Inspection system and inspection method

Abstract

To solve conventional problems, that is, to provide an inspection system and an inspection method that enable an excessive extraction point eliminated through automatic filtering processing to be found more easily than in prior art.SOLUTION: An inspection system of the invention of the present application is a system for inspecting a candidate ground point obtained by performing automatic filtering processing on point group data composed of a plurality of measurement points, including mesh setting means, lowest point extraction means, shade figure generation means, and display means. Among them, the shade figure generation means is means for generating a monochromatic shade figure based on a lowest point in each small area. Then, by visually inspecting the monochromatic shade figure and the candidate ground point displayed on the display means in an overlapping manner, an inspector is able to easily extract an excessive extraction point.SELECTED DRAWING: Figure 6

Description

[Technical Field]

[0001] The present invention relates to a technology for generating ground surfaces from point cloud data obtained by a laser scanner, and more specifically to an inspection system and inspection method that can extract measurement points that have been excessively removed by automatic filtering. [Background technology]

[0002] When measuring "ground" or "land features" (hereinafter collectively referred to as "ground, etc.") over a wide area, aerial photogrammetry has been the mainstream method up until now, but recently, various measurement methods have emerged, such as airborne laser measurement, measurement using satellite photos, and measurement using synthetic aperture radar, making it possible to select the most appropriate method depending on the situation. Note that "land features" here refers to "objects" excluding the ground.

[0003] Of these, airborne laser measurement is a method in which an aircraft flies over the target area to be measured and receives the reflected waves of laser pulses emitted by a laser scanner onto the ground and other surfaces within the target area. Aircraft are usually equipped with a positioning device such as a GNSS (Global Navigation Satellite System) and an inertial measurement device such as an IMU (Inertial Measurement Unit), and these GNSS and IMU can record the irradiation position (x, y, z) and irradiation attitude (ω, φ, κ) when the laser pulse is emitted.

[0004] When a laser pulse is emitted from the aircraft's laser scanner, the time of emission is recorded, and when the laser pulse is reflected from the ground or other surfaces, it is received by a sensor mounted on the aircraft, and the time of reception is also recorded. Therefore, the distance to the measurement point (the point where the laser pulse is reflected) can be obtained from the time difference between the emission time and the reception time. Since the irradiation position (x, y, z) and irradiation attitude (ω, φ, κ) at the time of laser pulse emission are also recorded, the 3D coordinate data of the laser pulse irradiation point (i.e., the measurement point) can be obtained. Furthermore, when the sensor mounted on the aircraft receives the laser pulse, it records the intensity of the reflected wave at that time. This reflection intensity is, so to speak, the magnitude of the energy of the received reflected wave (the amplitude of the laser pulse) and is measured directly as a voltage, and the magnitude of the energy can be obtained by converting this voltage.

[0005] As explained so far, airborne laser measurement is a method of obtaining measurement points by irradiating laser pulses onto the ground or other objects from an aircraft in flight. These laser pulses are emitted 100,000 to 2,000,000 times per second, resulting in a huge number of measurement points being obtained in a single measurement (flight). Furthermore, when measuring forests and other areas, it is obviously impossible to irradiate laser pulses only onto the ground through the trees (i.e., to aim the laser pulses at the ground). Therefore, in addition to measurement points that are reflected by the ground, measurement points that are reflected by tree leaves and trunks are also obtained. For convenience, measurement points that are reflected by the ground will be referred to as "ground points," and measurement points that are reflected by objects other than the ground (i.e., features) will be referred to as "non-ground points."

[0006] In many cases, airborne laser measurements are conducted to understand the ground conditions within a target area, and measurement points that reflect off tree leaves or trunks (non-ground points) are considered unnecessary data. For this reason, so-called filtering is performed to remove non-ground points. This filtering is typically performed in two stages: automatic filtering and manual filtering. Of these, automatic filtering, which is performed first, literally removes non-ground data automatically using a computer and specified software. However, automatic filtering is known to not provide a complete ground image at this stage, as it may leave ground points that should have been removed as non-ground points (hereinafter referred to as "unsampled points"), or remove ground points that should have been removed as non-ground points (hereinafter referred to as "over-sampled points").

