JP2014190962A - Data analysis device, data analysis method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To highly accurately and automatically detect a profile of a tunnel on the basis of three-dimensional coordinate data of a cloud of points extracted from a feature surface.SOLUTION: Subspace setting means 20 divides a profile neighborhood space that is along a cut plane and has a preset width in the depth direction orthogonal to the cut plane, into a grid shape, and sets a plurality of subspaces. Edge search means 22 searches for, within the cut plane for each subspace, a belt-like area having a predetermined width that is an edge approximate area in which a cloud of points projected on the cut plane are accumulated equal to or more than a criteria preset in the area and the cloud of points in the area has distribution width equal to or more than a threshold preset in the depth direction. The edge search means searches for a directional line along a distribution in the cut plane of the cloud of points belonging to the edge approximate area on the cut plane as the edge as a position of the surface in the tunnel in the cut plane.

Description

  The present invention relates to a data analysis apparatus, a data analysis method, and a program for detecting a cross-sectional shape of a tunnel based on three-dimensional coordinate data of a point group extracted from a feature surface.

  Conventionally, as a tunnel shape measurement method, a laser range finder is installed in the tunnel, the laser beam is sequentially irradiated to the inner peripheral surface of the tunnel at predetermined angles, and the time from laser emission to reception of reflected light The distance from the origin on which the distance measuring device is placed to the irradiation point is measured, and each irradiation point is converted into position data based on the measured distance and angle, thereby obtaining the cross-sectional shape of the tunnel. (Patent Documents 1 and 2).

  On the other hand, a mobile mapping system has recently been developed as a technique for acquiring three-dimensional point cloud data representing the shape of a feature along a road using a laser scanner mounted on a vehicle (Patent Document 3). In this system, a laser scanner mounted on an automobile irradiates a laser obliquely downward or obliquely upward from the top of the vehicle body. The optical axis of the laser is scanned laterally, a laser pulse is emitted at every minute angle within the scanning angle range, and the distance to the point where the laser pulse is reflected is measured. At that time, the laser emission direction, time, and the position / orientation of the vehicle body are measured. From these measurement data, point group data representing the three-dimensional coordinates of the reflection point of the laser pulse is obtained.

JP 2003-65555 A JP 2001-91249 A JP 2009-204615 A

  In the conventional tunnel shape measuring method, it is difficult to efficiently measure a wide area when a laser distance measuring device is installed in a tunnel and measured. In this regard, it has also been proposed to perform measurement while moving the laser distance measuring device with a carriage or the like. However, in this method of measuring while moving, if the carriage is stopped at each measurement point, it takes time and labor to measure at a short moving distance and obtain a cross-sectional shape with high density in the longitudinal direction of the tunnel. Take it. Further, if the laser irradiation angle is continuously changed in the transverse direction of the tunnel while being moved, the locus of the irradiation point becomes spiral. If the moving speed is low, it is possible to approximately grasp the cross-sectional shape of the tunnel from the helical shape measurement results, but it is not suitable for a wide range of measurements, while if the moving speed is high, the helical shape is measured. It becomes difficult to grasp the cross-sectional shape of the tunnel from the result.

  In this regard, the mobile mapping system can acquire a point cloud from a wide range around the travel route of the mounted vehicle. Since the point cloud is obtained with high density and distributed on the surface of the feature, the shape of the feature can be grasped based on the point cloud data. Specifically, an arbitrary cross-sectional shape of the tunnel can be reconstructed from the point cloud data. On the other hand, analysis such as interpretation of features based on point cloud data is performed manually using an editing tool or the like in 3D CAD, and automation is under development.

  The present invention provides a data analysis apparatus, a data analysis method, and a program capable of automating the operation of detecting the cross-sectional shape of a tunnel based on three-dimensional coordinate data of a point cloud extracted from the inner surface of the tunnel. With the goal.

  (1) A data analysis apparatus according to the present invention is an apparatus that detects a cross-sectional shape of a tunnel existing in a target space based on three-dimensional coordinate data of a point cloud extracted from a feature surface, and is a vertical cutting plane. Is virtually set in the target space, and a cross-section near space having a predetermined width in the depth direction along the cutting plane and perpendicular to the cutting plane in the target space is divided into a lattice shape and A subspace setting means for setting a plurality of subspaces arranged two-dimensionally along the cutting plane; and each subspace is set as an analysis unit space, and for each analysis unit space, in the tunnel in the cutting plane Edge search means for searching for an edge which is the position of the surface, and the edge search means is a band-like area having a predetermined width, and the point cloud projected onto the cutting plane is the area concerned. The edge registration area is searched for in the cutting plane for an edge registration area that has a distribution width equal to or greater than a preset threshold value in the depth direction, and the point group in the area gathers more than a preset reference in the depth direction. A direction line along the distribution in the cutting plane of the point group to which it belongs is obtained as the edge.

  (2) Another data analysis apparatus according to the present invention further includes edge tracking means for tracking the edge detected in the partial space outside the partial space, and the edge tracking means includes the analysis When the edge is detected from the unit space, a tracking space is set as the new analysis unit space adjacent to the analysis unit space in the direction of the extension line of the direction line corresponding to the edge, and the edge search means Searches the tracking space for the edge along the extension line.

  (3) In still another data analysis apparatus according to the present invention, the tracking space has a rectangular shape in the cutting plane, and a midpoint of one side of the rectangle is connected to an end of the detected edge. A rectangular parallelepiped-shaped space, and the edge tracking means sets, as the analysis unit space, the tracking space in which the center line of the rectangle passing through the midpoint coincides with the extension line of the edge, If the edge is not detected, the tracking space where the point cloud is most concentrated is searched as the tracking space as the analysis unit space by changing the direction of the center line in the cutting plane.

  (4) The data analysis device according to (2) or (3) further includes, for each partial space, one of the cross-sectional shapes of the tunnel from an edge group obtained by tracking starting from the edge of the partial space. A partial cross-sectional shape generating means for generating a partial cross-sectional shape to be a part can be provided.

  (5) In the data analysis device described in (4) above, when the edge tracking unit uses the direction line detected in the analysis unit space as an edge candidate, and the edge candidate intersects the partial cross-sectional shape The edge candidate may not be a new edge.

  (6) In the data analysis apparatus, the edge tracking means uses the direction line detected in the analysis unit space as an edge candidate, and the edge candidate tracked from the edge of the partial space When the edge or the other edge tracked from the edge intersects, the edge candidate may not be the new edge.

