high-quality captured images that each capture a part of the riverbed are captured by the camera. The camera associates and stores the geographical information at the time of capturing a part of the riverbed with the captured image.

[0025] Therefore, the image registration unit 17 registers each captured image in association with the geographical information at the time of capturing. At this time, the image registration unit 17 may register, as the geographical information representing the riverbed that is the detection target space, any one of the geographical information associated with each captured image.

[0026] The image combining unit 19 generates a mosaic image of the entire detection target space by combining the registered large number of captured images.

[0027] When the input data (image) used by the image combining unit 19 to generate the mosaic image satisfies a predetermined condition, the 3D data generation unit 21 performs SfM (Structure from motion) processing on this input data to generate 3D data of the captured detection target space, and generates an orthoimage from this 3D data. This orthoimage is an aerial view image of the detection target space (riverbed) from above. An orthoimage is an image that is orthogonally projected by correcting distortions due to the tilt and scale ratio of the camera, etc. based on photogrammetry techniques. Since the method of generating 3D data by SfM and SfM and generating an orthoimage based on this 3D data is well-known, the description here is omitted.

[0028] The orthoimage is generated to ensure the ground object detection accuracy (especially the detection accuracy of the size and its position) described later. Hereinafter, an inspection of the ground object (stones and gravel) that is the object to be detected existing in the detection target space is performed based on this orthoimage (aerial view image). The 3D data and the orthoimage are stored in the data storage and search unit 13.

[0029] The machine learning unit 23 detects the positions of the gravels photographed in the ortho-image. More specifically, the machine learning unit 23 uses a machine learning model to detect the positions where the gravels are photographed from the ortho-image. Since the method for generating the machine learning model is well-known, the description thereof is omitted here.

[0030] The spatial filter unit 25 performs spatial filtering processing on the ortho-image generated by the three-dimensional data generation unit 21 based on the region of interest set and input by the user via the user terminal 7. The region of interest is a region where the major axis / minor axis calculation unit 27 and the histogram generation unit 29 described later perform the calculation process of the major axis and minor axis of the gravel and the calculation process of the number of gravels, and is set in at least a part of the ortho-image. The spatial filtering processing means setting a region of interest for the ortho-image and limiting the region where the calculation process by the major axis / minor axis calculation unit 27 etc. should be performed.

[0031] The major axis / minor axis calculation unit 27 calculates the major axis and minor axis of the gravels detected by the machine learning unit 23 based on the output result of the machine learning unit 23, that is, the gravel detection result, from the ortho-image within the region of interest set by the spatial filter unit 25.

[0032] The histogram generation unit 29 generates a histogram of the gravel sizes and a cumulative histogram of the gravel sizes based on the calculation results of the major axis and minor axis of the gravels calculated by the major axis / minor axis calculation unit 27.

[0033] The map / image display unit 31 sends the ortho-image etc. displayed on the screen of the user terminal 7 to the web server 5 in response to a processing request from the user. The web server 5 sends a work screen incorporating the ortho-image to the user terminal 7, and the user terminal 7 displays this work screen on the screen of the web browser.

[0034] In particular, the map / image display unit 31 can generate a display control signal for displaying a screen that superimposes and displays the positions of the gravels detected by the machine learning unit 23 on the ortho-image.

[0035] In response to a request from the user, the report generation unit 33 creates an investigation report on the gravel in the region of interest using data such as ortho-images, the position detection results of gravel, the histogram of the gravel size, and the cumulative histogram of the gravel size. At this time, the report generation unit 33 accepts the comment input by the user via the user terminal 7 through the processing request reception unit 11, and creates an investigation report including this comment.

[0036] In response to a request from the user, the report display unit 35 sends the investigation report created by the report generation unit 33 to the Web server 5. The Web server 5 sends the investigation report to the user terminal 7, and the user terminal 7 displays this work screen on the screen of the web browser.

[0037] In response to a request from the user, the report output unit 37 sends the investigation report created by the report generation unit 33 to the Web server 5. The Web server 5 enables the user to download the investigation report to the user terminal 7.

