Construction site detection device and detection program

Abstract

To provide a detector and a detecting program for construction sites capable of improving business efficiency by automatically identifying construction sites and capable of grasping the construction sites by visualizing the construction sites in a target area on a map.SOLUTION: A construction site detector acquires observation images on the ground including construction sites from above through an image acquisition device 100 (an earth observation satellite 101, a flying object 102, and so on), and comprises a learning unit 201 to learn construction site identification information by machine learning for identifying the construction sites based on the observation images and a construction site detection unit 202 to detect the construction sites from the observation images based on the construction site identification information learned by the learning unit 201. The construction site detection unit 202 may include a temporary enclosure detection unit 203 to detect the temporary enclosures of the construction site from the observation images.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a detection device and a detection program for a construction site such as a building construction site or a demolition construction site. [Background technology]

[0002] In the real estate industry, construction industry, and the like, there are many cases where a construction site, a demolition site, or a vacant lot is identified, and sales activities and other business activities are carried out based on this identification information. Conventionally, the task of identifying construction sites, demolition sites, etc. has relied on so-called human wave tactics, in which a person in charge patrols the target area and visually inspects it.

[0003] Therefore, the task of identifying construction sites, etc., required a great deal of effort and time. Furthermore, because the task of identifying construction sites, etc., relied on human patrols and visual inspections, construction sites were easily overlooked or discovered late, resulting in missed business opportunities.

[0004] In addition, information obtained solely from construction signs and other sources can result in significant discrepancies between the planned construction period and the actual construction period, making it impossible to conduct business at the appropriate time depending on the industry. Furthermore, it is difficult in reality to keep track of properties that have become vacant on a continuous and timely basis.

[0005] Furthermore, there have been cases where people have purchased or moved into real estate without knowing that new construction or demolition work is planned in the area they are considering purchasing or in the area where they are moving, resulting in unexpected differences in the surrounding environment they had anticipated. Incidentally, Patent Document 1 proposes a technology for detecting changes in a building, such as new construction, reconstruction, expansion, or destruction of a building. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Patent Publication No. 2021-33584 Summary of the Invention [Problem to be solved by the invention]

[0007] Here, Patent Document 1 discloses a building change identification device and program that identifies differences for each unit area of ​​images taken from the sky at different times, and specifies changes in buildings.

[0008] If the technology of Patent Document 1 can detect changes due to new construction, reconstruction, expansion, or destruction of a building, it may be possible to detect construction sites for those changes. However, by the time changes to the building can be identified, several months or more have often passed since the start of construction or demolition work, making it difficult to collect information in a timely manner.

[0009] Therefore, the present invention has been made in consideration of the above-mentioned problems, and aims to provide a construction site detection device and detection program that can improve work efficiency by automatically identifying construction sites and enable construction sites in a target area to be visualized and understood on a map. [Means for solving the problem]

[0010] A construction site detection device according to one embodiment of the present invention comprises a learning unit that uses machine learning to learn construction site identification information for identifying a ground-based construction site contained in an observation image taken from the air, and a construction site detection unit that detects the construction site from the observation image based on the construction site identification information learned by the learning unit.

[0011] In addition, the observation image includes an image of a temporary fence used at an on-ground construction site, and the learning unit learns temporary fence identification information for identifying the temporary fence based on the observation image as the construction site identification information through machine learning, and the construction site detection unit can detect the temporary fence as the construction site from the observation image based on the temporary fence identification information learned by the learning unit.

[0012] The observation image may be an image taken from the sky in a direction inclined at a predetermined angle from a direction perpendicular to the ground.

[0013] The observation image may be an image taken by an imaging device mounted on a satellite or aircraft moving in the sky.

[0014] In addition, the satellite may be configured as a synthetic aperture radar (SAR) satellite, and the construction site detection unit may detect the construction site based on the observation images taken at multiple times by the imaging device of the SAR satellite.

[0015] The system may also be provided with a marking unit that marks the position of the construction site detected by the construction site detection unit, which is identified based on ground position information captured in the observation image acquired from the satellite or the aircraft, on a screen that displays map data including the construction site detected by the construction site detection unit.

[0016] In addition, the observation image may include an image of a temporary fence used at an on-ground construction site, and the learning unit may use machine learning to learn temporary fence identification information for identifying the temporary fence based on the observation image as the construction site identification information, the construction site detection unit may detect the temporary fence as the construction site from the observation image based on the temporary fence identification information learned by the learning unit, and the marking unit may mark and display the position of the temporary fence detected as the construction site by the construction site detection unit on the screen of the map data.

[0017] The vehicle may further include a location information display unit that displays location information of the construction site corresponding to the position of the construction site detected by the construction site detection unit on a screen of the map data.