  (7) The data analysis device described in (4) to (6) further includes a partial cross-sectional shape connection unit that connects between two end portions of the partial cross-sectional shape, and the edge search unit includes the edge search unit. Corresponding to the edge, defining an edge vicinity region that is a band-like region having a predetermined width with the edge as a center line and having a length corresponding to the edge, and the partial cross-sectional shape connecting means includes: When the extension line of the edge of one end of the partial cross-sectional shape intersects the edge vicinity region of the edge of the other end of the partial cross-sectional shape, the two end portions are connected. Can do.

  (8) The data analysis device described in (4) to (7) further includes a partial cross-sectional shape connecting unit that connects two end portions of the partial cross-sectional shape, and the edge search unit includes the end search unit. When the distance between the parts is equal to or less than a predetermined threshold value, the two end parts can be connected.

  (9) In the data analysis device described in (2) to (8) above, the edge tracking unit may start tracking from both ends of the edge detected in the partial space.

  (10) A data analysis method according to the present invention is a method for detecting a cross-sectional shape of a tunnel existing in a target space based on three-dimensional coordinate data of a point cloud extracted from a feature surface, and a vertical cutting plane. Is virtually set in the target space, and a cross-section near space having a predetermined width in the depth direction along the cutting plane and perpendicular to the cutting plane in the target space is divided into a lattice shape and A subspace setting step for setting a plurality of subspaces arranged two-dimensionally along the cutting plane; and setting each subspace as an analysis unit space, and for each analysis unit space, in the tunnel in the cutting plane An edge search step for searching for an edge which is a surface position, and the edge search step is a band-like region having a predetermined width and is projected onto the cutting plane. Search for an edge registration region in the cutting plane where the point group is gathered over a predetermined reference in the region, and the point group in the region has a distribution width equal to or greater than a predetermined threshold in the depth direction, A direction line along the distribution in the cutting plane of the point group belonging to the edge registration area is obtained as the edge.

  (11) A program according to the present invention is a program for causing a computer to perform data analysis for detecting a cross-sectional shape of a tunnel existing in a target space based on three-dimensional coordinate data of a point group extracted from a feature surface. The computer virtually sets a vertical cutting plane in the target space, and has a preset width in the depth direction along the cutting plane of the target space and perpendicular to the cutting plane. A subspace setting means for setting a plurality of subspaces that are two-dimensionally arranged along the cutting plane by dividing the near-section space into a lattice shape, and setting each subspace as an analysis unit space, For each unit space, the edge search means functions as an edge search means for searching for an edge that is a position of the surface in the tunnel in the cutting plane, A band-shaped area having a predetermined width, and the point group projected onto the cutting plane is gathered in a predetermined area or more in the area, and the point group in the area is preset in the depth direction. An edge registration area having a distribution width equal to or greater than a threshold is searched in the cutting plane, and a direction line along the distribution in the cutting plane of a point group belonging to the edge registration area is obtained as the edge.

  ADVANTAGE OF THE INVENTION According to this invention, the operation | work which detects the cross-sectional shape of a tunnel based on the three-dimensional coordinate data of the point group extracted from the inner surface of a tunnel is attained.

1 is a block diagram showing a schematic configuration of a tunnel shape extraction system according to an embodiment of the present invention. It is a typical perspective view which shows a reference line and a cutting plane. It is a general | schematic flowchart of the tunnel shape extraction process by the tunnel shape extraction system which concerns on embodiment of this invention. It is a schematic diagram explaining the partial space set along a cutting plane. It is a rough processing flow figure of a partial space setting means. It is an outline processing flow figure of an edge search means. It is a schematic diagram explaining the edge search process in a base point space. It is an outline processing flow figure of an edge tracking means and a figure registration means. It is a process flow figure of the outline of the continuous tracking process by an edge tracking means. It is a process flow figure of the outline of the continuous tracking process by an edge tracking means. It is a schematic diagram of a cutting plane showing an example of a general edge extracted in a partial space and a continuous tracking process using the edge as a base point. It is a schematic diagram explaining the 1st judgment method of the connection possibility of a continuous line segment pair. It is a schematic diagram explaining the 2nd determination method of the possibility of connection of a continuous line segment pair. It is a schematic flowchart of the edge fine adjustment processing by the edge shape adjusting means. It is a schematic diagram explaining the fine adjustment process of the edge by the edge shape adjusting means. It is a schematic diagram of the cutting plane which shows the general | schematic edge extracted in the level | step-difference part beside the road of a tunnel. It is a schematic diagram explaining the process which produces | generates a detailed edge from the general edge of FIG. It is a schematic diagram of the cutting plane which shows the detailed edge produced | generated from the general edge of FIG.

  Hereinafter, embodiments of the present invention (hereinafter referred to as embodiments) will be described with reference to the drawings.

  Hereinafter, a tunnel shape extraction system 2 according to an embodiment of the present invention (hereinafter referred to as an embodiment) will be described with reference to the drawings. This system is a data analysis device that detects a surface shape in a tunnel based on point cloud data representing a three-dimensional shape of a feature surface. The point cloud data is acquired by, for example, a laser scanner mounted on a vehicle traveling on the ground like the above-described mobile mapping system. In addition, measurement may be performed by installing a laser scanner on the ground. In order for the point cloud data to represent the three-dimensional shape of the feature surface, it is necessary to perform laser scanning at a density corresponding to the unevenness of the feature surface and the scale of the step. In this regard, laser scanning from the height of a vehicle, tripod, etc., for example, counts the shape of walls, roadside steps, tunnel attachments, etc. in a space that is about the size of a tunnel with multiple lanes for automobiles. A scanning density that can be captured with an accuracy of about a centimeter can be realized.

  FIG. 1 is a block diagram showing a schematic configuration of the tunnel shape extraction system 2. The system includes an arithmetic processing device 4, a storage device 6, an input device 8, and an output device 10. As the arithmetic processing device 4, it is possible to make dedicated hardware for performing various arithmetic processing of this system, but in this embodiment, the arithmetic processing device 4 uses a computer and a program executed on the computer. Built.

  A CPU (Central Processing Unit) of the computer constitutes the arithmetic processing unit 4, and a partial space setting means 20, edge search means 22, edge tracking means 24, surface element shape determination means 26, figure registration means 28, edge connection, which will be described later. It functions as the means 30 and the edge shape adjusting means 32.

  The storage device 6 is composed of a hard disk or the like built in the computer. The storage device 6 causes the arithmetic processing unit 4 to function as the partial space setting means 20, edge search means 22, edge tracking means 24, surface element shape determination means 26, figure registration means 28, edge connection means 30, and edge shape adjustment means 32. Programs and other programs, and various data necessary for the processing of this system are stored. For example, the storage device 6 stores point cloud data of the analysis target space as the processing target data. For example, the inside of a tunnel on an automobile road is a target space for analysis.