[0038] The land cover classification unit 39 infers the land cover of each region or each pixel in the ortho-image from the spectral radiation characteristics of each pixel of the ortho-image and the texture information of the sub-pixels. The processing itself of the land cover classification unit 39 is known, and for example, the technology disclosed in Japanese Patent Laid-Open No. 8-320930 can be applied.

[0039] The land cover analysis unit 41 examines the correlation between the spatial filter, which is the region of interest set and input by the user, and the land cover classification result classified by the land cover classification unit 39. As an example, the land cover analysis unit 41 extracts the estimation results of the land cover of each region or each pixel in the ortho-image within the spatial filter.

[0040] When creating an investigation report for the detection target space in the ground object detection system 1 that has not yet been analyzed, the spatial filter inference unit 43 sets the spatial filter that is considered to be optimal based on the land cover analysis result by the land cover analysis unit 41.

[0041] The report form registration unit 45 registers the form (sample) of the survey report that the user wishes to create and has input via the user terminal 7. The registered survey report form is stored in the data storage and search unit 13.

[0042] The report form analysis unit 47 analyzes the survey report form registered by the report form registration unit 45 and stored in the data storage and search unit 13, classifies the form into a character part, a graph part, a photo part, etc., and generates a template for a survey report having approximately the same ratio as this classification result.

[0043] <System operation> FIG. 2 shows the basic communication flow between the data processing server 3 and the user in the ground object detection system 1.

[0044] In FIG. 2, at step 51, the user 9 uploads a photographed image that serves as the basis for creating a survey report to the data processing server 3 via the user terminal 7. At step 53, the image registration unit 17 of the data processing server 3 registers the uploaded photographed image and stores it in the data storage and search unit 13.

[0045] At step 55, the image combining unit 19 combines the photographed images registered at step 53 to generate a mosaic image of the detection target space. Thereafter, the 3D data generation unit 21 performs SfM processing on the mosaic image generated by the image combining unit 19 to generate 3D data and an orthoimage. FIG. 3 shows an example of the orthoimage P generated by the 3D data generation unit 21.

[0046] At step 57, based on the orthoimage generated by the 3D data generation unit 21, the machine learning unit 23 detects the positions of the gravels photographed in the orthoimage. Then, at step 59, the map and image display unit 31 displays on the user terminal 7 a screen with the positions of the gravels detected in the orthoimage superimposed.

[0047] FIG. 4 shows an example of a screen display of the detection result of gravel on the user terminal 7. The position of the gravel T detected by the machine learning unit 23 is displayed as a bar B superimposed on the ortho-image P. The length of the bar B is proportional to the approximate size of the gravel T detected by the machine learning unit 23.

[0048] FIG. 5 shows the procedure for the machine learning unit 23 to generate a machine learning model. The model generator (teacher) designates the position where gravel is photographed in the ortho-image P by a rectangular region R. The machine learning unit 23 generates a machine learning model based on the feature amount of the rectangular region R.

[0049] In step 61, the user 9 makes a setting input of the region of interest for calculating the position and size of the gravel while looking at the screen displayed on the user terminal 7. The setting input of the region of interest in step 61 is performed using an input device such as a mouse provided on the user terminal 7, for example.

[0050] FIG. 6 shows an example of the setting input of the region of interest performed in step 61. First, the user 9 designates a length 81 of n meters (200 m as an example) upstream and downstream from the planned construction site 80 in the riverbed which is the detection target space. Next, the user 9 sets the width of the region of interest. At this time, regions where gravel counting is not performed, such as the road along the river, are excluded and designated. Further, the user 9 designates an area 82 where gravel counting is not performed (exclusion area), such as a group of concrete blocks. Thereby, the setting input of the region of interest 83 is performed. In addition, the user 9 adds a note of the area 84 where gravel for which detection omission by the machine learning unit 23 occurred in the region of interest 83 is photographed.

[0051] In step 63, the spatial filter unit 25 performs spatial filtering processing on the ortho-image based on the region of interest 83 set and input by the user. In step 65, the major axis / minor axis calculation unit 27 counts the major axis and minor axis of the gravel in the region of interest 83, and further counts the number of gravel.