[0018] The system may also be provided with a period display unit that performs fixed-point observation of a specified target area using the observation image and displays the detection period of the construction site on the screen of the map data based on the observation results of the fixed-point observation.

[0019] The marking unit may also mark and display the position of the construction site in a different color for each detection time based on the observation results.

[0020] Furthermore, when the construction site detection unit is no longer able to detect the construction site, the marking unit may continue to mark the position of the construction site for a certain period of time by changing the display color.

[0021] In addition, a construction site detection program according to another aspect of the present invention has a computer execute a learning step of learning, by machine learning, construction site identification information for identifying a construction site on the ground contained in an observation image taken from the air, and a construction site detection step of detecting the construction site from the observation image based on the construction site identification information learned in the learning step.

[0022] In addition, the observation image includes an image of a temporary fence used at an on-ground construction site, and in the learning step, temporary fence identification information for identifying the temporary fence based on the observation image is learned as the construction site identification information using machine learning, and in the construction site detection step, the temporary fence can be detected as the construction site from the observation image based on the temporary fence identification information learned in the learning step. [Effects of the Invention]

[0023] According to the present invention, it is possible to provide a construction site detection device and a detection program that can improve work efficiency by automatically identifying construction sites and that can visualize and understand construction sites in a target area on a map. [Brief explanation of the drawings]

[0024] [Figure 1] 1 is a functional block diagram showing a functional configuration of a detection system for a construction site according to an embodiment; [Figure 2] 1 is a schematic diagram illustrating a schematic configuration of a detection system for a construction site according to an embodiment. [Figure 3A] 10A to 10C are images showing an example of a temporary fence as an identification target. [Figure 3B] 10(d) to 10(g) are images showing an example of a temporary fence as an identification target. [Figure 3C] 10(h) to 10(j) are images showing an example of a temporary fence as an identification target. [Figure 4] 2A to 2C are explanatory diagrams showing an example of the configuration of a temporary fence detection unit in a construction site detection system according to an embodiment. [Figure 5] 1 is an enlarged explanatory diagram showing an example of a temporary fence detection unit in a construction site detection system according to an embodiment. FIG. [Figure 6] 10 is an enlarged explanatory diagram showing another example of a temporary fence detection unit in the construction site detection system according to the embodiment. FIG. [Figure 7] 10 is an enlarged explanatory diagram showing another example of a temporary fence detection unit in the construction site detection system according to the embodiment. FIG. [Figure 8] FIG. 10 is a schematic diagram illustrating the schematic configuration of a detection system for a construction site according to another embodiment. [Figure 9] 10 is a flowchart illustrating an example of the processing steps of a temporary fence model learning process executed in a construction site detection system according to an embodiment. [Figure 10A] 10 is a flowchart illustrating an example of a processing procedure for temporary fence detection processing executed by a construction site detection system according to an embodiment. [Figure 10B]10 is a flowchart illustrating an example of a processing procedure for temporary fence detection processing executed by a construction site detection system according to an embodiment. [Figure 11] 10 is a flowchart illustrating an example of a processing procedure for past data storage processing executed in the construction site detection system according to the embodiment. [Figure 12] An enlarged explanatory diagram showing an example of a temporary fence detection unit in a construction site detection system according to another embodiment. [Figure 13] An enlarged explanatory diagram showing an example of a temporary fence detection unit in a construction site detection system according to another embodiment. [Figure 14A] 10 is a flowchart showing an example of the processing procedure for temporary fence detection processing executed by the construction site detection system configured as shown in FIG. 2 according to another embodiment. [Figure 14B] 14 is a flowchart showing another example of the processing procedure for temporary fence detection processing executed by the construction site detection system configured as shown in FIG. 13 according to another embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0025] A detection system S1 for a construction site according to an embodiment of the present invention will be described with reference to FIGS.

[0026] In the following description of the drawings, the same or similar parts are denoted by the same or similar reference numerals. However, it should be noted that the drawings are schematic and the dimensional ratios may differ from those of the actual parts.

[0027] Therefore, specific dimensions should be determined with reference to the following explanation. Of course, the dimensional relationships and ratios may differ between the drawings.

[0028] (Outline of the detection system at the construction site) The functional configuration of a detection system S1 for a construction site according to an embodiment will be described with reference to the functional block diagram of FIG.

[0029] As shown in FIG. 1, the construction site detection system S1 can be configured, for example, from an image acquisition device 100 and a construction site detection device 200 configured from a personal computer PC1 (see FIG. 2) or the like.

[0030] Image acquisition device 100 can acquire images including construction sites such as construction sites, demolition sites, etc. Image acquisition device 100 can also acquire images including places that were previously construction sites but are now vacant lots after construction sites, demolition sites, etc. have been completed. The image acquisition device 100 can be a satellite or flying vehicle that carries an image capture device and moves through the sky.

