Underground structure detection device and underground structure detection method

Abstract

To accurately detect an underground buried object detected by a radar in an underground structure detection apparatus.SOLUTION: An underground structure detection apparatus detects information including a component constituting an underground structure, a type of the component, and position / posture information on the basis of a plurality of pieces of two-dimensional data indicating a cross section of ground and three-dimensional data indicating the underground structure. With respect to synthetic data obtained by synthesizing the two-dimensional data and the three-dimensional data, the position / posture information of a component represented by the synthetic data is corrected on the basis of an image feature grasped from the two-dimensional data. When a determination is made in a transverse direction with respect to a road, and the image feature is included in continuous images by a threshold or more, it is determined as a transverse pipe, and when a determination is mode in a longitudinal direction with respect to a road, and the image feature is included in continuous images by a threshold or less, it is determined as a longitudinal pipe, and the posture is corrected for each direction.SELECTED DRAWING: Figure 7

Description

[Technical Field]

[0001] The present invention relates to an underground structure detection device and an underground structure detection method, and more particularly to an underground structure detection device and an underground structure detection method suitable for detecting buried objects underground detected by radar with high accuracy. [Background technology]

[0002] In recent years, urban redevelopment and measures to remove utility poles have led to the active use of underground spaces. Water pipes, gas pipes, communication lines, electric wires, and other buried objects are present in underground spaces, and it is necessary to survey these buried objects before starting construction work such as water or electrical work. Surveys of underground spaces are conducted using ground-penetrating radar, which allows for investigations without excavating the ground.

[0003] A technology for measuring underground buried objects from the road using radar is disclosed, for example, in Patent Document 1. The 3D under-road diagnostic system described in Patent Document 1 uses radar from a moving object such as a vehicle to explore the area below the road surface. It also measures the road surface condition using a camera or laser scanner. The system then acquires 3D position information of the moving object with high precision and adds the 3D position information to road surface condition measurement data and under-road surface exploration data, integrating the information and enabling it to be used as integrated data.

[0004] In addition, in order to detect elements located deep underground, a method can be considered in which a three-dimensional model of the underground structure is generated based on multiple radar images, and elements located deep underground are clarified.

[0005] A technique for estimating the position and orientation of an object using 3D data and 2D images is disclosed, for example, in Patent Document 2. In the position and orientation estimation device described in Patent Document 2, a 3D sensor acquires the three-dimensional coordinates of an object, a first position and orientation estimation unit uses the 3D data to optimize six parameters (translation x, y, z and rotation φ, γ, θ), and a second position and orientation estimation unit 146 optimizes only three parameters (translation x, y and rotation θ) that can be estimated with high accuracy using 2D images based on the results. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Application Laid-Open No. 2015-90345 [Patent Document 2] Japanese Patent Application Publication No. 2019-109747 Summary of the Invention [Problem to be solved by the invention]

[0007] The 3D under-road diagnostic system described in Patent Document 1 is capable of accurately determining the positional relationship between the road surface condition and the under-road condition by integrating road surface condition camera images, road surface condition 3D point cloud data, and depth direction information. However, the 3D under-road diagnostic system described in Patent Document 1 does not have a function for determining the position and orientation of components of the underground structure.

[0008] In ground-penetrating radar surveys, it is possible that elements located deeper underground may not be clearly visible in the radar images. As a result, detection of underground structures using only radar images is limited to detecting elements near the road surface.

[0009] The technology described in Patent Document 2 optimizes a 3D model using 2D image information to accurately determine the orientation of an object. However, due to the distance measurement method, it can only measure the subject at the forefront, and therefore cannot be applied to underground exploration as is.

[0010] An object of the present invention is to provide an underground structure detection device and an underground structure detection method that can accurately detect underground buried objects detected by radar. [Means for solving the problem]

[0011] A feature of the underground structure detection device of the present invention is that it is an underground structure detection device that detects information including the components that make up the underground structure, the types of the components, and position and orientation information, based on multiple two-dimensional data showing cross sections of the ground and three-dimensional data showing the underground structure, and that corrects the position and orientation information of the components represented by the composite data, which is a combination of two-dimensional data and three-dimensional data, based on image features grasped from the two-dimensional data. [Effects of the Invention]

[0012] According to the present invention, it is possible to provide an underground structure detection device and an underground structure detection method that can accurately detect buried objects detected by radar. [Brief explanation of the drawings]

[0013] [Figure 1] 1 is a diagram showing a schematic configuration of an underground structure detection device and how an underground structure is detected. [Figure 2] FIG. 1 is a diagram showing the hardware and software configuration of the underground structure detection device. [Figure 3] FIG. 2 is a diagram illustrating measurement data and design drawing data. [Figure 4] 10 is a flowchart showing a series of processes performed by the underground structure detection device, from acquiring measurement data and design drawing data to outputting analysis results of components. [Figure 5] 10 is a flowchart showing the details of the 3D model generation and component element classification process. [Figure 6] 10 is a flowchart showing a process of estimating the position and orientation of a component. [Figure 7]10 is a flowchart showing a correction process using two-dimensional data according to the first embodiment. [Figure 8A] FIG. 10 is a diagram illustrating image features in a crossing pipe. [Figure 8B] FIG. 10 is a diagram illustrating image features in a longitudinal section of a pipe. [Figure 9] 10 is a flowchart showing details of a component identification process for component compound information. [Figure 10] FIG. 10 is a diagram illustrating a specific flow of component element composite information and component element type / position information. [Figure 11] 10 is a flowchart showing a process of determining a pair of different types of components. [Figure 12] FIG. 10 is a diagram showing an example of a component information display screen. [Figure 13] 10 is a flowchart showing a correction process using two-dimensional data according to the second embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0014] Hereinafter, each embodiment of the present invention will be described with reference to FIGS.

[0015] [Embodiment 1] Hereinafter, a first embodiment of the present invention will be described with reference to FIGS. In this embodiment, the term "underground structure" refers to a structure (state, position) in an underground space including buried objects such as water pipes, gas pipes, communication lines, and electric wires, and underground cavities. Furthermore, the term "component" refers to an element that constitutes the underground structure. In other words, the component is, for example, a buried object such as water pipes, gas pipes, communication lines, and electric wires, or an underground cavities.

[0016] First, the configuration of the underground structure detection device will be described with reference to FIGS. In the following explanation, the three-dimensional coordinate system will be described with the X-axis direction representing the direction in which the vehicle moves (towards the road), the Y-axis direction representing the width of the vehicle (width of the road), and the Z-axis direction representing the depth of the ground (vertical downward direction).

[0017] The underground structure detection device 1 shown in Fig. 1 is a device that detects an underground structure and outputs information about the detected underground structure (such as its position, size, and other auxiliary information about its components) to a user. That is, the underground structure detection device 1 detects the underground structure based on measurement data of the underground structure acquired from a vehicle 2 equipped with an underground exploration radar, for example. The underground structure detection device 1 displays the detected information about the underground structure on a display device (described later with reference to Fig. 2), for example.