[0052] The user 9 can modify the region of interest 83 set and input in step 61 in step S66. Hereinafter, with reference to FIGS. 7 to 10, a method for modifying the region of interest 83 will be described.

[0053] As shown in FIG. 7, the user 9 can draw a line segment, which is the planned construction site 80, on the aerial image so as to cross the river. (This line segment can simulate the planned position of the construction of the check dam.) Then, as the region of interest 83, two rectangular regions with a specified size of length 81 (for example, 200 m in length and the width being the length of the line segment) having the line segment as one side are automatically set on the upstream side and the downstream side of the line segment. These two rectangular regions are set as the region of interest 83.

[0054] Then, in step S65, the number and the percentage of the number of gravels of each of a plurality of size levels existing in the region of interest 83 are automatically calculated, and the aggregation result is displayed on the screen (see the lower right table in FIG. 7). In this aggregation result, a color mark unique to the size level of the gravel is displayed.

[0055] Furthermore, on the aerial image as well, each of the gravels existing in the region of interest 83 is displayed with a unique color of the size level to which it belongs. The user 9 can visually grasp how the gravels of which size levels are distributed in the region of interest.

[0056] By looking at the aggregation result, the planned position of the check dam construction on the aerial image, the distribution of the gravels by size in the region of interest 83, etc., the user 9 can easily judge the suitability of the planned position of the check dam construction.

[0057] The user can change the region of interest 83 as follows.

[0058] For example, as shown in FIG. 8, while fixing the planned position of the check dam construction (line segment 80), the directions of the regions of interest 83 on the upstream side and the downstream side can be individually changed (the rectangular regions change to parallelogram regions). It can cope with the bend of the river.

[0059] Also, as shown in Fig. 9, while fixing the planned position of the sediment control dam (line segment 80), the length 81 of the region of interest can be individually changed in the directions of the regions of interest on the upstream and downstream sides thereof. This makes it easier to handle cases where a sediment control dam is present nearby or where the width of a river or riverbed is changing.

[0060] Furthermore, as shown in Figs. 8 and 9, by designating an exclusion region 82 on the aerial image, the exclusion region 82 can be excluded from the region of interest 83.

[0061] Furthermore, although not shown in the drawings, the length and position of the planned dam construction position (line segment 80) can also be changed.

[0062] Furthermore, as shown in Fig. 10, when a region that does not correspond to gravel is automatically recognized and displayed as gravel on the aerial image, the user 9 can specify the misrecognized display (for example, the surrounding frame 84) and perform a correction to exclude it from the gravel. Conversely, although not shown in the drawings, when a region corresponding to gravel has been automatically missed in recognition, the region can be specified (for example, by surrounding it with a frame line) and a correction to add it to the gravel can be performed.

[0063] When the user 9 makes the above-described changes to the region of interest 83 or corrects misrecognition, aggregation is performed again each time in steps 63 and 65, and the display of the aggregation result, the display of the region of interest, and the display of the gravel in the region of interest are updated. Then, color-coded display according to the size of the gravel is performed, the number and percentage of gravel for each size level of the gravel existing in the region of interest 83 are calculated, and are displayed on the screen.

[0064] After the region of interest 83 is updated, the histogram generation unit 29 generates a histogram of the gravel size and a cumulative histogram of the gravel size based on the calculation results of the major axis and minor axis of the gravel calculated by the major axis / minor axis calculation unit 27.

[0065] FIG. 7 shows an example of a histogram generated by the histogram generation unit 29. FIG. 11 is an example of a screen displayed on the user terminal 7 based on an instruction from the user 9 after the histogram is generated by the histogram generation unit 29.

[0066] On this screen, a histogram 90 of the gravel size and a cumulative histogram 91 of the gravel size are displayed. Also, on the screen, there are provided a region 92 for displaying the number of gravels detected by the machine learning unit 23, a region 93 for displaying the average diameter of the gravels calculated by the major axis / minor axis calculation unit 27, a region 94 for displaying the diameter (D50) of the gravels corresponding to 50% of the total number of detected gravels, and a region 95 for displaying the diameter (D95) of the gravels corresponding to 95% of the total number of detected gravels.