[0031] The satellite constituting the image acquisition device 100 can be, for example, an earth observation satellite 101 that orbits the earth and captures observation images of the ground from above using an onboard imaging device. The earth observation satellite 101 may be, for example, an optical satellite equipped with an optical device such as a digital camera as an imaging device, or may be a radar satellite such as a synthetic aperture radar (SAR) satellite equipped with an SAR as an imaging device.

[0032] SAR (Synthetic Aperture Radar) is an active sensor that irradiates microwaves obliquely onto the earth's surface and receives backscattered waves from the surface. The microwaves emitted by the SAR satellite 101 can penetrate optical obstacles such as clouds that block visibility, and do not require sunlight to observe the Earth's surface, making observations possible in all weather conditions, including at night.

[0033] The imaging device of the earth observation satellite 101 can also repeatedly photograph the same area on the ground as a fixed-point observation each time it passes over that area. The construction site detection device 200 can acquire position information of the area on the ground that appears in the observation image taken by the earth observation satellite 101 based on the navigation trajectory information of the earth observation satellite 101. Construction sites can be detected by extracting feature points from images taken by optical satellites and recognizing patterns. Feature points can also be extracted by subtracting images taken at two different times through fixed-point observation. In the following description, the earth observation satellite 101 is an optical satellite.

[0034] The air vehicle constituting the image acquisition device 100 can be, for example, a manned or unmanned air vehicle 102 that takes observation images of the ground from the sky using an on-board imaging device. The manned or unmanned air vehicle 102 may be either a manned aircraft powered by a jet or propeller engine, or an unmanned aerial vehicle (UAV) powered by an engine or electric motor.

[0035] Unmanned aerial vehicles (UAVs) include multicopters with three or more rotors, including quadcopters with four rotors, which are widely used and known as drones.

[0036] The construction site detection device 200 can acquire the position information of the ground area that appears in the observation image taken by the imaging device of the manned aircraft 102 based on the flight trajectory information of the aircraft 102. The construction site detection device 200 can acquire position information of the area on the ground that appears in the observation image captured by the imaging device of the unmanned flying object 102 based on the remote control information of the flying object 102.

[0037] The detection device 200 at the construction site is configured to receive observation images of the ground taken from the sky by the image acquisition device 100 via a wireless line N1.

[0038] The construction site detection device 200 includes a learning unit 201 configured with an AI engine or the like that uses machine learning to learn construction site identification information for identifying construction sites based on received images. The construction site identification information can be information for determining whether feature points appearing in observed images of a construction site correspond to a construction site. The learning unit 201 can execute the learning step in a construction site detection program.

[0039] The learning unit 201 can use machine learning to learn, for each item, identification information for identifying items used at construction sites that are suitable for recognizing that they are construction sites, such as temporary fences for construction sites, sheets and panels for covering construction sites, as construction site identification information for identifying construction sites.

[0040] The sheets and panels may include, for example, protective sheets for covering construction, demolition and repair work sites, fireproof sheets, soundproof sheets, mesh sheets, painted sheets, and soundproof panels for temporarily enclosing construction sites.

[0041] The construction site detection device 200 also includes a construction site detection unit 202 that detects a construction site from an observation image of the ground based on the construction site identification information learned by the learning unit 201. In addition, if the learning unit 201 has a configuration corresponding to a temporary fence as the above-mentioned item, the construction site detection unit 202 can be configured to include a temporary fence detection unit 203 that detects a temporary fence as a construction site from an observation image of the ground. The construction site detection unit 202 can execute a construction site detection step in a construction site detection program.

[0042] Furthermore, the observation image acquired by the image acquisition device 100 can be an image captured from the sky in a direction tilted at a predetermined angle from the direction perpendicular to the ground. The predetermined angle can be, for example, 30°. By tilting the capture direction of the observation image, items erected on the ground, such as temporary fencing, can be more easily captured in the observation image, improving the detection efficiency of items suitable for recognizing a construction site, such as temporary fencing.

[0043] In addition, the construction site detection unit 202 can extract feature points from the observation image acquired from the image acquisition device 100, and determine whether or not a pattern corresponding to a construction site exists in the feature points based on the learning data of the learning unit 201, thereby detecting the construction site.

[0044] Furthermore, the construction site detection device 200 includes a marking unit (marking system) 204 that displays, by marking, the position of the construction site detected by the construction site detection unit 202 on the screen of map data M1 that indicates the area on the ground captured in the observation image. The marking unit 204 can be configured as a personal computer PC2 (see FIG. 2).