[0018] As shown in FIG. 2, the underground structure detection device 1 is connected to a vehicle 2 and a data center via a network 5. The database server 70 is connected to the user terminal 6 . The underground structure detection device 1 is not limited to acquiring measurement data of the underground structure from the vehicle 2. First, a hand-held ground-penetrating radar that can be moved by human power or Portable ground-penetrating radar etc. Measurement data of the underground structure may be obtained from the device.

[0019] Car 2 is Road 1, the underground space and road surface 41 are continuously measured by the measuring device 23. The measuring device 23 is equipped with an underground detection radar 231 and a camera 232. A plurality of underground detection radars 231 are provided in the width direction of the vehicle 2, and the underground space is continuously measured by irradiating the radar 24 in the Z-axis direction.

[0020] The underground exploration radar 231 measures the elements 3, namely, the manhole 31, the cavity 32, and the water pipe 33 (water pipe 33(1) and water pipe 33(2) in the illustration of FIG. 1 ), as the vehicle 2 moves, for example. The water pipe 33(1) and the water pipe 33(2) are both cross pipes (described later) with respect to the road, and the water pipe 33(2) is located at a position with a larger Z coordinate than the water pipe 33(1) (i.e., at a position deeper underground). The measurement data of the underground space by the underground exploration radar 231 is sent wirelessly to the underground structure detection device 1 by the radar data processing unit 22.

[0021] Furthermore, the camera 232 provided in the measuring device 23 continuously captures images of the road surface 41. The captured images show, for example, manholes 31 visible on the road surface 41. The captured images of the road surface are sent wirelessly to the underground structure detection device 1 by the camera data processing unit 21.

[0022] As shown in Figure 1, the underground structure detection device 1 includes a 3D model generation unit 11, a 3D model component detection unit 12, a road surface data component detection unit 13, a blueprint data component detection unit 14, a component information processing unit 15, and a display unit 16.

[0023] The three-dimensional model generating unit 11 is a functional unit that generates a three-dimensional model (a specific image of which will be described later) showing the underground structure based on the measurement data of the underground exploration radar 231.

[0024] The 3D model component detection unit 12 is a functional unit that detects components that make up the underground structure based on the 3D model generated by the 3D model generation unit 11. 3D detection data 201 (described later with reference to FIG. 5 ) is an example of the detection result of the 3D model component detection unit 12, and includes, for example, type information, position information, or shape information of the component. The 3D model component detection unit 12 sends the 3D detection data 201 to the component information processing unit 15.

[0025] The process by which the three-dimensional model component detector 12 generates position information and shape information of the components will be described in detail below with reference to FIGS.

[0026] The road surface data component detector 13 is a functional unit that detects components on the road surface based on road surface data 131 (described later with reference to FIG. 3) obtained by photographing the road surface 41 and obtained from the camera data processor 21.

[0027] The road surface data component detection unit 13 detects, for example, a manhole 31 from the road surface data 131. Road surface detection data 202 (described later with reference to FIG. 5 ), which is an example of the detection result of the road surface data component detection unit 13, includes, for example, type information, position information, or shape information of the component on the road surface. The road surface data component detection unit 13 sends the road surface detection data 202 to the component information processing unit 15.

[0028] The design drawing data component detecting unit 14 is a functional unit that detects components shown in design drawing data 141 (described later with reference to FIG. 3). The design drawing data component detecting unit 14 acquires design drawing data 141 (141(1) and 141(2) in the illustration of FIG. 3) of the underground structure.

[0029] The design drawing data 141 may be a digitized version of a paper design drawing. For example, a user may digitize the paper design drawing using a scanner or the like and input the digitized design drawing data 141 to the design drawing data component detection unit 14.

[0030] The detection result of the design drawing data component detection unit 14 (design drawing detection data 203 shown in FIG. 5) includes the type, position, or shape of the component shown on the design drawing. The design drawing data component detection unit 14 sends the design drawing detection data 203 to the component information processing unit 15.

[0031] The blueprint data component detection unit 14 may have an optical character recognition (OCR) function to detect the blueprint detection data 203 from character information shown on the blueprint. When the blueprint data 141 is composed of characters, the blueprint data component detection unit 14 may generate visually recognizable image data of the type and position of the component.

[0032] The component information processing unit 15 is a functional unit that processes information about the detected components, analyzes the 3D detection data 201, road surface detection data 202, and blueprint detection data 203, and finally outputs the type and position information of the component.

[0033] The component information processing unit 15 includes, as sub-functional units, a dimension adjustment unit 151, a component compound information generation unit 152, and a component type and position information generation unit 153.

[0034] The dimension adjustment unit 151 is a functional unit that converts the dimensions of the three-dimensional data, road surface data, and design drawing data as necessary. Details of each data will be explained later. The dimension adjustment unit 151 converts the dimensions of the road surface data 131 and the design drawing data 141, for example, based on the dimensions of the three-dimensional data 121. The dimension adjustment unit 151 sends the dimension-adjusted three-dimensional data 121, road surface data 131, and design drawing data 141 to the component element composite information generation unit 152.

[0035] The component compound information generation unit 152 is a functional unit that generates compound information related to the components. The component compound information is information for integrating and evaluating three-dimensional data, road surface data, and blueprint data, each of which represents a single component. The component compound information generation unit 152 includes a component type identification unit 154 as a sub-functional unit. The component type identification unit 154 is a functional unit that identifies the type of component from a pair of components included in the component compound information. The component type identification unit 154 calculates an evaluation index based on, for example, the size of the overlapping area of ​​the component pair, and identifies the type of component based on the calculated evaluation index. The process of identifying the type of component using each evaluation index will be described in detail below.

[0036] The component compound information generation unit 152 determines whether the component pair indicates the same component by comparing the evaluation index with a predetermined threshold using the component type identification unit 154, and sends the component compound information (information relating to the type of component and data) classified into the component as a result to the component type / position information generation unit 153.

[0037] The component type and position information generating unit 153 generates a type (manhole, cavity, Water pipes, etc. and a functional unit that generates position information in a three-dimensional coordinate system.

[0038] The component type / position information generating section 153 sends the component type and position information to the display section 16 as component result data 522. Details of the processing of the component information processing section 15 will be described later with reference to FIGS.

[0039] The display unit 16 is a functional unit that displays the type of component and the position information of the component as the detection results of the underground structure detection device 1. The display unit 16 displays the component result data 522 of the component information processing unit 15 on a display device such as a display monitor.

[0040] Next, the hardware and software configurations of the underground structure detection device will be described with reference to FIG. As shown in FIG. 2, the underground structure detection device 1 includes, for example, a storage device 51, a CPU (Central Processing Unit) 52, a main memory 53, and a communication device 54, which are connected by a bus.

