JP2019200760A - Automatic drawing method from point group, and point group processing program - Google Patents

Abstract

To provide a method for automatically acquiring or automatically recognizing information provided for drawing creation from a point group of a three-dimensional coordinates.SOLUTION: A method automatically extracts a point group showing three-dimensional coordinates of a solid structure surface of a builing or the like acquired by a three-dimensional measurement device, and a plane, an inclined plane, or an edge from a point group showing three-dimensional coordinates of a natural object surface. An area determination method extracts the plane, the inclined plane, or an edge region. The method includes: each point feature amount determination step for making a feature amount acquired by applying a three-dimensional spatial filter to height information on each of all points in the point group into a feature amount at each point; a step for determining a plane, inclined plane, or edge feature amount acquired by normalizing the feature amount; and a section extraction step for executing area division and labeling processing on the feature amount by threshold processing.SELECTED DRAWING: Figure 4

Description

  The present invention relates to automatic acquisition or automatic recognition of information used for drawing creation from a point group of three-dimensional coordinates obtained by a point group acquisition unit.

  Conventionally, when measuring a three-dimensional structure such as a building and a natural object with a three-dimensional measurement device, for example, a drone or a laser scanner is used, and three-dimensional measurement data can be acquired. A technique for extracting features from three-dimensional measurement data is disclosed (see Patent Documents 1 and 2).

Japanese Patent No. 5480914 JP 2017-49737 A

  According to the configuration of Patent Document 1, an edge or the like is extracted as a two-dimensional image from a three-dimensional point group, and occlusion (a state in which the object behind is obstructed by an object in front) is excluded as a non-surface area. It is said. Further, since there is a stage of handling as a two-dimensional image as intermediate processing, it is difficult to view the three-dimensional space from a multifaceted viewpoint.

  According to the configuration of Patent Document 2, a simultaneous equation that describes two planes is generated based on the result of obtaining a normal vector for a three-dimensional point group, and an edge is extracted by obtaining a solution of the simultaneous equations. Yes. However, the natural object which is the subject of the present invention is not always approximated by a plane and is difficult to apply.

  An object of the present invention is to provide a method for automatically extracting a plane, a slope, or an edge from a point group of three-dimensional coordinates, and a method for determining a plane, a slope, or an edge region by this automatic extraction method. To do.

In order to solve the above-described problem, the method of automatically drawing from a point cloud according to the present invention is a point cloud indicating a three-dimensional coordinate of a surface of a three-dimensional structure such as a building acquired by a three-dimensional measuring device, and a three-dimensional surface of a natural object. A method for automatically extracting a plane, a slope, or an edge from a point group indicating coordinates, and a region determination method for extracting a plane, a slope, or an edge region,
Each point feature amount determination step in which the feature amount obtained by applying a three-dimensional spatial filter to the height information of all the points in the point group is the feature amount at each point;
A plane, slope, or edge feature value determination step obtained by normalizing the feature value; and
This is a section extraction step for performing region division and labeling processing on the feature amount by threshold processing.

According to the present invention, the following effects can be obtained for each claim.

  According to the first aspect of the present invention, from a point group indicating a three-dimensional coordinate of a three-dimensional structure such as a building acquired by a three-dimensional measuring device and a point group indicating a three-dimensional coordinate of a natural object surface, By automatically extracting the slope or the edge, the drawing work based on the point cloud can be facilitated. Conventionally, in drawing work, the point cloud is manually sighted on the flat display to find the type and edge of the object, and the drawing is performed. However, since the 3D point cloud is mapped to the flat display, the depth is recognized. This tends to be physically difficult and contributes to an increase in labor and time for drawing. According to the present invention, the above-described problem can be solved by automatically extracting a plane, a slope, or an edge.

  According to the second aspect of the present invention, a feature amount indicating a plane, a slope, or an edge can be obtained from a three-dimensional point group.

  According to this invention of Claim 3, a plane, a slope, or an edge can be distinguished based on the said feature-value.

  According to the fourth aspect of the present invention, the feature amount obtained by the second aspect of the invention can be displayed.

  According to the present invention described in claim 5, the region determination result obtained in the invention of claim 3 can be displayed.

  According to the sixth aspect of the present invention, there is provided a point cloud processing program comprising the first aspect, the second aspect, the third aspect, the fourth aspect, or the fifth aspect.

It is a flowchart of a point cloud processing program. It is an image figure of a three-dimensional spatial filter. It is explanatory drawing which shows the weight of a three-dimensional spatial filter. It is a drawing substitute photograph which shows the color classification result of point cloud data.

  Hereinafter, an embodiment of the present invention will be described with reference to the accompanying drawings.

  The point cloud processing program according to the embodiment is hardware including arithmetic processing means such as a CPU (Central Processing Unit), and storage means such as a ROM (Read Only Memory) and a RAM (Random Access Memory), for example. The calculation means and specific means for cooperating with the point group processing program which is software realized on the calculation means are provided.

  A processor is a dedicated or general-purpose CPU, a GPU (Graphics Processing Unit), an application specific integrated circuit (ASIC (Application Specific Integrated Circuit)), a programmable logic device (SPLD (Simple Programmable Logic Device)). It is an arithmetic processing means capable of executing programs such as CPLD (Complex Programmable Logic Device) and field programmable gate array (FPGA (Field Programmable Gate Array)). The processor implements various functions by reading and executing a program stored in the storage means.

  Instead of storing and executing the program in the storage circuit, the program may be directly incorporated as a circuit constituting the processor.