[0045] In addition, if the construction site detection unit 202 has a temporary fence detection unit 203, the marking unit 204 can be configured to mark and display the position of the temporary fence detected by the temporary fence detection unit 203 as the location of the construction site on the screen of the map data M1.

[0046] The marking unit 204 can display the screen of the map data M1 and a marker for marking display on a display unit (not shown) of the personal computer PC2.

[0047] In addition, the construction site detection device 200 is equipped with a location information display unit 205 that displays location information of the construction site or temporary fence on the above-mentioned display unit (not shown) based on the location information of the construction site captured in the observation image obtained from the image acquisition device 100, etc.

[0048] When the learning unit 201 learns the identification information of items suitable for recognizing a construction site, such as sheets and panels, for each item, the construction site detection unit 202 can detect each item individually as a construction site. When the construction site detection unit 202 detects each item individually, the names of the items detected by the construction site detection unit 202 may be displayed on a display unit (not shown) of the personal computer PC2 together with the screen of the map data M1 as a result of the construction site type determination.

[0049] The construction site detection device 200 also includes a period display unit 206 that displays the period during which the construction site is detected by the construction site detection unit 202 (or the period during which the temporary fence is detected by the temporary fence detection unit 203).

[0050] The marking section 204 can be configured to mark and display each detection time in a different color. Furthermore, when the construction site detection unit 202 is no longer detecting a construction site, the marking unit 204 can be configured to mark and display the position of the construction site in a color different from that during the detection period of the construction site.

[0051] (Outline of the construction site detection system configuration) The schematic configuration of the detection system S1 for a construction site will be described with reference to FIG.

[0052] FIG. 2 is a schematic diagram showing the schematic configuration of a detection system S1 at a construction site according to an embodiment.

[0053] In this configuration example, an earth observation satellite 101 is used as the image acquisition device 100. The earth observation satellite 101 is configured to acquire observation images of the ground, including the temporary fencing at the construction site.

[0054] Furthermore, temporary fences can be detected (extracted) by the learning unit 201 configured by the personal computer PC1 of the detection device 200 at the construction site. In addition, the marking unit 204, which is configured using the personal computer PC2 of the construction site detection device 200, can mark and display the position of the temporary fence detected by the temporary fence detection unit 203 on the screen of the map data M1 on a display unit not shown.

[0055] The screen data of the map data M1 including the marking display can be stored as backup data in a hard disk drive or the like and displayed on a monitor or the like. The hard disk drive can be, for example, the hard disk drive of the personal computer PC2 of the detection device 200 at the construction site.

[0056] (Examples of temporary enclosures) Next, a brief description will be given of the temporary enclosure as an identification target with reference to FIGS. 3A to 3C.

[0057] Here, FIGS. 3A to 3C are images (a) to (j) showing examples of temporary fences as identification targets. For example, image (a) shows an example of a fence-like temporary enclosure, and image (b) shows an example of an openable temporary enclosure.

[0058] The temporary enclosures shown in the images (a) to (j) each take on a different form when photographed from above by an earth observation satellite 101 equipped with an imaging device (digital camera or the like). Then, by having the learning unit 201 learn the different shapes for each temporary fence, temporary fence identification information can be accumulated.

[0059] (Example of the configuration of the temporary fence detection unit) An example of the configuration of the temporary fence detection unit 203 in the construction site detection system S1 according to the embodiment will be described with reference to FIGS.

[0060] Here, Figures 4(a) to (c) are explanatory diagrams showing an example configuration of the temporary fence detection unit 203 in the construction site detection system S1 of the embodiment, Figure 5 is an enlarged explanatory diagram showing an example of the temporary fence detection unit in the construction site detection system S1, and Figures 6 and 7 are enlarged explanatory diagrams showing other examples of the temporary fence detection unit 203 in the construction site detection system S1.

[0061] In the configuration example shown in FIG. 5, the position information of the temporary fence is acquired from the position information indicating the area of ​​the image observed by the earth observation satellite 101.

[0062] Furthermore, observation images of the ground, including temporary fences, can be acquired by the earth observation satellite 101. Then, temporary fences can be detected (extracted) by image judgment based on temporary fence identification information learned by a learning unit 201 provided in a personal computer PC1 serving as a detection device 200 at the construction site.

[0063] The marking unit (marking system) 204 can mark and display the position of the temporary fence detected by the temporary fence detection unit 203 on the screen of the map data M1 on a display unit (not shown). Furthermore, the marking unit (marking system) 204 can mark and display the position of the temporary fence on the display unit in a different color for each detection time based on the detection result of the temporary fence by the temporary fence detection unit 203.

[0064] Furthermore, when the temporary fence detection unit 203 is no longer detecting a temporary fence, the marking unit 204 may be configured to mark and display the position of the temporary fence in a color different from that during which the temporary fence was detected.