[0041] The storage device 51 is a device for storing large amounts of data, such as a hard disk drive (HDD) that is a magnetic storage medium, a solid state drive (SDD) that is a nonvolatile semiconductor device storage medium, etc. Installed in the storage device 51 are, for example, a 3D model generation program 511, a 3D model component detection program 512, a road surface data component detection program 513, a design drawing data component detection program 514, a component information processing program 515, and a display program 516.

[0042] The 3D model generation program 511, the 3D model component detection program 512, the road surface data component detection program 513, the blueprint data component detection program 514, the component information processing program 515, and the display program 516 are programs that realize the functions of the 3D model generation unit 11, the 3D model component detection unit 12, the road surface data component detection unit 13, the blueprint data component detection unit 14, the component information processing unit 15, and the display unit 16, respectively.

[0043] Furthermore, the storage device 51 of the underground structure detection device 1 stores component element analysis data 521 and component element result data 522.

[0044] The component analysis data 521 is intermediate data calculated when analyzing the components based on the measurement data. The component result data 522 is data including the type of component and the three-dimensional coordinate position that will ultimately be displayed to the user.

[0045] The program may be recorded in a storage medium 530 such as a USB (Universal Serial Bus) memory.

[0046] The CPU 52 implements each function by loading each program installed in the storage device 51 into the main memory 53 and executing the program. The main memory 53 is, for example, a volatile storage medium such as a RAM (Random Access Memory).

[0047] Communication device 54 is connected to the vehicle 2, the database server 70, and the user terminal 6 via the network 5 so as to be able to perform two-way communication. The network 5 may be connected wirelessly or by wire.

[0048] The user terminal 6 is an information processing terminal equipped with a display device 61, and is, for example, a portable terminal such as a smartphone or a personal computer (PC). User terminal 6 The image sensor may be provided in the underground structure detection device 1 and connected via a display interface to display images.

[0049] The database server 70 is a server device that stores the measurement data collected by the vehicle 2 in a measurement data DB 71 and provides a function for external access. Data transfer from the vehicle 2 to the database server 70 may be processed sequentially via the network 5, for example, or may be transferred at regular intervals and processed in batches.

[0050] Furthermore, the measurement data may be stored in the storage device 51 of the underground structure detection device 1, in addition to being stored in an external device as described above.

[0051] Next, the data structure used in the underground structure detection device will be described with reference to FIG. FIG. 3(1) shows an example of two-dimensional data as measurement data obtained from the underground radar 231.

[0052] The three-dimensional model generating unit 11 generates a plurality of two-dimensional data 111 by coloring the measurement data acquired from the radar data processing unit 22 of the vehicle 2.

[0053] The three-dimensional model generating unit 11 sets, for example, a length H1 in the height direction (Z-axis direction) and a length L1 in the horizontal direction (X-axis direction) for each two-dimensional data 111. For example, the length H1, the length L1, and the number n of each two-dimensional data 111 are set as the setting values ​​for each two-dimensional data 111. storage device 51 The data is stored as component element analysis data 521.

[0054] The three-dimensional model generation unit 11 also stores two-dimensional processing parameters 112 (described later with reference to FIG. 5) used when generating the two-dimensional data 111. The two-dimensional processing parameters are parameters that describe the properties of the two-dimensional data, such as the coloring intensity value, the receiver sensitivity setting value, or the relative dielectric constant.

[0055] Each piece of two-dimensional data 111 is data that displays a cross section of the ground. Each piece of two-dimensional data 111 is generated for each underground radar 231. That is, each piece of two-dimensional data 111(1) to 111(n) is generated corresponding to each of the n underground radars 231. Note that each piece of two-dimensional data 111(i) is illustrated as corresponding to the i-th (i=1, ..., n) underground radar 231 when viewed from the front side of the vehicle 2 (the side with the larger Y value in the Y-axis direction).

[0056] In the two-dimensional data 111(1), for example, water pipe 33(1) and water pipe 33(2) are displayed as images. For a certain i, in the two-dimensional data 111(i), for example, manhole 31, cavity 32, and water pipes 33(1) and 33(2) are displayed. In the two-dimensional data 111(i), for example, water pipe 33(1) and water pipe 33(2) are displayed. Note that because radar 24 is irradiated from road surface 41, water pipe 33(2) displayed in the two-dimensional data 111(1) is less clear than water pipe 33(1).

[0057] Next, FIG. 3(2) shows an example of three-dimensional data as measurement data obtained from the underground exploration radar 231. The three-dimensional model generating unit 11 generates three-dimensional data 121 from the measurement data of the underground exploration radar 231 by using synthetic aperture processing. Synthetic aperture processing is a technique for acquiring high-resolution information using multiple receivers.

[0058] The three-dimensional model generating unit 11 sets, for example, a length H in the height direction (Z-axis direction), a length L in the horizontal direction (X-axis direction), and a length W in the depth direction (Y-axis direction) for each piece of three-dimensional data 121. The three-dimensional model generating unit 11 sets, for example, the length H, the length L, and the length W as the setting values ​​for the three-dimensional data 121. storage device 51 The data is stored as component element analysis data 521.

[0059] The three-dimensional model generation unit 11 also stores three-dimensional processing parameters 122 (described later with reference to FIG. 5) used when generating the three-dimensional data 121. The three-dimensional processing parameters 122 are parameters that describe the properties of the three-dimensional data, and are values ​​used when generating the three-dimensional data, such as coloring intensity values, receiver sensitivity settings, or relative dielectric constants.

[0060] The three-dimensional data 121 is data that expresses an underground structure using three-dimensional coordinates. The three-dimensional data 121 shows, for example, a water pipe 33(1) and a water pipe 33(2). Although not shown, in the example of this embodiment, the three-dimensional data 121 may also show a manhole 31 and a cavity 32.

[0061] Next, FIG. 3(3) shows an example of road surface data as measurement data obtained from the camera 232. The road surface data component detecting unit 13 acquires road surface data 131 from the camera data processing unit 21. For example, a lateral length L2 and a depth length W2 are set in the road surface data 131. The road surface data component detecting unit 13 sets the length L2 and the length W2 as the set values ​​of the road surface data 131. storage device 51 as the component element analysis data 521. The road surface data 131 is data indicating the state of the road surface, and this road surface data 131 expresses, for example, a manhole 31.

[0062] Next, FIG. 3(4) and FIG. 3(5) show an example of the design drawing data generated by the design drawing related information generating unit. The design drawing data 141(1) is a side view, and for example, a horizontal length L3 and a vertical length H3 are set. The design drawing data component detecting unit 14 sets the length H3 and the length L3 as the set values ​​of the design drawing data 141(1). storage device 51 The data is stored as component element analysis data 521.