  The input device 8 is a keyboard, a mouse, or the like, and is used for a user to operate the system.

  The output device 10 is a display, a printer, or the like, and is used to display the three-dimensional shape or cross-sectional shape of the tunnel obtained by this system to the user by screen display, printing, or the like.

  As will be described later, the basic process when extracting the tunnel shape in the present embodiment is a process of detecting the boundary line (edge) of the feature surface appearing on the cutting plane virtually set in the target space. For the cut plane, the feature surface in the band-like region along the edge is detected. Although the surface extracted with respect to one cutting plane is relatively local, it is possible to grasp the surface shape in a wider range by setting a plurality of cutting planes and analyzing each of them.

  The method for setting the cutting plane in the tunnel shape extraction system 2 is arbitrary, but is generally set to the cross section or vertical section of the tunnel. In the present embodiment, a case of a cross section will be described as an example. In the case of a cross section, a horizontal reference line is set along the crossing direction of the road, and the cutting plane is set to a vertical plane passing through the reference line. For example, a center line is set on a road, and the tunnel shape extraction system 2 performs a process of detecting a cross-sectional shape continuously or independently by generating a reference line 40 in the transverse direction by carving it at a certain distance (pitch). It is also possible to do this. FIG. 2 is a schematic perspective view showing a reference line and a cutting plane. The reference line 40 is set along the crossing direction of the road 42 and the tunnel 44. The reference line is set as the X axis, the vertical direction is set as the Y axis, and the longitudinal direction of the road is set as the Z axis. Therefore, the cutting plane 46 defines an XY plane.

  FIG. 3 is a schematic flowchart of tunnel shape extraction processing by the tunnel shape extraction system 2. Each means of the arithmetic processing unit 4 will be described with reference to FIG.

  The partial space setting means 20 sets a partial space that becomes a unit space for analysis by dividing a cross-section vicinity space that is a target of tunnel shape extraction processing in the set cutting plane out of the entire analysis target space. To do. The space near the cross section is a space having a dimension (depth) D in the Z-axis direction around the cutting plane. The depth D corresponds to the width of the band-like region on the feature surface detected corresponding to the above-described cutting plane. In this embodiment, for example, D is set to 30 cm. Incidentally, in FIG. 2, the horizontal cross section 48 in the cross section vicinity space is schematically represented by a rectangle.

  For example, the partial space setting means 20 divides the cross-section vicinity space in the horizontal direction and the vertical direction, and creates a rectangular parallelepiped partial space having sides along the X, Y, and Z axes. FIG. 4 is a schematic diagram for explaining the partial space, and shows the tunnel cross-sectional shape 50 and the arrangement of the partial space 52 in the cutting plane (XY plane). The space near the cross section is partitioned in a lattice shape (mesh shape), and the partial space 52 forms a two-dimensional array with the X direction and the Y direction as the array directions. In this embodiment, the dimension (width W) in the X-axis direction and the dimension (height H) in the Y-axis direction of the partial space are set to be the same as the depth D, for example, and the partial space is set to a cube. In the program constituting the partial space setting means 20, the width W, height H, and depth D are parameterized, and can be changed by the user using the input device 8, for example.

  The subspace setting means 20 sequentially creates subspaces that form a two-dimensional array. When the width W and the height H of the partial space are set as parameters, the number of partial spaces arranged in the X and Y directions is determined according to the size of the target space. The range of the index indicating the position of the partial space in each of the X and Y directions is determined based on the number of arrays. The partial space setting means 20 changes the index representing the two-dimensional array of the partial space in a predetermined order, and sets the partial space defined by the position (the coordinate range in each direction of X and Y) according to the set index. (S10).

  For example, as shown in FIG. 4, the lattice point at the substantially central portion of the tunnel cavity is set at the origin O, and the index in the X direction is set by increasing 1, 2, 3,... Decrease in the order of -1, -2, -3,... Further, the index in the Y direction is set by increasing in the order of 1, 2, 3,... Upward from the origin O and set by decreasing in the order of −1, −2, −3,. In this case, for example, the subspace setting means 20 (1) increments the Y-direction index sequentially from 1, and sequentially increments the X-direction index from 1 in the positive direction of the X axis for each value of the Y-direction index. (2) A scan that increments the Y-direction index sequentially from 1, and decrements the X-direction index in the negative direction of the X-axis sequentially from −1 at each value of the Y-direction index, (3) Scan in which the index in the Y direction is decremented sequentially from −1 and the index in the X direction is incremented sequentially from 1 in the positive direction of the X axis with each value of the index in the Y direction. (4) Y direction The index in the X direction is decremented in order from -1, and the index in the X direction is incremented in the negative direction of the X axis from -1 in each value of the Y direction index Perform the scanning to be incremented. In each scan, the subspace setting means 20 sequentially selects the subspaces, examines whether or not there is a point group in the subspace, and extracts a subspace including the point group as an edge search processing target.

  The partial space setting means 20 also performs scanning in which the scanning direction is changed by 90 ° in contrast to the above-described scanning, and extracts a partial space to be subjected to edge search processing even in this scanning. As a precaution, the scanning will be described. (1) The index in the X direction is incremented sequentially from 1, and the index in the Y direction is incremented from 1 in the positive direction of the Y axis at each value of the index in the X direction. (2) scanning in which the index in the X direction is incremented sequentially from 1 and the index in the Y direction is decremented sequentially from −1 in the negative direction of the Y axis at each value of the index in the X direction. 3) A scan in which the index in the X direction is decremented in order from −1, and the index in the Y direction is incremented sequentially from 1 in the positive direction of the Y axis at each value of the index in the X direction. Decrement the index in order from -1, and decrement the index in the Y direction in the negative direction of the Y-axis in order from -1 for each value of the index in the X direction. Is a scanning to door.

  The edge search means 22 analyzes the point group in each analysis unit space, and searches for an edge constituting the cross-sectional shape 50 of the tunnel 44 in the cutting plane (XY plane) 46 (S12, S14). In the search process, the edge search means 22 arranges various inspection lines on the XY plane of the analysis unit space, and examines whether the inspection line is a line (candidate line) that is an edge candidate. In the present embodiment, first, it is assumed that the surface of the feature in the partial space is approximately a plane, and a straight line (line segment) is extracted as an edge. Correspondingly, the shape of the inspection line is also a straight line or a line segment. The edge search means 22 sets a band-shaped area having a predetermined width along the inspection line, and an edge that is estimated to include the edge among the band-shaped areas that are variously set on the XY plane according to the inspection line. Search the register area. Specifically, the edge search means 22 obtains the number of data points (point group) located in the belt-like region and the height difference (coordinate difference in the Z-axis direction) of the data points belonging to the belt-like region (S12). The edge search means 22 extracts edges based on the conditions set for the result (S14). That is, when the point cloud gathers in a band-shaped area above a preset reference and the point cloud in the area has a depth difference equal to or greater than a preset threshold, the area is determined as an edge registration area, The inspection line that is the origin of the edge registration area is determined as an edge candidate. In step S12, whether or not the data point is in the band-like region is determined based on only the XY coordinate without considering the Z coordinate of the data point. For example, as a criterion for the number of point groups in the edge registration region among the edge extraction conditions, a condition that the number of point groups that gather near the inspection line is the largest can be set.