[0065] When the temporary fence is no longer detected, the marking display of the temporary fence's location is not terminated, but the display color is changed, thereby suggesting to the user viewing the screen of map data M1 on a display unit (not shown) that the construction may have been completed. A user who is informed that construction may have been completed can check whether the location has been cleared after construction was completed, for example, by checking whether there is a structure in the location where the marking display color has changed.

[0066] When the marking unit 204 performs a process of changing the display color of the location of a temporary fence that is no longer detected without terminating the marking display, this process can be considered to be a process of assisting the user in finding the location that has been cleared after construction.

[0067] In the configuration example shown in FIG. 6, the position information of the temporary fence is acquired from the position information indicating the area of ​​the image observed by the earth observation satellite 101.

[0068] Furthermore, observation images of the ground, including the temporary fence, can be acquired by the earth observation satellite 101. In this case, the earth observation satellite 101 can be controlled and adjusted so as to acquire observation images taken from the sky in a direction tilted at a predetermined angle (30° in the example shown in FIG. 6) from the direction perpendicular to the ground.

[0069] Then, the temporary fence can be detected (extracted) by image determination based on the temporary fence identification information learned by the learning unit 201 provided in the personal computer PC2 serving as the detection device 200 for the construction site.

[0070] The marking unit (marking system) 204 can mark and display the position of the temporary fence detected by the temporary fence detection unit 203 on the screen of the map data M1 on the display unit. Furthermore, the marking unit (marking system) 204 can mark and display the position of the temporary fence in a different color depending on when the temporary fence is detected.

[0071] Furthermore, when the temporary fence detection unit 203 is no longer detecting a temporary fence, the marking unit 204 may be configured to mark and display the position of the temporary fence in a color different from that during which the temporary fence was detected.

[0072] In the configuration example shown in FIG. 7, the position information of the temporary fence is acquired from position information indicating the area of ​​the images observed by the earth observation satellite 101 on different observation dates.

[0073] Furthermore, observation images of the ground, including temporary fences, can be acquired by the earth observation satellite 101. In this case, the temporary fence detection unit 203 can detect temporary fences (extract the difference between the presence and absence of temporary fences) based on observation images acquired by the earth observation satellite 101 at multiple times (two times, observation days A and B, in the example shown in FIG. 7).

[0074] The marking unit (marking system) 204 can mark and display the position of the temporary fence detected by the temporary fence detection unit 203 on the screen of the map data M1 on a display unit (not shown). Furthermore, the marking unit (marking system) 204 can mark and display the position of the temporary fence on the display unit in a different color for each detection time based on the detection result of the temporary fence by the temporary fence detection unit 203.

[0075] Furthermore, when the temporary fence detection unit 203 is no longer detecting a temporary fence, the marking unit 204 may be configured to mark and display the position of the temporary fence in a color different from that during which the temporary fence was detected.

[0076] (Construction site detection system according to another embodiment) FIG. 8 is a schematic diagram showing the schematic configuration of a detection system S1 at a construction site according to another embodiment.

[0077] The difference from the embodiment shown in FIG. 2 is that an SAR satellite equipped with a synthetic aperture radar (SAR) is used as the earth observation satellite 101 instead of an optical satellite. This makes the system less susceptible to the effects of weather, etc., and allows for stable and accurate detection of temporary fences at construction sites.

[0078] (Temporary enclosure model learning process) The processing steps of the temporary fence model learning process executed by the construction site detection device 200 of the construction site detection system S1 (S1a) according to this embodiment will be described with reference to the flowchart of FIG.

[0079] When this process starts, first in step S101, training data is acquired and input into the model, and then the process proceeds to step S102.

[0080] In step S102, the type of temporary enclosure is identified, and the process proceeds to step S103. In step S103, it is determined whether or not learning data has been input. If "No", the process returns to step S101, and if "Yes", the process proceeds to step S104. In step S104, the trained model is confirmed and the process ends.

[0081] (Temporary fence detection processing) The processing procedure of the temporary fence detection processing executed by the construction site detection device 200 of the construction site detection system S1 (S1a) according to this embodiment will be described with reference to the flowcharts of FIGS. 10A and 10B.

[0082] When this process starts, as shown in FIG. 10A, first, in step S201, an observation image (a satellite image from an earth observation satellite 101 or an aerial image from an aircraft 102) is acquired by the image acquisition device 100, and the process proceeds to step S202.

[0083] In step S202, the data is input into the trained model and the process proceeds to step S203. In step S203, a temporary fence is detected and the process proceeds to step S204. If a temporary fence cannot be detected, the process essentially proceeds to step S209.