[0067] In step 67, the report generation unit 33 creates an investigation report. FIG. 12 shows an example of the investigation report created by the report generation unit 33. The investigation report has a character part 100 in which bibliographic information and detection results are described, a photo part 101 in which an image with the detection results of gravels superimposed on the ortho image is displayed, and a graph part 102 in which two types of histograms are displayed.

[0068] In step 69, the report display unit 35 displays the investigation report on the screen of the user terminal 7. In step 71, the user 9 edits the investigation report by inputting comments on the investigation report via the user terminal 7. The comments input by the user 9 are added to the character part 100.

[0069] In step 73, the report generation unit 33 stores the investigation report in the data storage / search unit 13. In step 75, the report output unit 37 sends the investigation report to the Web server 5 in a downloadable state, and the user 9 downloads the investigation report.

[0070] <Effect of the System> According to this embodiment configured as described above, based on the captured image, the position of the gravel is detected, the size and number of the gravels captured in the set and input region of interest are counted, and further, a histogram of the gravel sizes is generated, and an investigation report can be created based on these.

[0071] Therefore, according to this embodiment, it is possible to perform the work of creating a report describing the size and number of gravels simply and in a short time, and as a result, it is possible to perform the report creation work at low cost.

[0072] <Other functions of the system> In addition to the functions and operations described above, the ground object detection system 1 of this embodiment has the following functions and operations.

[0073] First, since geographical information is associated with the captured image, geographical information can also be associated with the region of interest. When there is a setting input of the region of interest, the spatial filter unit 25 stores the region of interest subjected to the spatial filtering process and the geographical information associated with this region of interest in the data storage and search unit 13 when performing the spatial filtering process corresponding to this region of interest.

[0074] Then, when an aerial image that can be determined to be the same (a large number of captured images with overlapping geographical information) is input, the spatial filter unit 25 searches the inside of the data storage and search unit 13 using the geographical information of the aerial image as a search key, and reads out information regarding the corresponding region of interest. After that, the spatial filter unit 25 performs the spatial filtering process based on the region of interest that has already been stored without waiting for the setting input of the region of interest.

[0075] Also, the land cover classification unit 39 estimates the land cover of each region or each pixel in the orthoimage, and the land cover analysis unit 41 examines the correlation between the region of interest and the land cover classification result. Then, the land cover analysis unit 41 learns the characteristics of the land cover classification in the spatial filter.

[0076] Furthermore, when the captured image registered by the image registration unit 17 is a captured image that is not already associated with the geographical information stored in the data storage and search unit 13, the spatial filter inference unit 43 infers the region of interest that the user 9 will input from the land cover classification result by the land cover classification unit 39 based on the learning result of the land cover analysis unit 41. In addition, the machine learning unit 23 presents the detection result of gravel in the spatial filter inferred by the spatial filter inference unit 43 and the detection result of gravel in the entire orthoimage.

[0077] By such an operation, for the detection target space for which an investigation report has been created once, the setting input of the region of interest can be omitted, and the investigation report creation work can be performed more quickly.

[0078] Also, if the investigation report is repeatedly created, the spatial filter can be inferred based on the learning result of the land cover analysis unit 41, and similarly, the setting input of the region of interest can be omitted, and the investigation report creation work can be performed more quickly. In particular, since the captured image is also associated with the user, if the learning result of the land cover analysis unit 41 is associated with the user, the preferences and habits of the user's region of interest setting can be learned, and the optimal region of interest can be inferred for each user.

[0079] Then, the report format analysis unit 47 analyzes the investigation report registered by the report format registration unit 45, classifies the inside of the format into a character part 100, a graph part 102, a photo part 101, etc., and creates a template of the investigation report. Therefore, the report generation unit 33 can create an investigation report according to the template created by the report format analysis unit 47.