[0007] Therefore, an inspector visually inspects the ground points extracted by automatic filtering to find unextracted or over-extracted points, and then performs manual filtering to remove the unextracted points and return the over-extracted points to ground points. However, it is extremely difficult to extract unextracted or over-extracted points simply by visually inspecting the ground points after automatic filtering (hereinafter referred to as "candidate ground points"). Therefore, various techniques have been proposed to enable easy and appropriate manual filtering. For example, Patent Document 1 discloses an invention in which manual filtering is performed after superimposing a color shaded relief map on the candidate ground points to find unextracted points, and after superimposing a monochrome shaded relief map on the candidate ground points to find over-extracted points. [Prior art documents] [Patent documents]

[0008] [Patent Document 1] Japanese Patent Application Laid-Open No. 2008-310785 Summary of the Invention [Problem to be solved by the invention]

[0009] According to the invention of Patent Document 1, unsampled points are clearly displayed on color shaded relief maps, and oversampled points are clearly displayed on monochrome shaded relief maps, allowing inspectors to easily identify unsampled and oversampled points. However, when automatic filtering is performed in densely wooded areas, for example, signals that actually reach the ground are often treated as oversampled points, i.e., reflected by trees. Even when manually filtering is performed by superimposing candidate ground points on a conventional monochrome shaded relief map, inspectors may not be able to clearly identify whether or not a point is oversampled. For example, Figure 8 shows a screenshot of candidate ground points obtained by automatic filtering of a densely wooded area, superimposed on a conventional monochrome shaded relief map. The outlined area shows oversampled points, but visually identifying these as oversampled points is extremely difficult. In this case, inspectors must return to the original point cloud data (so-called raw data) to check the data, which increases the time and effort required for inspection.

[0010] The object of the present invention is to solve the conventional problems, that is, to provide an inspection system and an inspection method that can more easily find over-extracted points that have been removed by automatic filtering processing compared to the prior art. [Means for solving the problem]

[0011] The present invention focuses on the fact that, in order to extract over-extracted points, the lowest point of a small area set within the range of point cloud data (measurement range) is extracted, and a monochrome shaded relief map is generated based on this lowest point, and is an invention based on an idea that has not been seen before.

[0012] The inspection system of the present invention inspects candidate ground points obtained by performing automatic filtering on point cloud data consisting of multiple measurement points, and includes a mesh setting means, a lowest point extraction means, a shaded relief map generation means, and a display means. The mesh setting means sets multiple small areas within the range of the point cloud data, and the lowest point extraction means extracts the lowest point with the smallest elevation value among the measurement points included in the small area. The shaded relief map generation means generates a monochrome shaded relief map based on the lowest point associated with each small area, and the display means superimposes the monochrome shaded relief map and the candidate ground points. By visually checking the monochrome shaded relief map and the candidate ground points superimposed on the display means, an inspector can extract measurement points that were excessively removed by the automatic filtering process.

[0013] The inspection system of the present invention may further include an area selection means for selecting a portion of the range of the point cloud data as a selected area, in which case the shaded relief map generation means generates a monochrome shaded relief map based on a small area related to the selected area selected by an operator.

[0014] The inspection system of the present invention may further include a size setting unit for setting the size of the small regions. When multiple selection regions are selected by an operator, an individual small region size can be set for each of the selected regions.

[0015] The inspection method of the present invention is a method for inspecting candidate ground points using the inspection system of the present invention, and includes an automatic filtering step, a mesh setting step, a lowest point extraction step, a shaded relief map generation step, a superimposition display step, and an excessive point extraction step. In the automatic filtering step, candidate ground points are extracted by performing an automatic filtering process on point cloud data. In the mesh setting step, a plurality of small areas are set in a measurement range by a mesh setting means. In the lowest point extraction step, the lowest points of the measurement points included in the small areas are extracted by a lowest point extraction means. In the shaded relief map generation step, a monochrome shaded relief map is generated by the shaded relief map generation means based on the lowest points of each small area. In the superimposition display step, the monochrome shaded relief map and the candidate ground points are superimposed on the display means. In the excessive point extraction step, measurement points that were excessively removed by the automatic filtering process are extracted by visually checking the monochrome shaded relief map and the candidate ground points superimposed on the display means. [Effects of the Invention]