[0084] In step S204, the first detection date, target location, and temporary fence type for the detection target are recorded, and then the process proceeds to step S205. In step S205, it is determined whether this is the first detection or whether the detection has occurred again within a period exceeding the set input period 1 since the first detection. If the determination result is "No", the process proceeds to step S206, and if the determination result is "Yes", the process proceeds to step S208.

[0085] In step S206, the detected object is marked in color A on the map on which the satellite aerial image was acquired, and the process proceeds to step S207. In step S207, the recorded information of the detection target is displayed on the system as new information during the set input period 1, and the process ends.

[0086] On the other hand, in step S208, the detected object is marked in B color on the map from which the satellite aerial image was acquired, and the process proceeds to step S209. In step S209, it is determined whether the temporary enclosure object is no longer detected. If the determination result is "No," the process proceeds to step S210, where the process ends with no change in the recorded information.

[0087] If the answer is "Yes", the process proceeds to step S211 in FIG. 10B, where it is determined whether the period of time that has elapsed since it was determined that the signal was no longer detected exceeds the set input period 2.

[0088] If the determination result is "No", the process proceeds to step S212, where the detected object is marked in C color on the map from which the satellite aerial image was acquired, and the process ends. If the answer is "Yes", the process proceeds to step S213, where the marking display of the temporary enclosure target is terminated, and the process ends.

[0089] (Past data storage processing) The processing procedure of the past data storage process executed in the construction site detection system S1 (S1a) according to this embodiment will be described with reference to the flowchart of FIG.

[0090] When this process starts, first, temporary enclosure detection processing is executed in step S301, and then the process proceeds to step S302. In step S302, images are periodically acquired by fixed point observation from a satellite, and the process proceeds to step S303. In step S303, the images and marking information acquired by the fixed point observation and each recorded data are saved, and the process ends.

[0091] (Construction site detection system according to another embodiment) FIG. 12 is a schematic diagram showing the schematic configuration of a detection system S1 at a construction site according to another embodiment.

[0092] 6 is that instead of taking observation images from the sky in a direction tilted from the perpendicular direction, the imaging device C1 of the earth observation satellite 101 takes multiple stereoscopic observation images of the same area on the ground from multiple angles, and the construction site detection device 200 detects the construction site (or its temporary fence) from the stereoscopic observation images. The stereoscopic observation images can be obtained, for example, by having the imaging device C1 take images of the same area on the ground from a distant point on the orbit around the earth observation satellite 101.

[0093] This allows 3D (three-dimensional) stereoscopic map data M1 to be obtained from the observation image, making it easy to see the sides of the temporary fence, and the construction site detection unit 202 (temporary fence detection unit 203) can accurately detect the temporary fence at the construction site.

[0094] (Construction site detection system according to another embodiment) FIG. 13 is a schematic diagram showing the schematic configuration of a detection system S1 at a construction site according to another embodiment.

[0095] The difference from the embodiment shown in Figure 6 is that after the construction site detection device 200 detects the position of the temporary fence using observation images from the imaging device C1 of the Earth observation satellite 101, which is an optical satellite, it monitors whether the temporary fence at the detected position continues to be detected using observation images from the imaging device C1 of the Earth observation satellite 101, which is a SAR satellite.

[0096] When the view of observation images taken by an optical satellite is obstructed on cloudy or rainy days, the construction site detection device 200 may be unable to detect the temporary fence from the observation images. Therefore, by continuously detecting the position of a temporary fence that has already been detected from observation images taken by an SAR satellite, which is a radar satellite, the temporary fence can be detected with high accuracy even on cloudy or rainy days when the view is obstructed.

[0097] (Construction site detection system according to another embodiment) Figures 14A and 14B are flowcharts showing an example of the processing steps for temporary fence detection processing executed by the construction site detection device 200 in a construction site detection system S1 according to another embodiment. Figure 14A shows a case in which the temporary fence detection unit 203 detects a temporary fence using only observation images from the Earth observation satellite 101, an optical satellite, in the configuration of the construction site detection system S1 in Figure 2, and Figure 14B shows a case in which the temporary fence detection unit 203 detects a temporary fence using observation images from the Earth observation satellite 101, an optical satellite, and observation images from the Earth observation satellite 101, an SAR satellite, in the configuration of the construction site detection system S1 in Figure 13.

[0098] The difference between the processing procedure of the temporary fence detection process shown in Figures 14A and 14B executed by the construction site detection device 200 of this embodiment and the processing procedure of the temporary fence detection process shown in Figure 10B is the timing at which the color of the marking display of the detection target (temporary fence) is changed from color B to color C when the construction site detection device 200 can no longer detect the position of the temporary fence.