[0063] The design drawing data 141(2) is a top view, and for example, a horizontal length L3 and a depth length W3 are set. The design drawing data component detection unit 14 sets the length L3 and the length W3 as the dimensions of the design drawing data 141(2). storage device 51 The result is stored as component element analysis data 521.

[0064] The design plan data 141(1) is data showing a design plan of the component 3 when the underground space is viewed from the front side, and this design plan data 141(1) shows, for example, a manhole 31, a water pipe 33(1), and a water pipe 33(2). The design plan data 141(2) is data showing a design plan of the component 3 when the underground space is viewed from the road surface side (Z-axis direction), and this design plan data 141(2) shows, for example, a manhole 31 and a water pipe 33(1).

[0065] Next, the processing performed by the underground structure detection device will be described with reference to FIGS.

[0066] First, a series of processes from the acquisition of measurement data and design drawing data by the underground structure detection device to the output of analysis results of the components will be described with reference to FIG. First, the 3D model generation unit 11 and road surface data component detection unit 13 of the underground structure detection device 1 acquire measurement data from the vehicle 2, and the design drawing data component detection unit 14 acquires design drawing data 141 of the underground structure (S1). In step (S1), the three-dimensional model generating unit 11 acquires two-dimensional data 111 and three-dimensional data 121 as measurement data.

[0067] Next, the 3D model generation unit 11 of the underground structure detection device 1 generates a 3D model from the 2D data and 3D data, and the 3D model component detection unit 12, road surface data component detection unit 13, and blueprint data component detection unit 14 detect components based on the 3D model, components based on the road surface data, and components based on the blueprint data, respectively, and send them to the component information processing unit 15 as 3D detection data 201, road surface detection data 202, and blueprint detection data 203 (S2). Details of step (S2) will be described later with reference to FIG.

[0068] Next, the dimension adjustment unit 151 of the component information processing unit 15 makes the dimensions of the three-dimensional detection data 201, the road surface detection data 202, and the design drawing detection data 203 correspond to each other (S3). The dimension adjusting section will be described in detail later.

[0069] Next, the component element combined information generator 152 of the component element information processor 15 generates component element combined information of the three-dimensional detection data 201, the road surface detection data 202, and the blueprint detection data 203 (S4).

[0070] Next, the component compound information generating unit 152 of the component information processing unit 15 identifies the type of component for the component compound information of the three-dimensional detection data 201, road surface detection data 202, and blueprint detection data 203 (S5).

[0071] The component type identification process for the component compound information will be described in detail later with reference to FIGS.

[0072] Next, the component type and location information generating unit 153 generates the type and location information of the component as the analysis result of the component from the component compound information (S6). The component type and position information generating unit 153 may re-determine that a plurality of components that have been determined to be different components by the component identification process (S5) of the component compound information are the same component.

[0073] That is, for example, when a plurality of components that have been determined to be different components are connected in series, the component type and position information generating unit 153 determines that they are the same component. For example, when a water pipe that is diagonal to the X-axis direction and a water pipe that is aligned along the Y-axis direction are connected in series, Component composite information generation unit 152 In this case, the component type and position information generation unit 153 may, for example, detect a location where a water pipe diagonal to the X-axis direction and a water pipe aligned along the Y-axis direction are connected, and determine that they are one water pipe.

[0074] Next, the display unit 16 displays a screen including the type and position information of the component 3 (S7). The screen displayed by the display unit 16 will be described in detail later with reference to FIG.

[0075] Next, the details of the 3D model generation and component classification process, road surface classification process, and blueprint classification process will be explained using FIG. This process corresponds to S2 in FIG. First, the three-dimensional model generating unit 11 generates two-dimensional data 111 by displaying an underground cross section in gray scale according to the amplitude of the reflected wave of the underground exploration radar 231 of the radar 24 (S101).

[0076] 3 (n is a predetermined integer, and i is an integer satisfying 1≦i≦n) is image data showing a cross section of the underground that has been colored according to the amplitude of the reflected wave of the radar 24. For example, the coloring process may be such that the whiter the color, the greater the amplitude of the reflected wave of the radar.

[0077] In addition, the process of generating two-dimensional data may include a process of obtaining clear image data of the two-dimensional data 111, and for example, image processing such as noise removal or edge enhancement may be performed by applying a filter to the two-dimensional data 111.

[0078] Furthermore, the three-dimensional model generating unit 11 generates three-dimensional data 121 by using synthetic aperture processing on the measurement data measured by each underground exploration radar 231 (S102).

[0079] The process of generating three-dimensional data may include a process of obtaining clear image data. For example, image processing such as noise removal or edge enhancement may be performed by filtering the three-dimensional data 121.

[0080] Next, the three-dimensional model generation unit 11 synthesizes the two-dimensional data 111 and the three-dimensional data 121 (S103). In the data synthesis process S103, the three-dimensional model generation unit 11 adjusts the dimensions of each piece of two-dimensional data 111 based on, for example, the following (Equation 1) and (Equation 2). L1=s1*L … (Formula 1) H1=s2*H … (Formula 2) s1 and s2 are scales that indicate the correspondence between the three-dimensional data 121 and the two-dimensional data 111. That is, s1 is the horizontal magnification between the three-dimensional data 121 and the two-dimensional data 111, and s2 is the vertical magnification between the three-dimensional data 121 and the two-dimensional data 111.

[0081] After adjusting the dimensions of the 2D detection data and the 3D data, the 3D model generation unit 11 may set the luminance value of the point where the positional information matches as the intermediate value between the two. Alternatively, for example, a predetermined ratio value a may be determined, and the luminance values ​​of both data may be combined using a calculation formula such as (luminance value of a certain coordinate) = a × luminance value of the 3D data + (1-a) × luminance value of the 2D data so that the combined result does not exceed the original maximum value. Alternatively, both pieces of data in grayscale format may be newly assigned to one of the RGB channels of the color pixel (for example, the luminance value of the 3D data may be assigned to R, the luminance value of the 2D data to G, and B may be set to 0).

[0082] Next, the 3D model generation unit 11 generates a 3D model (S104). For example, the 3D model is generated by previously learning 3D training data as teacher data and 2D training processing parameters and 3D training processing parameters used when generating the 3D training data.

[0083] The three-dimensional model is a learning model for predicting to which type a component in three-dimensional data belongs, based on, for example, the three-dimensional data and three-dimensional processing parameters.

[0084] The 3D model component detection unit 12 classifies the components indicated by the synthetic data by inference based on the generated 3D model (S105). That is, using the 3D model and the AI ​​inference function, it predicts the type that the component in the 3D model is relatively likely to be among manhole 31, cavity 32, water pipe 33(1), and water pipe 33(2). The three-dimensional model may be generated in advance through learning.