[0080] Accordingly, even when creating an investigation report based on a predefined format of an investigation report, it is possible to create an investigation report in a format compliant with the predefined format without taking the trouble to perform coding based on the predefined format or specify a format based on a GUI (Graphic User Interface). Moreover, if different investigation report formats are registered by the report format registration unit 45, it is possible to easily create an investigation report based on the registered format. Modification example

[0081] Note that the present invention is not limited to the above-described embodiments and includes various modification examples. For example, the above-described embodiments have been described in detail for easy understanding of the present invention and are not necessarily limited to those having all the configurations described. Also, a part of the configuration of one embodiment can be replaced with the configuration of another embodiment, and the configuration of another embodiment can be added to the configuration of one embodiment. Further, for a part of the configuration of each embodiment, addition, deletion, or replacement with other configurations is possible.

[0082] Also, each of the above configurations, functions, processing units, processing means, etc. may be realized in hardware, for example, by designing a part or all of them with an integrated circuit. Further, each of the above configurations, functions, etc. may be realized in software by a processor interpreting and executing a program for realizing each function. Information such as a program, table, file, etc. for realizing each function can be placed in a memory, a recording device such as a hard disk, SSD, or a recording medium such as an IC card, SD card, DVD.

[0083] Also, control lines and information lines show those considered necessary for explanation and do not necessarily show all control lines and information lines on the product. In reality, it may be considered that almost all configurations are interconnected.

Explanation of reference numerals

[0084] 1…Ground object detection system 3…Data processing server 5…Web server 7…User terminal 9…User 11…Processing request reception unit 13…Data storage / search unit 15…Analysis / calculation unit 17…Image registration unit 19…Image combination unit 21…3D data generation unit 23…Machine learning unit 25…Spatial filter unit 27…Major / minor axis calculation unit 29…Histogram generation unit 31…Image display unit 33…Report generation unit 35…Report display unit 37…Report output unit 39…Land cover classification unit 41…Land cover analysis unit 43…Spatial filter inference unit 45…Report format registration unit 47…Report format analysis unit 83…Region of interest 90…Histogram 91…Cumulative histogram 100…Character part 101…Photo part 102…Graph part P…Orthophoto image T…Gravel

Claims

1. A 3D data generation unit that generates an ortho-image of a detection target space including the ground based on a plurality of captured images of the ground including ground objects; A position detection unit that detects the position of the ground object based on the ortho-image; A display unit that superimposes and displays the position of the ground object on the ortho-image; An interest area setting unit that sets an interest area on the ortho-image; A spatial filter unit that performs spatial filtering processing on the ortho-image based on the interest area; An output unit that counts the number of ground objects within the interest area of the ortho-image subjected to the filtering process and outputs the counting result; An exclusion area setting unit that sets an exclusion area within the interest area where the counting of the ground objects is not performed A ground object detection system having the above.

2. A 3D data generation unit that generates an ortho-image of a detection target space including the ground based on a plurality of captured images of the ground including ground objects; A position detection unit that detects the position of the ground object based on the ortho-image; A display unit that superimposes and displays the position of the ground object on the ortho-image; An interest area setting unit that sets an interest area on the ortho-image; Means for adding, based on human input, an area where a ground object that has been undetected by the position detection unit within the interest area is captured; A spatial filter unit that performs spatial filtering processing on the ortho-image based on the interest area; An output unit that counts the number of ground objects within the interest area of the ortho-image subjected to the filtering process and outputs the counting result A ground object detection system having the above.

3. A 3D data generation unit that generates an ortho-image of a detection target space including the ground based on a plurality of captured images of the ground including ground objects; A position detection unit that detects the position of the ground object based on the ortho-image; A display unit that superimposes and displays the position of the ground object on the ortho-image; An interest area setting unit that sets an interest area on the ortho-image; A spatial filter unit that performs spatial filtering processing on the ortho-image based on the interest area; A land cover classification unit that estimates the land cover of the ortho-image and classifies the ortho-image into areas having the same land cover; A land cover analysis unit that classifies and analyzes the land cover within the interest area and the area Based on the analysis result of the land cover analysis unit, estimate the region of interest for the ortho-image on which the positions of the ground objects that have not yet accepted the setting input of the region of interest for the ortho-image are superimposed, and a spatial filter inference unit that performs spatial filtering processing on the ortho-image, a counting unit that counts the number of the ground objects within the region of interest of the ortho-image on which the filtering process has been performed, and outputs the counting result A ground object detection system having the above components.

Images