[0099] In the processing procedure of Figure 10B, when the detection device 200 at the construction site no longer detects the position of the temporary fence, the color of the marking display of the detection target (temporary fence) is changed from color B to color C, and when the elapsed period from the day when it was determined that the detection device 200 at the construction site no longer detected the position of the temporary fence exceeds the set input period 2, the marking display of the detection target (temporary fence) is terminated.

[0100] In the processing procedure of Figure 14A, even if the detection device 200 at the construction site no longer detects the position of the temporary fence, the color of the marking display of the detection target (temporary fence) remains color B until the elapsed period since detection stopped exceeds set input period 2 (step S211A).

[0101] When the set input period 2 is exceeded, the color of the marking display of the detection target (temporary enclosure) is changed from color B to color C, and when the set input period 3, which is longer than the set input period 2 (for example, one month), is exceeded, the marking display of the detection target (temporary enclosure) is terminated (step S213).

[0102] According to this processing procedure, even if the temporary fence becomes temporarily undetectable from the observation image of the optical satellite Earth observation satellite 101 due to weather or other factors, the color of the marking display will not be immediately changed to C, which indicates that construction has been completed and the site has been cleared, but rather the color of the marking display can be changed to C after determining whether the temporary fence can be detected again from the observation image of the optical satellite Earth observation satellite 101.

[0103] In the processing procedure of Figure 14B, when it is determined that the detection device 200 at the construction site can no longer detect the position of the temporary fence, it determines whether the observation is being affected by weather such as rain or clouds based on whether the temporary fence can be detected from the observation image of the SAR satellite (step S210A), and if the observation is being affected by weather, the color of the marking display of the detected object (temporary fence) remains color B (step S210B).

[0104] If the observation is not affected by weather, the color of the marking display of the detection target (temporary fence) is changed from color B to color C, and when the set input period 2 has elapsed since the day it was determined that the detection device 200 at the construction site no longer detected the position of the temporary fence, the marking display of the detection target (temporary fence) is terminated (step S213).

[0105] According to this processing procedure, even if the temporary fence temporarily becomes undetectable from the observation image of the Earth observation satellite 101, which is an optical satellite, due to the influence of weather or the like, the color of the marking display is not immediately changed to C, which indicates that construction has been completed and the site has been cleared, for example, but rather, it is possible to check whether the temporary fence can be detected from the observation image of the Earth observation satellite 101, which is an SAR satellite (radar satellite), and then change the color of the marking display to C.

[0106] In the above embodiment, an example has been described in which the marking unit 204 marks and displays the position of the temporary fence detected by the temporary fence detection unit 203 of the construction site detection unit 202 as the position of the construction site.

[0107] However, the construction site detection unit 202 also detects sheets and panels other than temporary fencing, such as protective sheets, fireproof sheets, soundproof sheets, mesh sheets, painted sheets, and soundproof panels for temporarily fencing a construction site, as items suitable for recognizing that the site is a construction site.

[0108] Therefore, the marking unit 204 can mark and display the position of each item detected by the construction site detection unit 202 as the position of the construction site, in the same way as in the case of a temporary fence. Furthermore, the marking unit 204 can change the display color of markings relating to items other than temporary enclosures, or can terminate the display, in the same way as for temporary enclosures.

[0109] As described above, according to this embodiment, it is possible to provide a construction site detection device 200 and a detection program that can improve work efficiency by automatically identifying construction sites and that can visualize and grasp construction sites in a target area on a map.

[0110] The above describes the construction site detection system and detection program of the present invention based on the illustrated embodiment, but the present invention is not limited to this, and the configuration of each part can be replaced with any configuration having similar functions.

[0111] For example, when items suitable for identifying a construction site, such as temporary fencing, are no longer detected, the color of the marking display can be changed to a different color to confirm that construction has ended. Furthermore, the location of a newly detected construction site can be marked in different colors within a certain period and after the passage of a certain period, making it possible to identify the new construction site.

[0112] Furthermore, even after items suitable for identifying a construction site, such as temporary fences, are no longer detected, by continuing to display the markings for a certain period of time, it is possible to identify properties that have been newly converted into vacant lots or newly constructed buildings.