[0016] The inspection system and inspection method of the present invention have the following effects. (1) Inspectors can extract over-extracted points without having to go back and check the original point cloud data. As a result, manual filtering can be performed more easily and at lower cost than with conventional techniques. (2) Since fewer measurement points are used than in the past, monochrome shaded relief maps can be created more quickly. (3) Since monochrome shading maps are created using fewer measurement points than conventional methods, they can be manipulated more easily. [Brief explanation of the drawings]

[0017] [Figure 1] (a) is a side view showing a schematic representation of all measurement points obtained by laser measurement, and (b) is a side view showing a schematic representation of the measurement points after automatic filtering. [Figure 2] 1 is a block diagram showing the main configuration of an inspection system according to the present invention; [Figure 3] FIG. 4 is a plan view schematically showing a plurality of small regions set in a measurement range. [Figure 4] (a) is a plan view showing all measurement points included in a small area, and (b) is a plan view showing the lowest point extracted from the small area. [Figure 5] A screenshot showing candidate ground points obtained as a result of automatic filtering of densely treed areas superimposed on a monochrome low-land shaded relief map. [Figure 6] FIG. 2 is a flowchart showing an example of a main processing flow of the inspection system. [Figure 7] 1 is a flow chart showing the flow of main steps in an inspection method according to the present invention. [Figure 8] A plan view in which candidate ground points obtained as a result of automatic filtering of densely treed areas are superimposed on a conventional monochrome shaded relief map. DETAILED DESCRIPTION OF THE INVENTION

[0018] An example of an embodiment of the inspection system and inspection method of the present invention will be described with reference to the drawings.

[0019] 1.Overview As mentioned above, laser measurements are often performed to understand the ground (surface shape) within a target area (hereinafter simply referred to as the "measurement area"). For this reason, filtering is performed to remove "non-ground points" that measure features other than the ground from all measurement points obtained by measuring the target area (hereinafter referred to as the "original data"), and extract only "ground points" that measure the ground. This filtering is typically performed in two stages: automatic filtering and manual filtering, and automatic filtering mechanically (automatically) removes non-ground data to generate "candidate ground points."

[0020] However, it is known that in the process of generating "candidate ground points" through automatic filtering, "over-extracted points" occur, where points that should have been ground points are removed as non-ground points. Figure 1 is a side view showing a schematic diagram of measurement points obtained by laser measurement, where (a) shows all measurement points (i.e., the original data), and (b) shows the measurement points after automatic filtering (i.e., the candidate ground points). As shown in this figure, measurement points that pass through trees and reach the ground are often deemed non-ground points by automatic filtering, meaning they are prone to being over-extracted points.

[0021] Therefore, manual filtering is performed to visually identify over-sampled points while comparing the original data with the candidate ground points, and then to return the over-sampled points as ground points. In this case, it is effective to generate a monochrome shaded relief map based on the original data and compare the monochrome shaded relief map (i.e., the original data) with the candidate ground points. Monochrome shaded relief maps are a well-known technique that shades the terrain by shining light onto a 3D ground model from a specific direction and displays the shading in shaded areas. Conventional monochrome shaded relief maps are generated using all measurement points (original data). In contrast, one of the technical features of the present invention is the use of a monochrome shaded relief map generated using measurement points at low elevations.

[0022] 2.Inspection system Next, the inspection system of the present invention will be described in detail with reference to the drawings. The inspection method of the present invention is a method of inspecting ground points after automatic filtering using the inspection system of the present invention, so the inspection system of the present invention will be described first, and then the inspection method of the present invention will be described.

[0023] 2 is a block diagram showing the main components of an inspection system 100 of the present invention. As shown in this figure, the inspection system 100 is configured to include mesh setting means 101, lowest point extraction means 102, shaded relief generation means 103, and display means 104, and can also be configured to include area selection means 105, size setting means 106, automatic filtering means 107, model generation means 108, point cloud data storage means 109, etc.