  The edge search means 22 may search only the edge where the point cloud is most concentrated in the analysis unit space, or may be a process assuming a multiple edge in which a plurality of edges exist in the analysis unit space.

  The edge tracking means 24 tracks the edge detected in the partial space outside the partial space (S16). When edge direction lines existing in the partial space (base point space) serving as the base point are detected, the edge tracking unit 24 creates a new analysis unit space adjacent to the base point space in the extension line direction before and after the direction line. Set the tracking space as. In this case, the edge search means 22 searches for an edge along the extension line in the tracking space. When an edge is detected in the tracking space, the edge tracking means 24 further sets the tracking space by extending the direction line of the edge in the same direction as the direction in which the tracking space is obtained, and the edge search means 22 searches for an edge in the tracking space. In this embodiment, the analysis unit space set as the tracking space has a rectangular parallelepiped shape as in the partial space. The dimension (W × H × D) can be the same as or different from the partial space described above. For example, since the direction and position of the edge being tracked are registered in advance, the dimension in the direction orthogonal to the tracking direction may be made narrower than the above partial space as a setting for making the dimension different from the above partial space. Is possible.

  The surface element shape determining means 26 obtains the shape of the part (surface element) detected as an edge on the cutting plane in the base point space or the tracking space in the feature surface in consideration of the coordinates in the depth direction of the point cloud. (S14, S16). Specifically, the surface element shape determination means 26 projects the data points having the (X, Y) coordinates in the band-like region corresponding to the edge onto a plane perpendicular to the cutting plane with the edge as a base line. The shape of the surface element is obtained based on the distribution range of the point cloud. The shape is a trapezoid in which the upper base and the lower base, which are two parallel sides, are set in the Z-axis direction and the distribution range of the point group on the projection plane is approximated.

  The graphic registration unit 28 functions as a partial cross-sectional shape generation unit that generates, for each partial space, a partial cross-sectional shape that becomes a part of the cross-sectional shape of the tunnel from an edge group obtained by tracking starting from the edge of the partial space. Have When the continuous tracking of the edges is interrupted, the graphic registration unit 28 connects the series of edges obtained by the edge tracking unit 24 to create a continuous line segment (partial cross-sectional shape), information indicating the continuous line segment, and The graphic data of each edge constituting the continuous line segment is registered in the storage device 6 (S18). A continuous line segment can be a single line (broken line) in which a series of edges are actually connected to each other at this stage, or a plurality of line segments that are arranged at this stage but are not connected. In other words, the process of interconnecting them into one line may be performed separately. In the latter case, for example, the interconnection can be performed by an edge shape adjustment process described later.

  The edge connecting means 30 joins the continuous line segments that are tracked and generated by the edge tracking means 24 for each partial space from which edges are extracted based on their positional relationship (S20). The density of point clouds obtained from the feature surface is not uniform. Specifically, in the mobile mapping system, a laser scanner mounted on an automobile irradiates a laser in an obliquely downward direction or an obliquely upward direction from the upper part of the vehicle body. The optical axis of the laser is scanned in the horizontal direction, and laser pulses are emitted at every minute angle within the scanning angle range. For this reason, the distance between the laser scanner and the laser reflection point on the surface of the feature changes depending on the laser emission direction, the distance between the laser scanner and the surface of the feature, and the like. As the distance from the laser scanner increases, the distance between the laser reflection points with respect to the difference in laser pulse emission angle increases, and the density of the point group decreases. In addition, there may be places where point clouds cannot be obtained due to the presence of water on the surface or shadows when viewed from the laser scanner. In places where the density of point clouds is low or where no point clouds exist, continuous tracking is performed by selecting many subspaces as the base point space by changing the scanning direction related to the selection of the base point space as described above. However, there is a possibility that a continuous edge cannot be extracted at a place to be connected. The edge connecting means 30 connects the continuous line segments extracted in the vicinity thereof at such a location. In addition, regarding the process of connecting the continuous line segments, the continuous line segments can be actually connected to form one line (folded line) at this stage. The process of actually connecting may be performed separately as the process of determining the combination of the end portions of the continuous line segments to be performed. In the latter case, for example, it is possible to perform processing for connecting the end portions with edges in edge shape adjustment processing described later.

  The edge shape adjusting means 32 positions the edge extracted in a straight line in the partial space as a rough edge of the feature surface, and represents a more detailed edge shape based on the line segment that is the rough edge. Is obtained (S22).

  Hereinafter, a processing example of the tunnel shape extraction system 2 will be described in more detail. 5, 6, and 8 to 10 separately show a flow of a series of processes executed by the arithmetic processing device 4. First, FIG. 5 is a schematic process flow diagram of the subspace setting means 20. The initial setting of the analysis unit space in the tunnel shape extraction system 2, that is, the processing of the portion related to the creation processing of the partial space is shown. When the tunnel shape extraction system 2 starts, for example, extraction processing of a tunnel shape such as a cross section of a tunnel (S30), the partial space setting unit 20 uses a tracking processing base space for dividing a partial space obtained by dividing the cross-section vicinity space into a mesh shape. Are set sequentially. For example, when the size of the partial space is set, the partial space in each of the X and Y directions depends on the size in the XY plane of the cross-section vicinity space having the depth D and set along the cutting plane. For example, the index range indicating the position of the partial space in each direction is determined. On the other hand, when the number of divisions (index width) in each direction of the cross-section near space is set, the size of the partial space is calculated and set according to the size of the target space.

  The subspace setting means 20 designates an index representing a two-dimensional mesh-like arrangement of the subspace in the order as described above, and is defined by a position (coordinate range in each direction of X and Y) according to the set index. A rectangular parallelepiped partial space is set as a base point space (S32). If the partial space of the mesh designated by the index is an unprocessed one (in the case of “No” in S34), the partial space setting means 20 sets the point group in the base point space to, for example, a hard disk or the like. The data is taken from the storage device 6 into a work area such as a RAM (Random Access Memory) (S36). When the number of point groups captured in the base point space is equal to or greater than the threshold value n (in the case of “Yes” in S38), and the point groups captured in the base point space are not flat (in the case of “No” in S40) , An edge search is performed using the partial space as a base point space (processing proceeds to node B in FIG. 6). For example, in the process S38, the threshold value n of the point group can be set to 5 or more. Further, the fact that the point group is not flat can be defined as the difference between the maximum value and the minimum value of the coordinate values in the Z direction of the point group being not less than a preset value, for example, not less than 2 cm.