[0113] For example, when a user searches for a location that has become vacant after construction work is completed, a construction site detection device 200 and a detection program can be provided that do not terminate the marking display that indicates the location of a construction site that is no longer detected, but instead change the display color to visualize locations that may have become vacant after construction work is completed, making them easier for the user to recognize, and thereby assisting the user in finding the vacant site. [Explanation of symbols]

[0114] S1 Construction Site Detection System 100 Image acquisition device 101 Earth Observation Satellite 102 Flying Objects 200 Construction site detection device (personal computers PC1, PC2, etc.) 201 Learning Department 202 Construction Site Detection Unit 203 Temporary Fence Detection Unit 204 Marking section 205 Local Information Dispatch Department 206 Period Representation Department

Claims

1. a learning unit that learns, by machine learning, construction site identification information for identifying a construction site on the ground included in an observation image from an observation image of the ground taken from above; a construction site detection unit that detects the construction site from the observation image based on the construction site identification information learned by the learning unit; a marking unit that marks the position of the construction site detected by the construction site detection unit on a screen that displays map data including the detected construction site; Equipped with The marking unit marks and displays the position of the construction site in a different color for each detection time based on the observation results of fixed-point observation of a predetermined target area using the observation image. Construction site detection equipment.

2. a learning unit that learns, by machine learning, construction site identification information for identifying a construction site on the ground included in an observation image from an observation image of the ground taken from above; a construction site detection unit that detects the construction site from the observation image based on the construction site identification information learned by the learning unit; a marking unit that marks the position of the construction site detected by the construction site detection unit on a screen that displays map data including the detected construction site; Equipped with When the construction site detection unit is no longer able to detect the construction site, the marking unit changes the display color of the marking for the position of the construction site and continues the marking for a certain period of time. Construction site detection equipment.

3. The observation image includes an image of a temporary fence used at a construction site on the ground, the learning unit learns temporary fence identification information for identifying the temporary fence based on the observed image as the construction site identification information through machine learning; The construction site detection device described in claim 1 or claim 2, characterized in that the construction site detection unit detects the temporary fence as the construction site from the observation image based on the temporary fence identification information learned by the learning unit.

4. 4. The construction site detection device according to claim 1, wherein the observation image is an image taken from the sky in a direction inclined at a predetermined angle from a direction perpendicular to the ground.

5. 5. The construction site detection device according to claim 1, wherein the observation image is an image taken by an imaging device mounted on a satellite or an aircraft moving in the sky.

6. The satellite is a synthetic aperture radar (SAR) satellite, 6. The construction site detection device according to claim 5, wherein the construction site detection unit detects the construction site based on the observation images taken at multiple times by the imaging device of the SAR satellite.

7. The construction site detection device according to claim 5 or 6, characterized in that the marking unit marks and displays on the screen of the map data the position of the construction site detected by the construction site detection unit, which is identified based on ground position information shown in the observation image obtained from the satellite or the aircraft.

8. The observation image includes an image of a temporary fence used at a construction site on the ground, the learning unit learns temporary fence identification information for identifying the temporary fence based on the observed image as the construction site identification information through machine learning; the construction site detection unit detects the temporary fence as the construction site from the observation image based on the temporary fence identification information learned by the learning unit; 8. The construction site detection device according to claim 7, wherein the marking unit marks and displays the position of the temporary fence detected by the construction site detection unit as the construction site on a screen of the map data.

9. The construction site detection device according to claim 7 or 8, further comprising a location information display unit that displays location information of the construction site corresponding to the position of the construction site detected by the construction site detection unit on a screen of the map data.

10. A construction site detection device as described in any one of claims 7 to 9, characterized in that it is characterized by having a period display unit that performs fixed-point observation of a specified target area using the observation image and displays the detection period of the construction site on the screen of the map data based on the observation results of the fixed-point observation.

11. The construction site detection device according to claim 10, characterized in that, when the construction site detection unit is no longer able to detect the construction site, the marking unit continues to mark the location of the construction site for a certain period of time by changing the display color.

12. a learning step of learning, by machine learning, construction site identification information for identifying a construction site on the ground included in an observation image taken from the sky; a construction site detection step of detecting the construction site from the observation image based on the construction site identification information learned in the learning step; a marking step of marking the position of the construction site detected in the construction site detection step on a screen displaying map data including the detected construction site in a different color for each detection time based on the observation results of fixed-point observation of a predetermined target area using the observation image; to the computer, Construction site detection program.

13. a learning step of learning, by machine learning, construction site identification information for identifying a construction site on the ground included in an observation image taken from the sky; a construction site detection step of detecting the construction site from the observation image based on the construction site identification information learned in the learning step; a marking step of marking the position of the construction site detected in the construction site detection step on a screen displaying map data including the detected construction site, and, when the construction site is no longer detected in the construction site detection step, continuing the marking display of the position of the construction site for a certain period of time by changing the display color; to the computer, Construction site detection program.

14. The observation image includes an image of a temporary fence used at a construction site on the ground, In the learning step, temporary fence identification information for identifying the temporary fence based on the observed image is learned as the construction site identification information by machine learning; A construction site detection program as described in claim 12 or 13, characterized in that in the construction site detection step, the temporary fence is detected as the construction site from the observation image based on the temporary fence identification information learned in the learning step.

Images