[0024] The mesh setting means 101, lowest point extraction means 102, shaded relief generation means 103, area selection means 105, and size setting means 106 that make up the inspection system 100 can be manufactured as dedicated units, or a general-purpose computer device can be used. That is, the computer device executes calculations using a predetermined program to perform processing specific to each means. This computer device is equipped with a processor such as a CPU, memories such as ROM and RAM, and some also include input means such as a mouse and keyboard, and a display, and can be configured, for example, as a personal computer (PC) or server. When a computer device including a display is used, the display can also be used as the display means 104.

[0025] The point cloud data storage means 109 can be a storage device of a general-purpose computer (for example, a personal computer) or can be built in a database server. When built in a database server, it can be placed on a local network (LAN: Local Area Network) or can be a cloud server that stores data via the Internet.

[0026] Below, each of the main elements that make up the inspection system 100 of the present invention will be described in detail.

[0027] (Mesh setting method) The mesh setting means 101 is a means for setting "small regions MS" in the measurement range. Here, the small regions MS are divided regions (so-called meshes) formed by dividing the measurement range (however, a planar range) by, for example, orthogonal grids, as shown in Fig. 3, and are made up of information indicating two-dimensional positions (planar positions), but do not have height information such as altitude.

[0028] The mesh setting means 101 can be configured to set small regions MS for the entire measurement range, or it can be configured to set small regions MS only for specified regions, such as mountainous areas with many trees. In this case, it is preferable that the inspection system 100 is equipped with region selection means 105. The region selection means 105 can use a pointing device (such as a mouse, touch panel, pen tablet, touchpad, trackpad, or trackball) or a keyboard; that is, the operator selects a desired region (hereinafter referred to as the "selected region") using a pointing device or the like. The mesh setting means 101 then sets a small region MS for the selected region.

[0029] The dimensions of the small region MS (hereinafter referred to as "mesh size") can be set according to the conditions of the measurement range (such as the terrain undulations and density of trees) and the requirements for manual filtering (such as the required accuracy and the given work time), and can be, for example, 20 cm x 20 cm. This mesh size can be fixed (unchangeable) at a predetermined value, or can be set (changeable) by the operator as needed. In this case, the inspection system 100 should preferably be equipped with a size setting means 106. The operator selects the desired mesh size using a keyboard, pointing device, or the like, and the mesh setting means 101 sets the small region MS using that mesh size. Note that if two or more selected regions within the measurement range are selected by the region selection means 105, the size setting means 106 should preferably be able to set a unique mesh size for each selected region.

[0030] (lowest point extraction means) The lowest point extraction means 102 is a means for reading out the original data stored in the point cloud data storage means 109 and extracting the measurement point (hereinafter referred to as the "lowest point") that has the smallest elevation value among the measurement points contained in the small area MS. For example, in Figure 4, 16 small areas MS are shown, and in (a) all the measurement points (black dots and white dots) are shown in each small area MS, while in (b) one lowest point (white dot) extracted in each small area MS is shown.

[0031] (Shaded relief generation means) The shaded relief map generating means 103 generates a monochrome shaded relief map based on the lowest points extracted by the lowest point extraction means 102. For convenience, the monochrome shaded relief map used in the present invention will be referred to as a "lowland monochrome shaded relief map" to distinguish it from conventional monochrome shaded relief maps. More specifically, the three-dimensional "ground model" generated by the model generating means 108 is shaded by shining light from a specific direction on the terrain, and the shading is displayed in grayscale to generate a lowland monochrome shaded relief map. The ground model here is a model typified by a DSM (Digital Surface Model) or a DEM (Digital Elevation Model), and the model generating means 108 generates this ground model based only on multiple lowest points (white points in FIG. 4). For example, the model generating means 108 can be configured to generate a ground model using a triangulated irregular network (TIN) consisting of multiple lowest points. Alternatively, instead of TIN, a nearest neighbor method, inverse distance weighting (IDW), Kriging, averaging, or the like may be used.