  On the other hand, when the number of point clouds captured in the base point space is less than the threshold value n (in the case of “No” in S38), and when the point cloud captured in the base point space is determined to be flat (in S40). In the case of “Yes”, the search is not performed because there is no feature surface intersecting the cutting plane or the edge search cannot be performed with high accuracy. In these cases, the process is finished for the base point space, and the process proceeds to the process using the next partial space as the base point space (the process returns to node A).

  When the processing has been completed for all the partial spaces arranged in the mesh shape in the scanning order described above (“Yes” in S34), the arithmetic processing device 4 finishes the edge extraction processing (S42).

  FIG. 6 is a schematic processing flow diagram of the edge search means 22 and shows processing related to edge search in the base point space. The edge search means 22 searches for an edge in the base point space (S50).

  FIG. 7 is a schematic diagram for explaining the edge search processing S50 in the base point space. FIG. 7 is a diagram when the base point space is viewed from a direction (Z-axis direction) orthogonal to the cutting plane. An example of the arrangement of the rectangle corresponding to the base point space 60 and the point group in the base point space in the XY plane is as follows. It is shown. The edge search means 22 selects arbitrary two data points Pα and Pβ from the point group in the base point space, and sets a line segment L0 in the XY plane having these two points as both ends as an edge inspection line. The band-like areas EA having the width w (total width 2w) are set on both sides of the line segment L0. The edge search means 22 counts the number of data points whose coordinates on the XY plane are located in the band-like area EA. Whether or not the data point is in the area EA can be determined by whether or not the length of the perpendicular line from the data point to the line segment L0 on the XY plane is equal to or less than w, for example. In addition, as an edge condition, it is imposed that the three-dimensional shape represented by the point group in the belt-like area EA is not flat. Specifically, the edge search means 22 determines that the data point is not flat if the depth difference ΔZ of the data points included in the band-shaped area EA is equal to or greater than a preset threshold value γ. w and γ are parameters. For example, w can be set to about 3 cm and γ can be set to about 2 cm.

  The edge search means 22 performs the above-described determination by setting inspection lines for all combinations of two data points in the base point space, and the most data points gather in the area EA and ΔZ is equal to or greater than the threshold value γ. The line segment L0 is selected as an edge in the search process in the base point space. The edge becomes a registration candidate (edge candidate) in the linear list. If there is an edge (“Yes” in S52), the process proceeds to S54. On the other hand, when the edge is not found (in the case of “No” in S52), the process is finished for the base point space, the process proceeds to node A in FIG. 5, and the process proceeds to the process using the next partial space as the base point space.

  The edge search means 22 applies the least square method to the data points in the band-like area EA (edge registration area) corresponding to the line segment L0 selected as the edge in step S50, and follows the obtained regression line. Then, an optimization process for correcting the position of the line segment L0 is performed (S54).

  The surface element shape determining means 26 obtains the shape of the surface element of the feature existing at the edge obtained in step S54. Specifically, the data points in the belt-like area EA are projected onto a plane perpendicular to the cutting plane with the edge as a base line, and the shape of the surface element is obtained by approximating the distribution range of the projected point group with a trapezoid (S56). .

  The edge search means 22 checks whether a figure corresponding to the obtained edge (coincidence or intersecting direction line) exists in the figure already registered (S58). If the figure is not registered (in the case of “No” in S60), the edge (edge candidate) is assumed to be a new edge, and the process proceeds to node D in FIG. 8 and includes the position of the edge and the surface element shape. (Graphic data) is stored in the linear list, and the continuous tracking process is started. On the other hand, if the figure is registered (in the case of “Yes” in S60), the edge is already detected and is not a new edge, and the edge information obtained in the base point space is not recorded, The process proceeds to node A of No. 5, and the process proceeds to a process in which the next partial space is set as a base point space. The edge search means 22 defines an edge vicinity area which is a band-like area having a predetermined width with the edge as a center line and a length corresponding to the edge, and is stored in a linear list. The data also includes information representing the edge vicinity region.

  FIG. 8 is a schematic process flow diagram of the edge tracking means 24 and the figure registration means 28. The edge tracking means 24 generates a linear list for buffering edge information. The edge tracking means 24 creates two linear lists in response to the tracking process being performed in opposite directions from both ends of the edge. Here, tracking from one end of the edge is referred to as tracking in the A direction, and tracking from the other end of the edge is referred to as tracking in the B direction. The edge tracking means 24 creates a linear list for tracking in the A direction and a linear list for tracking in the B direction (S80). The graphic data of the edge obtained in the base point space can be stored, for example, in a linear list in the A direction (S82). Thereafter, the edge tracking means 24 performs a continuous tracking process S84 in the A direction and a continuous tracking process S86 in the B direction.

  The edge information obtained by the continuous tracking process in each of the A direction and the B direction is stored in a linear list in the A direction and the B direction. When the two-way tracking of the edge obtained in a certain base point space is completed, the figure registering unit 28 joins the edge endpoints stored in the linear list in the A direction and the linear list in the B direction according to the tracking order. A continuous line segment is created (S88). The created continuous line segment is registered as a figure in the storage device 6 (S90). The process proceeds to node A in FIG. 5 and proceeds to a process in which the next partial space is set as a base point space.

  FIG. 9 and FIG. 10 are schematic process flow diagrams of the continuous tracking process by the edge tracking means 24. When an edge is detected in the base point space or the tracking space set in advance, the edge tracking means 24 sets a tracking space adjacent to the analysis unit space in which the edge detection is performed in advance (S100). . The new tracking space is arranged by aligning the rectangular center line, which is a cross section in the cutting plane, with the direction of the edge detected in the preceding analysis unit space. In addition, when an edge is detected in the tracking space set in advance, the direction line of the edge is extended in the same direction as the direction in which the tracking space is obtained, and further tracking is performed.

  When the edge tracking means 24 sets the tracking space, the edge search means 22 takes the point cloud in the tracking space from the storage device 6 into a work area such as a RAM (S102), and uses it as an extension of the edge of the preceding analysis unit space. It is checked whether there is an edge in the tracking space set along (S104). If point clouds above the reference are gathered in the tracking space set along the extension line and the height difference is equal to or larger than the threshold γ, it is determined that an edge exists in the tracking space (“Yes” in S106). If). In this case, the surface element shape determining means 26 performs a process similar to the process S56 performed on the edge in the base point space, so that the trapezoid serving as the surface element of the feature existing in the edge extracted in the tracking space is obtained. Processing to determine is performed (S108).