  Alternatively, an arithmetic processing circuit capable of executing a program may be configured by combining a plurality of independent processors. A storage circuit for storing a program may be provided for each processor individually, or programs corresponding to functions of a plurality of processors may be provided in an aggregated manner.

  The point cloud processing program may be incorporated into the point cloud information processing obtained by the point cloud information acquisition means. For example, the point cloud information may be acquired by a point survey information acquisition means such as a drone surveying, a ground laser scanner, or a hand scanner and used for the analysis.

  FIG. 1 is a flowchart of the point cloud processing program. S101 is a step of inputting a point cloud, and connecting means for connecting to an external storage means such as a flash memory, a magnetic storage means such as a hard disk, a LAN (Local Area Network), or a WAN (Wide Area Network). It is.

  S102 is a step of extracting feature amounts based on the point group input in S101 and acquiring features of the point group indicating a plane, a slope, or an edge.

  S103 is a step of normalizing (−1 to 1) the point cloud feature values obtained in S102. By applying normalization, a plane, a slope, or an edge can be selectively handled.

  S104 is a step of performing labeling by classifying the cut point group in the data section defined in advance based on the feature amount obtained in S103.

  S105 is a step of outputting the classification result obtained in S104 to an external storage means, a magnetic storage means, a connection means such as a LAN, or a display means such as a liquid crystal display based on the feature amount normalization result obtained in S103. It is a process of displaying data. For example, if the point cloud distribution is closer to a flat surface (as the evaluation value is closer to −1), it is displayed in cold blue, etc., and as the slope is steep (as the evaluation value is closer to 1), it is expressed in warm red. By displaying on a display means such as a liquid crystal display, a flat surface, a slope, or an edge can be visually understood.

  Hereinafter, the configuration of S102 in FIG. 1 will be described in detail. In step S102, a feature amount is calculated by applying a three-dimensional spatial filter to each point of the point group. Equation 1 is a weight function of the three-dimensional spatial filter. FIG. 2 shows an image diagram of the three-dimensional spatial filter. FIG. 3 shows the applied weights. Here, the weight function is defined for each of the x-axis, y-axis, and z-axis, and the input parameters x, y, and z indicate the coordinates of each point in the point group.

  Equation 2 is a function indicating the height of the point group. The height is, for example, a height value such as an altitude value, and the type is not limited. Equation 3 calculates the height distribution of each target point and its neighboring points on each axis (x axis, y axis, z axis) as an evaluation value. Equation 4 converts the evaluation value (three-dimensional vector) of the height change of each axis into an evaluation value (scalar).

The processing so far is performed for each point in the point group, and an evaluation value at each point is obtained. That is, a point group having a relatively small change in height (such as a plane when compared with a slope) has a relatively low evaluation value and a relatively large height change (such as a slope when compared with a plane). The point cloud has a relatively high evaluation value.

Hereinafter, the configuration of S103 in FIG. 1 will be described in detail. The evaluation value calculated in S102 in the preceding stage is a relatively high evaluation value with respect to an outlier (generated due to point cloud noise or the like). Further, since the maximum value depends on the distribution status of the point cloud input in S101, it is assumed that the distribution of the entire evaluation value is a normal distribution, and the evaluation value is set to −1 by applying the normalization based on Equation 5. The plane, the slope, or the edge portion can be extracted by mapping between ˜1.

Hereinafter, the configuration of S104 in FIG. 1 will be described in detail. In the normalized evaluation values calculated by the pre-processing S103, by extracting only points that become evaluation values within a specific range from the set of evaluation values of all points as shown in Equation 6, a plane or a slope, Alternatively, each edge can be extracted.

  Hereinafter, the configuration of S105 in FIG. 1 will be described in detail. The extraction result in the previous step S104 can be output to an external storage means, a magnetic storage means, a connection means such as a LAN, or output to a display means such as a liquid crystal display. Data output or data display may be performed simultaneously, or may be selectively performed upon receipt of a user selection, or may be selected by some trigger means. The point group input in S101 is a three-dimensional coordinate of x coordinate, y coordinate, and z coordinate, and may not have color information (RGB) and luminance information. Color information can be added to the input point group based on the normalized feature amount calculated in S103. For example, if the point cloud distribution is closer to the plane (as the evaluation value is closer to −1), it is displayed in cold blue or the like, and as the slope is steep (as the evaluation value is closer to 1), it is displayed in warm red. This makes it possible to visually express the point cloud distribution. The display can be performed by a display means such as a liquid crystal display. An embodiment is shown in FIG.

S101 Point group input S102 Plane / slope / edge feature quantity calculation S103 Feature quantity normalization S104 Classification / grouping S105 Data output

Claims (6)

  This is an automatic drawing method from a point cloud, and automatically detects a plane, a slope, or an edge from a point cloud that indicates the three-dimensional coordinates of the surface of a three-dimensional structure such as a building or a natural object acquired by a three-dimensional measuring device. A method of automatic operation drawing from a point cloud characterized by extracting.   2. The automatic operation drawing method according to claim 1, further comprising a feature quantity extraction step of obtaining a plane, slope, or edge feature quantity from the point group.   3. The automatic operation drawing method according to claim 2, further comprising a region determining step of extracting a plane, a slope, or an edge region based on the feature amount.   A plane, slope, or edge feature quantity display method for displaying the feature quantity according to claim 2.   A region determination display method for displaying the region determination result of claim 3.   A point cloud processing program comprising claim 1, claim 2, claim 3, claim 4, or claim 5.