[0032] (Display means) The display means 104 can superimpose the lowland monochrome shaded relief map generated by the lowest point extraction means 102 and the candidate ground points obtained by automatic filtering. For example, a display provided in a computer device can be used. Figure 5 is a screen diagram that schematically shows the superimposed lowland monochrome shaded relief map and candidate ground points. Comparing Figure 5 with Figure 8, differences are particularly apparent in the area enclosed by the border. The presence of ground undulations can be seen in Figure 4, whereas this undulation cannot be seen in Figure 8. The presence of ground undulations indicates a high probability that the area is ground, and if there are no measurement points within that area, it can be inferred that there is a high probability that an over-extraction point has occurred. In other words, while it is difficult to find over-extraction points by superimposing a conventional monochrome shaded relief map (Figure 8) on candidate ground points, over-extraction points can be found by superimposing the lowland monochrome shaded relief map of the present invention (Figure 5) on candidate ground points.

[0033] (Processing flow) The main processing of the inspection system 100 will be described in detail below with reference to Fig. 6. Fig. 6 is a flow diagram showing an example of the flow of the main processing of the inspection system 100, in which the processing to be executed is shown in the center column, what is necessary for that processing is shown in the left column, and what results from that processing is shown in the right column.

[0034] As shown in Figure 6, first, automatic filtering is performed on the point cloud data using automatic filtering means 107 (Step 201 in Figure 6). This automatic filtering can be performed using various conventional software. Note that the inspection system 100 of the present invention can be configured to include the automatic filtering means 107, but can also be configured not to include it. When the automatic filtering means 107 is not included, the inspection system 100 performs a series of processes using candidate ground points obtained as a result of performing automatic filtering on the point cloud data.

[0035] When candidate ground points are obtained by performing automatic filtering on the point cloud data, a selected area is set within the measurement range using the area selection means 105 (Step 202 in FIG. 6), and a mesh size is set for the selected area using the size setting means 106 (Step 203 in FIG. 6). At this time, if two or more selected areas are selected within the measurement range, the size setting means 106 can also set a unique mesh size for each selected area. Note that the inspection system 100 of the present invention does not necessarily need to be equipped with the area selection means 105 or the size setting means 106, and therefore it is possible to skip the steps of setting the selected area (Step 202) and setting the mesh size (Step 203) and proceed to the next process.

[0036] Once the selected region and mesh size are set, the mesh setting means 101 sets a small region MS for the selected region with the specified mesh size (Step 204 in Fig. 6). At this time, if the region selection means is not set, the small region MS is set for the entire measurement range, and if the mesh size is not set, the small region MS is set with the specified (default) mesh size.

[0037] Once the small regions MS are set, the lowest point extraction means 102 extracts the lowest points from each small region MS (Step 205 in Fig. 6), and the model generation means 108 generates a ground model based on these lowest points (Step 206 in Fig. 6). Next, the shaded relief generation means 103 generates a lowland monochrome shaded relief map based on the ground model (Step 207 in Fig. 6), and displays the lowland monochrome shaded relief map and candidate ground points superimposed on the display means 104 (Step 208 in Fig. 6). Then, the operator finds over-extracted points while visually checking the lowland monochrome shaded relief map and candidate ground points superimposed on the display means 104, and performs manual filtering to return the found over-extracted points to ground points (Step 209 in Fig. 6).

[0038] 3. Inspection method Next, the inspection method of the present invention will be described with reference to Figure 7. Note that the inspection method of the present invention is a method for inspecting ground points after automatic filtering using the inspection system 100 described up to this point, and therefore, we will avoid overlapping explanations with those described for the inspection system 100 and will only describe the details unique to the inspection method of the present invention. In other words, the details not described here are the same as those described in "2. Inspection System".

[0039] Figure 7 is a flow diagram showing the flow of the main steps of the inspection method of the present invention. When carrying out the inspection method of the present invention, as shown in this figure, first, automatic filtering is performed on the point cloud data (Step 301 in Figure 7). After automatic filtering is performed on the point cloud data to obtain candidate ground points, a selected area is set within the measurement range using the area selection means 105 (Step 302 in Figure 7), and a mesh size is set for the selected area using the size setting means 106 (Step 303 in Figure 7). Note that in cases where the inspection system 100 of the present invention does not include the area selection means 105 or the size setting means 106, the process proceeds to the next step without setting the selected area (Step 302) and setting the mesh size (Step 303).