[0085] Next, the 3D model component detection unit 12 performs a position / orientation estimation process to estimate shape information including position information, direction, and orientation of the classified components (S106). That is, as the position / orientation estimation process, the 2D data 111 and information on the classified components are read from the main memory 43, and the position / orientation estimation process is repeatedly performed. The position and orientation estimation process will be described in detail later with reference to FIG.

[0086] Meanwhile, the road surface data component detection unit 13 classifies the components 3 in the road surface data 131 (S111). In the road surface classification process S111, for example, an image processing method such as contour detection or color detection is used to classify the components shown in the road surface data 131 as manholes 31.

[0087] Then, the underground structure detection device 1 sends the classification results of the components in the road surface data 131 and the position information and shape information of the components 3 in the road surface data 131 to the component information processing unit 15 as road surface detection data 202.

[0088] Next, the design drawing data component detection unit 14 classifies the components in the design drawing data 141(1) and 141(2) (S121). In the design drawing classification process S121, for example, the component 3 in the design drawing data 141(1) is classified into a manhole 31 and a water pipe 33(1), and the components in the design drawing data 141(2) are classified into a manhole 31, a water pipe 33(1), and a water pipe 33(2).

[0089] Then, the underground structure detection device 1 sends the classification results of the components in the design drawing data 141(1) and 141(2) and the position information and shape information of the components in the design drawing data 141(1) and 141(2) to the component information processing unit 15 as design drawing detection data 203.

[0090] The blueprint classification process S121 may use, for example, an object detection framework to calculate label and position data for underground components. For example, at least one of Faster R-CNN (Region-based Convolutional Neural Network) and SSD (Single Short Multibox Detector) may be used as the object detection framework. The blueprint classification process S121 is not limited to using methods such as Faster R-CNN and SSD to calculate component type, position, and shape information. It may also use OCR to detect component type, position, and shape information.

[0091] Next, the process of estimating the position and orientation of the constituent elements will be described with reference to FIGS. 6 to 8B. This is the process corresponding to S106 in FIG. First, in the position and orientation estimation process, it is determined whether the component is a "pipe" (S701). Here, "pipe" is a concept that includes not only water pipes and gas pipes, but also buried objects protected by tubular objects, such as electric wires, and is intended to be long, extending objects that are installed to connect predetermined locations.

[0092] Next, based on the results of the classification process of the components in S105, the direction of the pipe is confirmed, and it is determined whether the extension direction is offset from the direction of vehicle movement (X-axis direction) by more than a predetermined threshold angle in the X-axis direction, or whether it is offset from the width direction of the vehicle (Y-axis direction) by more than a predetermined threshold angle in the Y-axis direction (S702). If it is offset (S702: Yes), proceed to S703; if not (S702: No), proceed to S704. A prerequisite for this determination is that most of these buried objects are characterized by being installed in the direction in which the vehicle 2 is traveling, i.e., under the road, parallel to the extension direction of the road, or in the width direction of the vehicle, i.e., perpendicular to the extension direction of the road. Therefore, if this condition is not met, it is considered that the pipe orientation detection is likely to be incorrect as a result of the classification of the components in S105.

[0093] It should be noted that the determination in S702 need not be limited to the example of determining based on the angles of the components. For example, the process may proceed to S703 when the road surface data 131 is referenced and it is determined that the road surface contains moisture (the color is darker). Also, for example, the process may proceed to S703 when the design drawing data 141 is referenced and the posture of the buried object at that location differs. Also, for example, if the detection data is accompanied by data indicating location information such as GPS, the process may proceed to S703 when information from the same location in a past inspection is referenced and there is a deviation of more than a predetermined threshold.

[0094] Next, the three-dimensional model component detection unit 12 performs position and orientation correction processing using the two-dimensional data (S703). The details of this process will be explained later with reference to FIG.

[0095] Next, the three-dimensional model component detection unit 12 stores the position and orientation estimation results as component analysis data 521 (S704).

[0096] Next, the position and orientation correction process using two-dimensional data will be described in detail with reference to FIGS. 7, 8A, and 8B. This is the process corresponding to S703 in FIG. First, the 3D model component detection unit 12 determines whether or not the 2D image within the target range indicated in the classification result of the component contains image features (image features of a crossing pipe) that appear when the component object is in the crossing direction (x direction) (S801).If the feature is present (S801: Yes), the process proceeds to S802; if the feature is not present (S801: No), the process proceeds to S805.

[0097] For example, as shown in FIG. 8A, if two-dimensional data 111(1) through 111(4) are included in the buried object area estimated by the three-dimensional classification process, the features of the crossing pipe are checked sequentially from the target area of ​​two-dimensional data 111(1) through two-dimensional data 111(4). If even one feature is detected, the result of S801 is determined to be Yes. The image feature of the crossing pipe includes a pattern with a quadratic curve shape, where the brightness shading is upwardly convex, as shown in FIG. 8A. The detection method is, for example, by comparing the degree of match with a pre-prepared pattern. Because the size and width of the pattern vary depending on the material and size of the buried object, multiple pre-prepared patterns may be prepared. Note that the X and Z coordinates of the detected feature may be recorded in the component analysis data 521. Alternatively, the features of all crossing pipes included in the two-dimensional image within the target area may be detected.

[0098] Next, the three-dimensional model component detection unit 12 counts the number of pieces of two-dimensional data in which the features of the crossing pipe overlap and straddle the same position (S802).

[0099] For example, starting with two-dimensional data 111(1), the system determines whether image features of a crossing pipe seen in two-dimensional data 111(1) are found in the two-dimensional data 111(2)... and counts the number of such features. If a crossing pipe feature is found at approximately the same position (within a predetermined threshold range) as the coordinates where the feature was detected in two-dimensional data 111(1), the continuity count is increased by 1, the next adjacent image is checked, and the count is incremented. If a crossing pipe feature is not found at the same coordinate position, the system stops counting the continuity count and terminates this process. Taking into account cases where crossing pipes are installed diagonally, it may be determined that the crossing pipe regions are contiguous if they are shared by more than a predetermined threshold or percentage.

[0100] Next, the 3D model component detection unit 12 determines whether the number of consecutive images determined to be in the same position is equal to or greater than a predetermined threshold (S803). If the number is equal to or greater than the predetermined threshold (S803: Yes), the process proceeds to S804. If the number is less than the predetermined threshold (S803: No), the process ends. The predetermined threshold may be specified as a specific number such as 3, or may be specified as a percentage such as 80% of the number of 2D images to be confirmed. The reason for determining the number of consecutive images in this manner is that, in the case of a crossing pipe, as shown in FIG. 8A, it is likely that the pipe extends across the entire width of the road, and if this image feature is seen only sporadically, it is likely that the pipe is a block-shaped buried object.

[0101] Next, the 3D model component detection unit 12 corrects the position and orientation information of the components based on the detection results from the 3D classification process, based on the information of the 2D data having image features in the transverse direction (S804), and ends the process.