  The edge tracking means 24 checks whether or not the extracted edge (edge candidate) intersects with an edge stored on the linear list (S110). Here, the edge registered in the linear list originates in the same base point space as the edge newly extracted from the tracking space. When a figure with a closed edge is formed as in a cross section of a tunnel, the closed figure may be extracted by continuous tracking processing from a certain base point space. Therefore, in process S110, it is determined whether or not there is an intersection (including a match) between the newly extracted edge and the edge on the linear list. If an intersection is detected (“Yes” in S112), In order to avoid the self-loop or the infinite loop, the newly extracted edge is not added to the linear list, and the continuous tracking process is terminated (S114).

  When no intersection is detected (in the case of “No” in S112), the edge tracking means 24 further checks whether the newly extracted edge has already been registered in the graphic (S116). When the graphic is registered (in the case of “Yes” in S118), the newly extracted edge is not added to the linear list, and the continuous tracking process is terminated (S120).

  On the other hand, if the figure is not registered (in the case of “No” in S118), the edge tracking unit 24 adds the edge information about the current tracking space to the linear list (S122), and returns to the process S100. Set a new tracking space that follows the current tracking space. In this way, the edge tracking means 24 tracks the edges by sequentially setting the tracking space.

  If the edge tracking means 24 cannot detect an edge in the tracking space set along the extension line of the detected edge (“No” in S106), the process proceeds to node F in FIG. 10 to extend the detected edge. Search processing in a direction deviating from the line is performed. In the search process, the tracking space is set by changing the central axis in the cutting plane from the position of the extension line to the clockwise direction and the counterclockwise direction. For example, the change range (± θ) of the direction of the central axis from the direction of the extension line and the number of divisions k are set, and a plurality of tracking spaces having different angles in the cutting plane by θ / k are set. An edge is searched for in the space, and an edge having the most point cloud is extracted from the edges obtained in a plurality of tracking spaces set by changing the angle (S130).

  When the edge is extracted by this (when “Yes” in S132), the surface element shape determination means 26 determines the surface element of the feature existing in the edge extracted in the tracking space in the same manner as the processing S108. The process which determines the trapezoid which becomes is performed (S136). Further, the edge tracking means 24 performs the same processing as the processing S110 to S122. That is, the presence / absence of intersection between the extracted edge and the edge on the linear list is checked (S138), and it is checked whether the extracted edge is already registered in the figure (S144). When an intersection is detected (“Yes” in S140) and a figure is registered (“Yes” in S146), the newly extracted edge is not added to the linear list. Then, the continuous tracking process is ended (S142 and S148), and the process returns to the processes S84 and S86 of FIG. On the other hand, if no intersection is detected (in the case of “No” in S140) and no graphic is registered (in the case of “No” in S146), the edge tracking unit 24 uses the extracted edge information as a linear list. (S150), the process returns to the node E in FIG. 9 to set a new tracking space following the current tracking space. If no edge is extracted (in the case of “No” in S132), the process returns to S84 and S86 of FIG. 8 in which the continuous tracking process is called as a subroutine (S134).

  FIG. 11 is a schematic diagram of a cutting plane showing an example of an edge 102 extracted in a certain partial space 100 and a continuous tracking process using the edge 102 as a base point. In the tracking process, the edge 106 is searched by sequentially setting the tracking space 104 in the clockwise direction and the counterclockwise direction from the partial space 100. A series of edges 102 and 106 on the right side of the tunnel cross-sectional shape 50 of FIG. 11 constitute one continuous line segment.

  The arithmetic processing device 4 finishes the edge search in the base point space and the continuous tracking process in the tracking space, and when the extraction of the continuous line segment representing the partial cross-sectional shape of the tunnel in the cutting plane is completed (S42 in FIG. 5), the edge Processing for connecting the continuous line segments is performed by the connecting means 30 (processing S20 in FIG. 3).

  The edge connecting means 30 (partial cross-sectional shape connecting means) connects between the ends of two continuous line segments. The edge connecting means 30 selects two continuous line segments extracted corresponding to the cutting plane from the storage device 6, and determines whether or not the two continuous line segments can be connected by two methods. This determination is made for all continuous line segment pairs in the cutting plane.

  FIG. 12 is a schematic diagram illustrating a first determination method for determining whether or not a continuous line segment pair can be connected. In the first technique, the extension line L1 of the edge E1 located at the end of one continuous line segment of the selected pair of continuous line segments is the edge vicinity region R2 of the edge E2 located at the end of the other continuous line segment. Whether or not connection is possible is determined based on whether or not they intersect, and when it is determined that connection is possible, the ends PB1 and PA2 of the edges E1 and E2 are connected and two continuous line segments are integrated into one. .

  Here, the determination as to whether or not the connection is possible can be determined not only in one direction as to whether the extended line L1 from the end portion PB1 of the edge E1 intersects the edge vicinity region R2 of the edge E2, but also in both determinations. . That is, whether or not connection is possible can be determined in consideration of whether an extension line (not shown) of the edge E2 from the end portion PA2 intersects the edge vicinity region R1 of the edge E1. For example, it can be determined that the connection is possible if the intersection is determined in any one of the two directions, and can be determined that the connection is possible when the intersection is determined in both directions.

  Further, since each continuous line segment has two ends, there are four combinations of end portions in the two continuous line segments. Specifically, when the opposite end of the continuous line segment having the edge E1 at one end is defined as an edge E1 'and the opposite end of the continuous line segment having the edge E2 at one end is defined as an edge E2', the edges E1, E1 ' Combinations of these and the edges E2 and E2 ′ can be considered. Of these, the edge connecting means 30 basically determines whether or not connection is possible only for the combination having the smallest distance between the end portions, but may determine whether or not connection is possible for a plurality of combinations, for example, all combinations.

The width (dimension in the direction perpendicular to the edge) w _N (not shown) of the edge vicinity region can be a constant value. On the other hand, for example, it may be changing the width w _N in accordance with the density of the point cloud of the corresponding analyzing unit space and edge registration region an edge is extracted. For example, it is considered to be high accuracy of the position of the edges extracted density of the points from the high analysis unit space and edge registration area, can the density is set smaller the higher the width w _N.