[0040] Once the selected region and mesh size have been set, a small region MS is set for the selected region with the specified mesh size using the mesh setting means 101 (Step 304 in Fig. 7). At this time, if the region selection means has not been set, a small region MS is set for the entire measurement range, and if the mesh size has not been set, a small region MS is set with a specified (default) mesh size.

[0041] Once the small region MS is set, the lowest point is extracted from the small region MS using the lowest point extraction means 102 (Step 305 in Fig. 7), and a ground model is generated based on the lowest point using the model generation means 108 (Step 306 in Fig. 7). Next, a lowland monochrome shaded relief map is generated based on the ground model using the shaded relief map generation means 103 (Step 307 in Fig. 7), and the lowland monochrome shaded relief map and candidate ground points are superimposed and displayed on the display means 104 (Step 308 in Fig. 7). Then, an operator visually checks the lowland monochrome shaded relief map and candidate ground points superimposed and displayed on the display means 104, and finds over-extracted points, and performs manual filtering to return the found over-extracted points to ground points (Step 309 in Fig. 7). [Industrial Applicability]

[0042] The inspection system and inspection method of the present invention can be used to obtain ground elevations in various locations, such as mountainous areas, coastal areas, and urban areas, and are particularly suitable for use in locations with forests. Since the present invention can obtain ground elevations with high accuracy, it can be effectively used in planning social infrastructure and disaster prevention plans, and is an invention that can be expected to not only be used industrially but also make a significant contribution to society. [Explanation of symbols]

[0043] 100 Inspection system of the present invention 101 (Inspection system) mesh setting means 102 (Inspection system) minimum point extraction means 103 (Inspection system) shaded relief map generator 104 (Inspection system) display means 105 Area selection means (of inspection system) 106 (Inspection System) Sizing Method 107 Automatic filtering means (of inspection systems) 108 (Inspection system) model generation means 109 (Inspection system) point cloud data storage means MS small area

Claims

1. A system for inspecting candidate ground points obtained by performing automatic filtering processing on point cloud data consisting of multiple measurement points, a mesh setting means for setting a plurality of small regions within the range of the point cloud data; a lowest point extraction means for extracting a lowest point having a minimum elevation value from among the measurement points included in the small area; a shaded area generating means for generating a monochrome shaded area based on the lowest point of each of the small areas; a display means for displaying the monochrome shaded relief map and the candidate ground points in a superimposed manner; By visually checking the monochrome shaded relief map and the candidate ground points superimposed on the display means, the measurement points that have been excessively removed by the automatic filtering process can be extracted. An inspection system characterized by:

2. further comprising area selection means for selecting a part of the range of the point cloud data as a selected area by an operator's operation; the shaded map generating means generates the monochrome shaded map based on the small area associated with the selected area selected by an operator operation.

2. The inspection system according to claim 1.

3. a size setting unit that is operated by an operator to set the size of the small region; When a plurality of the selection areas are selected, the size of the small area can be set for each of the selection areas.

3. The inspection system according to claim 2.

4. A method for inspecting the candidate ground points using the inspection system according to any one of claims 1 to 3, comprising: an automatic filtering step of extracting the candidate ground points by performing an automatic filtering process on the point cloud data; a mesh setting step of setting a plurality of small regions in the measurement range by the mesh setting means; a lowest point extraction step of extracting the lowest point from among the measurement points included in the small region by the lowest point extraction means; a shaded map generating step of generating the monochrome shaded map based on the lowest point of each of the small regions by the shaded map generating means; a superimposed display step of superimposing the monochrome shaded relief map and the candidate ground points on the display means; and an excess point extraction step of extracting the measurement points that have been excessively removed by the automatic filtering process by visually checking the monochrome shaded map and the candidate ground points superimposed on the display means. An inspection method characterized by:

Images