[0102] For example, when image features of a crossing pipe are detected consecutively from two-dimensional data 111(1) to two-dimensional data 111(4), the trajectory of the X and Z coordinate values ​​of the brightest point among the features of the crossing pipe from two-dimensional data 111(1) to two-dimensional data 111(4) may be linearly approximated to determine the posture of the crossing pipe. Also, when multiple corresponding crossing pipes are detected, the information may be replaced with information about multiple crossing pipes.

[0103] If the result in S801 is No, the 3D model component detection unit 12 determines whether the 2D data within the range contains image features of a longitudinal pipe (S805). If the image features are present (S805: Yes), the process proceeds to S806. If the image features are not present (S805: No), the process ends. The image features of a longitudinal pipe include, for example, a pattern in which low-brightness points are sandwiched between high-brightness points, as shown in FIG. 8B. The detection method is, for example, to compare the degree of match with a pre-prepared pattern. Upon detection, the X and Z coordinates may be recorded in the component analysis data 521. Alternatively, all longitudinal pipe features contained in the 2D data within the target range may be detected.

[0104] Next, the three-dimensional model component detection unit 12 counts the number of pieces of two-dimensional data in which the features of the longitudinal section pipe overlap and straddle the same position (S806).

[0105] For example, starting with two-dimensional data 111(1), the system determines whether image features of a longitudinal pipe seen in two-dimensional data 111(1) are found in the subsequent two-dimensional data 111(2)... and counts the number of such features. If a longitudinal pipe feature is found at approximately the same position (within a predetermined threshold range) as the coordinates at which a feature was detected in two-dimensional data 111(1), the continuity count is increased by 1, the next adjacent image is checked, and the count is incremented. If no longitudinal pipe feature is found at the same coordinate position, the system stops counting the continuity count and terminates this process. Taking into account cases where longitudinal pipes are installed at an angle, it may be determined that the longitudinal pipes are contiguous if the longitudinal pipe regions are common to a predetermined threshold or percentage or more.

[0106] Next, the 3D model component detection unit 12 determines whether the number of consecutive images determined to be in the same position is equal to or greater than a predetermined threshold (S807). If it is equal to or greater than the predetermined threshold (S807: Yes), the process proceeds to S808; if it is less than the predetermined threshold (S807: No), the process ends.

[0107] The reason for making the determination based on the number of consecutive sheets is that, as shown in FIG. 8B, the longitudinal pipe extends in the direction of travel of the vehicle 2, and therefore buried objects with a large number of consecutive sheets in the vertical direction are judged to be of a plate-like shape rather than a pipe-like shape. Therefore, the predetermined threshold may be set to a fixed value, such as 2. Furthermore, when considering extension in the diagonal direction, the search range may be divided into regions by a predetermined number of X coordinates, and it may be confirmed that the number of consecutive sheets in all divided regions is equal to or less than the predetermined threshold.

[0108] Next, the 3D model component detection unit 12 corrects the position and orientation information of the components based on the detection results from the 3D classification process, based on the information of the 2D data having image features in the longitudinal direction (S808), and ends the process.

[0109] For example, when image features of a longitudinal pipe are detected consecutively in two-dimensional data 111(2) and two-dimensional data 111(3), the posture of the buried object is determined by connecting the endpoints of the area in which pixels with a brightness higher than a predetermined value are consecutive in the extension direction among the features of the longitudinal pipe in two-dimensional data 111(2) and two-dimensional data 111(3). Also, for example, when multiple longitudinal pipes are detected in the corresponding area, they may be replaced with multiple crossing pipe information.

[0110] Next, the dimension adjustment process will be described in detail. This is the process corresponding to S3 in FIG.

[0111] As a dimension adjustment process, the dimension adjustment unit 151 makes the coordinate systems of each piece of data correspond to each other before adjusting the dimensions of each piece of three-dimensional data 121, road surface data 131, and design drawing data 141. That is, the dimension adjustment unit 151 makes the road surface data 131 and the design drawing data 141 correspond to the X, Y, and Z axis directions of the three-dimensional data 121.

[0112] The dimension adjustment unit 151 adjusts the dimensions of the road surface data 131 based on, for example, the following (Equation 3) and (Equation 4). L2=r1*L … (Formula 3) H2=r2*H … (Formula 4) r1 and r2 are scales that indicate the correspondence between three-dimensional data 121 and road surface data 131. That is, r1 is the magnification between three-dimensional data 121 and road surface data 131 in the horizontal direction, and r2 is the magnification between three-dimensional data 121 and road surface data 131 in the vertical direction.

[0113] Furthermore, the dimension adjustment unit 151 adjusts the dimensions of the design drawing data 141(1) based on, for example, the following (Equation 5) and (Equation 6). L3=m1*L … (Formula 5) H3=m2*H … (Formula 6) m1 and m2 are scales that indicate the correspondence between the three-dimensional data 121 and the design drawing data 141(1). That is, m1 is the horizontal magnification between the three-dimensional data 121 and the design drawing data 141(1), and m2 is the vertical magnification between the three-dimensional data 121 and the design drawing data 141(1).

[0114] Furthermore, the dimension adjustment unit 151 adjusts the dimensions of the design drawing data 141(2) based on, for example, the above-described (Equation 5) and the following (Equation 7). Note that when the horizontal dimension of the design drawing data 141(2) is set to a length other than the length L3, the dimension adjustment unit 151 may adjust the dimensions of the design drawing data 141(2) using an equation other than (Equation 5). W3=m3*W…(Equation 7) m3 is a scale indicating the correspondence between the three-dimensional data 121 and the design drawing data 141(2). That is, m3 is the magnification in the depth direction between the three-dimensional data 121 and the design drawing data 141(2).

[0115] The coordinate integration process of the dimension adjustment unit 151 will be described using as an example the coordinates (x1, y1) of the manhole 31 shown in Fig. 3 of the road surface data 131. The dimension adjustment unit 151 adjusts the coordinates (x1, y1) of the manhole 31 in the road surface data 131 to other coordinates corresponding to the three-dimensional data 121. The dimension adjustment unit 161 adjusts the coordinates of the manhole 31 from (x1, y1) to (x1 / r1, y1 / r2) based on, for example, the above (Equation 3) and (Equation 4).

[0116] Next, the component identification process for the component compound information will be described with reference to FIGS. This is the process corresponding to S5 in FIG.

[0117] The component compound information generator 152 selects a same-type pair between the components (S51). For example, a same-type pair of components 3 classified as the same type is selected between the components in the two-dimensional data 111 and the three-dimensional data 121. Specifically, as shown in Fig. 10, the component compound information generator 152 selects, for example, a water pipe 33(1) that is common to the two-dimensional data 111(i) and the three-dimensional data 121 as a pair of components.