  FIG. 13 is a schematic diagram illustrating a second determination method for determining whether or not a continuous line pair can be connected. In the second method, when the distance DN between the ends of the selected continuous line segment pair is equal to or less than a predetermined threshold ε, the ends are connected. Specifically, first, a combination having the smallest distance between the end portions is selected from the four combinations of the end portions of the two continuous line segments. In FIG. 13, among the four combinations, the distance between the end portion PB1 of the edge E1 and the end portion PA2 of E2 is the minimum, and if the distance is equal to or less than ε, the continuous line segment pair is connected at the end portion. If the distance is larger than ε, the continuous line segment pair is not connected by the second method.

  Note that the first method is effective when the point group is relatively sparse and the distance DN between the ends of the pair of continuous line segments is long. On the other hand, the second method is effective when the point group is relatively dense and the distance DN between the ends of the pair of continuous line segments is short. Further, by using the first method and the second method together, it is also possible to connect, for example, continuous line segment pairs that could not be connected by the first method by the second method.

  By connecting the continuous line segments by the edge connecting means 30, for example, the cross-sectional shape of the tunnel can be easily obtained as a closed figure. With the processing so far, the cross-sectional shape of the tunnel can be expressed by a broken line connecting edges extracted for each partial space and tracking space. The tunnel shape extraction system 2 can also output this as the cross-sectional shape of the tunnel, but the tunnel shape extraction system 2 can further adjust and output the cross-sectional shape of the tunnel by the edge shape adjusting means 32. The edge shape adjusting means 32 positions the edge extracted in the partial space or tracking space by the edge searching means 22 as a rough edge, and represents a more detailed edge shape based on the line segment that is the rough edge. The fine adjustment processing of the edge for obtaining is performed (processing S22 in FIG. 3).

  FIG. 14 is a schematic flowchart of edge fine adjustment processing S22 by the edge shape adjusting means 32. FIG. 15 is a schematic diagram for explaining edge fine adjustment processing by the edge shape adjusting means 32. FIG. 15 is a view of the partial space 70 as viewed from the direction orthogonal to the cutting plane (Z-axis direction), corresponding to the rectangle corresponding to the partial space 70 and the approximate edge L0 detected in the partial space 70. The example of the detailed edge Lf produced | generated is shown. In FIG. 15, the space 70 is a partial space, but the edge fine adjustment processing is performed in the same manner even if this is the tracking space.

When the edge search means 22 obtains a line segment L0 indicating an approximate edge, the edge shape adjusting means 32 delimits the line segment by a certain length from one end (S160), and details are generated in each section generated thereby. A point (representative point) through which the edge passes is searched, and instead of the line segment L0 indicating the approximate edge connecting PA and PB in the XY plane, a fold line Lf connecting the representative points obtained in each section is used as the edge. Here, the number of sections is N. When the representative point P _K is obtained from the selected section K, the arithmetic processing unit 4 obtains the coordinates (X _K , Y _K ) of the representative point P _{K in} the cutting plane as representative point information.

  The edge shape adjusting means 32 increments the section number K by 1, selects a section (S162, S164), and searches for a representative point in each section (S168). The search is performed in a compartment 74 obtained by dividing the edge peripheral space 72 for each section. The edge peripheral space 72 has a shape projected onto the cutting plane and is a region near the edge, and can be set to a space having a depth corresponding to the surface element in the Z direction. Specifically, the shape of the edge peripheral space 72 on the cutting plane is a rectangle in which both ends PA and PB of the approximate edge are located at the midpoints of the two opposing sides. The dimension in the direction perpendicular to the approximate edge of the rectangle can be set in consideration of the curvature assumed for the cross-sectional shape of the tunnel to be detected and the size of the unevenness. Its size needs to be large enough to accurately capture the edge at the cutting plane of the tunnel. On the other hand, from the viewpoint of making it difficult for point clouds that do not originate from the tunnel edge to be erroneously detected as representative points. Is preferably smaller than, for example, the width W or height H of the partial space. The compartment 74 is generated by dividing the edge peripheral space 72 by a plane that passes through the boundary point of the section where the outline edge is divided and is orthogonal to the outline edge. The edge shape adjusting means 32 takes in the point cloud data in the compartment 74 corresponding to the section for each section, and obtains the position where the point cloud is most concentrated in the cutting plane of the compartment 74. The crowded position is determined by the number of point groups included in a circle centered on an arbitrary point of the point group. For example, the radius of the circle can be 2 cm. If the number of points crowded at the position is greater than or equal to a predetermined number (“Yes” in S170), the position is added as a representative point to the edge representative point list (S172), and the next section The process proceeds (S164). On the other hand, if the number of points crowded at the position is less than the predetermined number (in the case of “No” in S170), the representative point of the edge does not exist in the section and is not added to the edge representative point list. The process proceeds to the next section (S164). When the processing is completed for all the sections (in the case of “Yes” in S166), the edge shape fine adjustment processing is ended.

In the example shown in FIG. 15, the line segment L0 indicating the general edge is divided into, for example, four sections, and the edge peripheral space 72 represented by a dotted rectangle with L0 as the center line is also divided into four sections. Comparted into compartments 74a-74d. For example, representative points P _k , P _{k + 1} , and P _{k + 2} are extracted in the compartments 74a, 74b, and 74d, and a broken line P _k P _{k + 1} P _{k + 2} that connects them in the order of the compartments becomes the detailed edge Lf. Incidentally, in the example shown in FIG. 15, the edge peripheral space 72 is divided at regular intervals from the PA side, and as a result, it is shown that the last section can be smaller than the other sections.

  The obtained detailed edge information is registered in the storage device 6 in association with the schematic edge graphic data, for example. The shape of the surface element corresponding to the detailed edge Lf is determined so that, for example, the shape obtained by projecting it onto the plane perpendicular to the cutting plane with the approximate edge L0 as the baseline matches the shape of the surface element with respect to the approximate edge. Can do.

  FIGS. 16 to 18 are schematic views of cutting planes for explaining the process of generating the detailed edge from the general edge, and show the step portion 108 (FIG. 11) beside the road of the tunnel cross-sectional shape 50. FIGS. FIG. 16 shows a schematic edge 122 extracted from each tracking space 120 set in the step portion 108. In FIG. 17, an edge peripheral space 130 having a rectangular cross-sectional shape with the approximate edge 122 as the center line is divided at a predetermined interval from one end of the approximate edge 122 to set a compartment 132. In each compartment, representative points are searched, and the obtained representative points 134 are sequentially connected to generate a detailed edge 140 of a broken line. FIG. 18 shows the detailed edge 140. The detailed edge 140 shown in FIG. 18 better represents the shape of the stepped portion 108 than the schematic edge 122 shown in FIG. In other words, the tunnel shape extraction system 2 can extract the cross-sectional shape of the tunnel with high accuracy by obtaining the detailed edge, and thereby, for example, can detect the shape change of the surface inside the tunnel with high sensitivity.