[0118] Next, the component type identification unit 154 of the component compound information generation unit 152 calculates a same-type evaluation index, which is an evaluation index between the same-type pairs of components selected by the same-type pair selection process S51 (S52). For example, the component type identification unit 154 calculates the same-type evaluation index between the water pipe 33(1) in the two-dimensional data 111(i) and the water pipe 33(1) in the three-dimensional data 121 shown in Figure 10 to be "0.8".

[0119] It should be noted that the component type identification unit 154 may use, for example, the degree of overlap between the selected component pairs (3D IoU: 3-Dimensional Intersection over Union) when calculating the evaluation index. Component type identification unit 154 The method is not limited to calculating the evaluation index based on the degree of overlap between the selected component pairs, but may also take a method such as evaluating the same-type evaluation index higher when the numerical values ​​of the corresponding dimensions of each component are highly similar.

[0120] The component type identification unit 154 of the component compound information generation unit 152 determines whether the same-type pairs indicate the same component based on the same-type evaluation index (S53). For example, if the same-type evaluation index is greater than a predetermined threshold (S53: Yes), the component compound information generation unit 152 determines that the same-type pairs indicate the same component (S54).

[0121] The predetermined threshold is a value for determining whether a pair of components indicate the same component. The predetermined threshold may be input by the user to a component information display screen (described later) of the underground structure detection device 1. For example, when the predetermined threshold is "0.7", the component type identification unit 154 determines that the water pipe 33(1) in the two-dimensional data 111(i) shown in FIG. 10 and the water pipe 33(1) in the three-dimensional data 121 indicate the same component.

[0122] The predetermined threshold may be set for each component. An input field for the predetermined threshold may be displayed on a component information display screen (described later) shown on the display device 61. The user may change the predetermined threshold for each underground structure detection process to adjust the accuracy of determining the type of component.

[0123] If the same-type evaluation index is equal to or less than a predetermined threshold (S53: No), the component type identification unit 154 determines that the same-type component pairs indicate different components, and performs a different-type pair discrimination process (S55). In this case, the component compound information generation unit 152 determines whether pairs of components classified into different types (hereinafter referred to as "different-type pairs") indicate the same component. Details of the different-type pair discrimination process will be described later with reference to FIG. 11.

[0124] If the same-type evaluation index is equal to or less than a predetermined threshold (S53: No), the component type identification unit 154 may select a same-type pair by executing the same-type pair selection process S51 again. That is, in this case, if the same-type evaluation index is equal to or less than a predetermined threshold (S53: No), the component type identification unit 154 selects one of the same-type pairs determined in S53.

[0125] Then, the component type identification unit 154 selects a new same-type pair by newly selecting a component that is classified into the same type as one of the selected same-type pairs. If there is no component that can be selected for the new same-type pair, the component compound information generation unit 152 may execute a different-type pair determination process S55.

[0126] Finally, the component type identification unit 154 records the analysis results regarding the component type in the component analysis data (S56).

[0127] Next, the process of determining pairs of different types of components will be described with reference to FIG. This is the process corresponding to S55 in FIG.

[0128] The component type identification section 154 of the component compound information generation section 152 selects a heterogeneous pair (S551). In this process, the component type identification unit 154, for example, sets one of the same-type pair selected in the same-type pair selection process S51 as a predetermined component 3, and sets the other as a corresponding different-type component.

[0129] The component type identification unit 154 calculates a heterogeneous evaluation index, which is an evaluation index of the heterogeneous pair (S552). If the evaluation index is greater than a predetermined threshold (S553: ​​Yes), the component compound information generation unit 152 determines which of the heterogeneous pair should belong as a component of the component compound information (S554 to S556). A heterogeneous evaluation index is an index that evaluates something that has the opposite properties to a homogeneous evaluation index.

[0130] The component type identification unit 154 acquires an accuracy index from the 3D model component detection unit 12, the road surface data component detection unit 13, and the blueprint data component detection unit 14 that detected the heterogeneous pairs (S554). The accuracy index is an index that indicates the accuracy of the component classification results for, for example, the 3D model detection data 201 output by the 3D model component detection unit 12, the road surface detection data 202 output by the road surface data component detection unit 13, and the blueprint data output by the blueprint data component detection unit 14. The accuracy index is generated, for example, in the 3D model-based component classification process S105, the road surface data-based component classification process S111, and the blueprint data-based component classification process S121 shown in FIG. 5, and sent to the component information processing unit 15. The accuracy index can be calculated by obtaining a numerical value that highly evaluates good results according to a learning model based on past measurement data.

[0131] Next, the component type identification unit 154 determines a priority order based on the acquired accuracy index (S555). The component type identification unit 154, for example, determines the classification of the component with a higher accuracy index as correct (S556). That is, the component compound information generation unit 152 determines, for example, the component with a low classification accuracy in the heterogeneous pair as a misclassification, and determines the classification of the component with a high classification accuracy in the heterogeneous pair as correct, and sets these as the analysis results.

[0132] If the heterogeneous evaluation index is smaller than the predetermined threshold (S553: ​​No), the component type identification unit 154 determines that the components are different (S557). Note that if the heterogeneous evaluation index is smaller than the predetermined threshold (S553: ​​No), the component type identification unit 154 may select a new heterogeneous pair and perform the heterogeneous evaluation index determination S553 again.

[0133] That is, the component type identification unit 154 selects a new component that is classified into a different type from the specified component from among the heterogeneous pairs of the specified component and other components, and performs the determination S553 of the heterogeneous evaluation index again using the newly selected heterogeneous pair.

[0134] Next, an example of the component information display screen will be described with reference to FIG. The component information display screen 51 is the screen output in S7 of FIG.

[0135] The display unit 16 displays, for example, the three-dimensional model state display map 35 and the explanation column 36 of the components 3 (manhole 31, cavity 32, water pipe 33) on the display device 61 as a component information display screen 51. The component information display screen 51 comprises a three-dimensional state display map 35 and a component information display field 36 .

[0136] The three-dimensional status display map 35 is an area that displays the shape and positional relationship of components in three-dimensional space within a three-dimensional bounding box based on component result data 522, which is the result of analysis by the underground structure detection device 1.

[0137] The component information display field 36 is an area that displays the type and coordinate data of each component. For example, the component information display field 36 displays the type of buried object in the format: [center X coordinate, center Y coordinate, center Z coordinate, length, width, height]. Furthermore, in addition to the type and coordinates of the component, the component information display field 36 may also display calculation results such as the length, width, and volume of the buried object, as well as other supplementary information.

[0138] As described above, the underground structure detection device 1 includes the component information processing unit 15, and thereby detects the components of an underground structure based on the two-dimensional data 111, the three-dimensional data 121, the road surface data 131, and the design drawing data 141. This allows the underground structure detection device 1 to comprehensively detect the components of an underground structure, and improves the detection accuracy of the underground structure compared to conventional methods.

[0139] In particular, the underground structure detection device 1 performs processing to correct the position and orientation of components based on the image features of the components from the two-dimensional data 111 for the composite data obtained by combining the two-dimensional data 111 and the three-dimensional data 121, thereby making it possible to obtain accurate position and orientation.

[0140] Furthermore, the underground structure detection device 1 can suppress dimensional discrepancies between data by performing a dimension adjustment process, thereby improving the accuracy of detecting components based on the 3D detection data 201, the road surface data 134, and the design drawing data 144.

[0141] Furthermore, the underground structure detection device 1 can automatically determine the type and three-dimensional location of buried objects through classification processing of the components, which reduces the amount of work required compared to manually classifying the components.

[0142] Furthermore, the underground structure detection device 1 can prevent data of components with incorrect classification results from being imported by determining which information of a heterogeneous pair should be prioritized based on the accuracy index, thereby enabling the underground structure detection device 1 to improve the accuracy of classification of components.

[0143] Furthermore, the underground structure detection device 1 displays the type and position of the constituent element on the constituent element information display screen 51. This makes it easier for the user to understand what is located at what position in the three-dimensional space.

[0144] In addition, the underground structure detection device 1 is not limited to generating information on the type and position of components 3 based on 2D radar data, 3D radar data, road surface data, and blueprint data, but may also detect components of an underground structure by performing processing similar to that described above even when only 3D radar data and road surface data are input.

[0145] [Embodiment 2] Hereinafter, the second embodiment of the present invention will be described with reference to FIG. The configuration and processing of the underground structure detection device of this embodiment are generally similar to those of the first embodiment, so the differences will be mainly explained.

[0146] This embodiment differs from the first embodiment only in the position and orientation estimation process using two-dimensional data in S703 of Fig. 6, and performs the process of Fig. 13 instead of the process of Fig. 7 of the first embodiment. The difference from the process of Fig. 7 of the first embodiment is that both the crossing direction and the traveling direction are determined.

[0147] S801 to S803 are the same as those in the first embodiment.

[0148] In this embodiment, if No is selected in S803, the process proceeds to S805, and if Yes is selected in S803, the crossing direction information is recorded in the component element analysis data 521 (S810).

[0149] S805 to S807 are the same as those in the first embodiment.

[0150] The same processing is performed, but in this embodiment, if No is selected in S805, the process proceeds to S830.

[0151] If the answer is Yes in S807, the information on the longitudinal direction is recorded in the component element analysis data 521 (S820).

[0152] Finally, based on the transverse direction information of S810 and the longitudinal direction information of S820, and on information regarding the position and orientation of the components, the correction results are reflected. For example, if there are image features of a continuous transverse pipe in the two-dimensional data 111(2) to (4) and image features of a longitudinal pipe in the two-dimensional data 111(1), the results of both are adopted for correction. Also, for example, if an assumed extension of the transverse pipe intersects with part of the detection result of the longitudinal pipe, information indicating a high possibility of an intersection or branch occurring along the way may be included in the detection result.

[0153] In the correction process using 2D data in the first embodiment, only correction was performed in either the transverse or longitudinal direction, but in the procedure of this embodiment, image features in the transverse and longitudinal directions are captured, and in some cases corrections in both the transverse and longitudinal directions are performed. This enables the 3D model component detection unit 12 to accurately detect buried pipes with L- or T-shaped cross sections where the longitudinal and transverse sections intersect. [Explanation of symbols]

[0154] 1... underground structure detection device, 11... 3D model generation unit, 12... 3D model component detection unit, 13... road surface data component detection unit, 14... blueprint data component detection unit, 15 component information processing unit, 16... display unit, 2... vehicle, 21... camera data processing unit, 22... radar data processing unit, 23... measuring device, 24... radar, 3... component, 31... manhole, 32... cavity, 33... water pipe

Claims

1. An underground structure detection device that detects information including the components that make up an underground structure and the types of the components and position / posture information based on a plurality of two-dimensional data showing cross sections of the ground and three-dimensional data showing an underground structure, wherein a first judgment condition for composite data obtained by combining the two-dimensional data and the three-dimensional data is that image features shown by the components in the two-dimensional data are included in consecutive images in a first direction to a degree equal to or greater than a predetermined first threshold, and when the first judgment condition is met, the device corrects the position / posture information of the components represented by the composite data based on the two-dimensional data connected in the first direction, An underground structure detection device characterized in that a second judgment condition is that the image features indicated by the component in the two-dimensional data are included in consecutive images below a predetermined second threshold in a second direction, and when the second judgment condition is met, the position and orientation information of the component represented by the composite data is corrected based on the two-dimensional data connected in the second direction.

2. the component is a tube; the first direction is a direction transverse to a road; 2. The underground structure detection device according to claim 1, wherein the second direction is a longitudinal direction relative to a road.

3. 2. The underground structure detection device according to claim 1, wherein a determination is made based on both the first determination condition and the second determination condition.

4. the component is a tube; An underground structure detection device as described in claim 1, characterized in that when the judgment is true for both the first judgment condition and the second judgment condition, the output includes information that the type of pipe is a T-shaped pipe or an L-shaped pipe.

5. Furthermore, a means for photographing the state of the road surface to obtain road surface data; means for obtaining design data from a design drawing relating to an underground structure; The underground structure detection device described in claim 1, characterized in that the type and position of the components are determined based on three-dimensional detection data that detects components from composite data that combines the two-dimensional data and the three-dimensional data, road surface detection data that detects components from the road surface data, and blueprint detection data that detects components from the blueprint data.

6. An underground structure detection method for an underground structure detection device that detects information including components constituting an underground structure, types of the components, and position / attitude information of the components, based on a plurality of two-dimensional data showing cross sections of the ground and three-dimensional data showing the underground structure, comprising: creating composite data by combining the two-dimensional data and the three-dimensional data; determining, as a first determination condition, whether or not image features indicated by the components in the two-dimensional data are included in continuous images in a first direction at a rate equal to or greater than a predetermined first threshold value, for composite data obtained by combining the two-dimensional data and the three-dimensional data; When the first determination condition is met, correcting position and orientation information of the constituent element represented by the composite data based on the two-dimensional data connected in the first direction; determining, as a second determination condition, whether or not image features indicated by the components in the two-dimensional data are included in a series of images equal to or smaller than a predetermined second threshold value along a second direction, for composite data obtained by combining the two-dimensional data and the three-dimensional data; An underground structure detection method characterized by having a step of correcting the position and orientation information of the components represented by the composite data based on the two-dimensional data connected in the second direction when the second judgment condition is met.

7. the component is a tube; the first direction is a direction transverse to a road; 7. The underground structure detection method according to claim 6, wherein the second direction is a longitudinal direction relative to a road.

Images