  In the above embodiment, a program that causes a computer to operate as each means of the tunnel shape extraction system 2 is stored in the storage device 6 and the computer reads and executes the program. However, in other configurations, the program is a network or the like. In this case, the tunnel shape extraction system 2 includes a communication device, and the communication device acquires a program from a network or the like and provides it to the arithmetic processing device 4 or a storage device. 6 is stored. Further, the program can be provided by being stored in a recording medium such as a CD-ROM (Compact Disc Read Only Memory).

  2 tunnel shape extraction system, 4 arithmetic processing device, 6 storage device, 8 input device, 10 output device, 20 partial space setting means, 22 edge search means, 24 edge tracking means, 26 surface element shape determination means, 28 graphic registration means , 30 Edge connecting means, 32 Edge shape adjusting means, 50 Tunnel cross-sectional shape, 52, 70, 100 Partial space, 72, 130 Edge peripheral space, 74, 132 compartments, 122 Outline edges, 134 representative points, 140 Detailed edges.

Claims (11)

A data analysis device for detecting a cross-sectional shape of a tunnel existing in a target space based on three-dimensional coordinate data of a point cloud extracted from a feature surface,
A vertical cutting plane is virtually set in the target space, and a cross-section near space having a predetermined width in the depth direction along the cutting plane and orthogonal to the cutting plane in the target space is formed in a lattice shape A partial space setting means for setting a plurality of partial spaces divided and two-dimensionally arranged along the cutting plane;
Each subspace is set as an analysis unit space, and for each analysis unit space, an edge search means for searching for an edge that is a position of a surface in the tunnel in the cutting plane,
The edge search means is a band-shaped region having a predetermined width, and the point group projected onto the cutting plane is gathered over a predetermined reference in the region, and the point group in the region is An edge registration area having a distribution width equal to or greater than a predetermined threshold in the depth direction is searched in the cutting plane, and a direction line along the distribution in the cutting plane of a point group belonging to the edge registration area is the edge. Asking for,
A data analysis device characterized by The data analysis apparatus according to claim 1,
Edge tracking means for tracking the edge detected in the partial space outside the partial space;
The edge tracking means, when the edge is detected from the analysis unit space, a tracking space as a new analysis unit space adjacent to the analysis unit space in the direction of the extension line of the direction line corresponding to the edge. Set
The edge searching means searches the tracking space for the edge along the extension line;
A data analysis device characterized by In the data analysis device according to claim 2,
The tracking space is a rectangular parallelepiped space in which the shape in the cutting plane is a rectangle, and the midpoint of one side of the rectangle is connected to the end of the detected edge,
The edge tracking means is
Setting the tracking space in which the center line of the rectangle passing through the midpoint coincides with the extended line of the edge as the analysis unit space;
When the edge is not detected in the tracking space, the tracking space in which the most point cloud is gathered inside by changing the direction of the center line in the cutting plane as the tracking space as the analysis unit space Exploring,
A data analysis device characterized by In the data analysis device according to claim 2 or 3,
Each partial space has a partial cross-sectional shape generating means for generating a partial cross-sectional shape that becomes a part of the cross-sectional shape of the tunnel from an edge group obtained by tracking starting from the edge of the partial space. Data analysis device. The data analysis apparatus according to claim 4, wherein
The edge tracking means uses the direction line detected in the analysis unit space as an edge candidate, and when the edge candidate intersects the partial cross-sectional shape, the edge candidate is not a new edge. Data analysis device. In the data analysis device according to any one of claims 2 to 5,
The edge tracking means uses the direction line detected in the analysis unit space as an edge candidate, and the edge candidate tracked from the edge of the subspace is tracked from the edge of the subspace or the edge A data analysis apparatus characterized by not making an edge candidate a new edge when intersecting with another edge. In the data analysis device according to any one of claims 4 to 6,
A partial cross-sectional shape connecting means for connecting two end portions of the partial cross-sectional shape;
The edge search means defines an edge vicinity region corresponding to the edge, having a predetermined width with the edge as a center line, and a band-like region having a length corresponding to the edge,
The partial cross-sectional shape connection means, when the extension line of the edge of one end of the partial cross-sectional shape intersects the edge vicinity region of the edge of the other end of the partial cross-sectional shape, Connecting between the
A data analysis device characterized by In the data analysis device according to any one of claims 4 to 7,
A partial cross-sectional shape connecting means for connecting two end portions of the partial cross-sectional shape;
The edge search means, when the distance between the end portions is equal to or less than a predetermined threshold, connecting the both end portions;
A data analysis device characterized by In the data analysis device according to any one of claims 2 to 8,
The data analysis apparatus, wherein the edge tracking means starts tracking from both ends of the edge detected in the partial space. A data analysis method for detecting a cross-sectional shape of a tunnel existing in a target space based on three-dimensional coordinate data of a point cloud extracted from a feature surface,
A vertical cutting plane is virtually set in the target space, and a cross-section near space having a predetermined width in the depth direction along the cutting plane and orthogonal to the cutting plane in the target space is formed in a lattice shape A subspace setting step for setting a plurality of subspaces divided and two-dimensionally arranged along the cutting plane;
An edge search step for setting each subspace as an analysis unit space, and searching for an edge that is a position of a surface in the tunnel in the cutting plane for each analysis unit space;
The edge search step is a band-shaped region having a predetermined width, and the point group projected onto the cutting plane is gathered over a predetermined reference in the region, and the point group in the region is An edge registration area having a distribution width equal to or greater than a predetermined threshold in the depth direction is searched in the cutting plane, and a direction line along the distribution in the cutting plane of a point group belonging to the edge registration area is the edge. Asking for,
A data analysis method characterized by A program for causing a computer to perform data analysis for detecting a cross-sectional shape of a tunnel existing in a target space based on three-dimensional coordinate data of a point cloud extracted from a feature surface,
A vertical cutting plane is virtually set in the target space, and a cross-section near space having a predetermined width in the depth direction along the cutting plane and orthogonal to the cutting plane in the target space is formed in a lattice shape A partial space setting means for setting a plurality of partial spaces divided and two-dimensionally arranged along the cutting plane; and
Each subspace is set as an analysis unit space, and for each analysis unit space, function as edge search means for searching for an edge that is a position of a surface in the tunnel in the cutting plane,
The edge search means is a band-shaped region having a predetermined width, and the point group projected onto the cutting plane is gathered over a predetermined reference in the region, and the point group in the region is An edge registration area having a distribution width equal to or greater than a predetermined threshold in the depth direction is searched in the cutting plane, and a direction line along the distribution in the cutting plane of a point group belonging to the edge registration area is the edge. Asking for,
A program characterized by

Images