Point cloud data reduction device, point cloud data reduction method, and point cloud data reduction program

Abstract

A point cloud data reduction device is provided that can compress point cloud data at a high compression rate while suppressing loss of necessary information. [Solution] A point cloud discrimination unit 5 compares target point cloud data with base point cloud data and classifies the target point cloud into a fixed point cloud and a non-fixed point cloud. The point cloud discrimination unit 5 sets point clouds at corners or edges in the fixed point cloud as an exceptional point cloud, and sets point clouds in clusters obtained by clustering the fixed point cloud that represent thin objects narrower than a predetermined width or objects smaller than a predetermined size as exceptional point clouds. A downsampling unit 6 compresses the target point cloud data by downsampling point clouds from the fixed point cloud that are not set as exceptional point clouds.

Description

[Technical Field]

[0001] The present invention relates to a point cloud data reduction device, a point cloud data reduction method, and a point cloud data reduction program. [Background technology]

[0002] Point cloud data acquired by measuring three-dimensional space with a 3D laser scanner is large in volume, so it uses a lot of memory capacity to store the point cloud data and takes a long time to process. Therefore, point cloud data is sometimes compressed using methods such as PCC (Point Cloud Compression), the latest international standard for three-point cloud compression technology developed by MPEG (Moving Picture Experts Group). [Prior art documents] [Non-patent literature]

[0003] [Non-Patent Document 1] Free encyclopedia "Wikipedia", [Searched on August 29, 2024], Internet<URL: https: / / ja.wikipedia.org / wiki / Zstandard> [Non-patent document 2] Harris 3D: a robust extension of the Harris operator for interest point detection on 3D meshes, [Retrieved August 29, 2024], Internet<URL: https: / / www.ivan-sipiran.com / papers / SB11b.pdf> [Non-patent document 3] Chihiro Matsunuma and two others, University of Tokyo, "Detection of cylindrical and rectangular surfaces using edge extraction from large-scale point clouds," [Retrieved August 29, 2024], Internet<URL: http: / / www.ddm.mi.uec.ac.jp / papers / smt_10a-1.pdf> Summary of the Invention [Problem to be solved by the invention]

[0004] When point cloud data of around 150 MB is compressed using the data compression algorithm of mpeg-pcc-tmc13, which is a PCC method, or Zstandard, which is a lossless compression algorithm (see Non-Patent Document 1), the compression rate is, for example, about 30% to 60%, and the capacity reduction effect is limited. It is desired to compress point cloud data at a high compression rate while suppressing the loss of necessary information.

[0005] The present invention aims to provide a point cloud data reduction device, a point cloud data reduction method, and a point cloud data reduction program that can compress point cloud data at a high compression rate while suppressing the loss of necessary information. [Means for solving the problem]

[0006] A first aspect of the present invention provides a point cloud data reduction device comprising: a point cloud discrimination unit that compares target point cloud data indicating a target point cloud to be subjected to data compression with base point cloud data indicating a base point cloud as a reference, and classifies the target point cloud into a fixed point cloud, which is a point cloud that is included in common at the same position in both the base point cloud and the target point cloud, and a non-fixed point cloud, which is a point cloud that is not included in the base point cloud; sets point clouds at corners or edges in the fixed point cloud as exceptional point clouds that are not subject to downsampling; and sets point clouds within clusters obtained by clustering the fixed point cloud that indicate thin objects that are narrower than a predetermined width or objects that are smaller than a predetermined size as exceptional point clouds; and a downsampling unit that compresses the target point cloud data by downsampling point clouds from the fixed point cloud that are not set as exceptional point clouds.

[0007] A second aspect of the present invention provides a point cloud data reduction method for compressing the target point cloud data by comparing target point cloud data indicating a target point cloud to be subjected to data compression with base point cloud data indicating a base point cloud as a reference, and based on the result of the comparison between the target point cloud data and the base point cloud data, classifying the target point cloud into a fixed point cloud, which is a point cloud that is included in common at the same position in both the base point cloud and the target point cloud, and a non-fixed point cloud, which is a point cloud that is not included in the base point cloud, setting point clouds at corners or edges in the fixed point cloud as exception point clouds that are not subject to downsampling, setting point clouds within clusters obtained by clustering the fixed point cloud that indicate thin objects that are narrower than a predetermined width or objects that are smaller than a predetermined size as the exception point group, and downsampling point clouds from the fixed point cloud that are not set as the exception point group.

[0008] A third aspect of the present invention provides a point cloud data reduction program that causes a computer to execute the steps of comparing target point cloud data indicating a target point cloud to be subjected to data compression with base point cloud data indicating a base point cloud as a reference; classifying the target point cloud into a fixed point cloud, which is a point cloud that is included in common at the same position in both the base point cloud and the target point cloud, and a non-fixed point cloud, which is a point cloud that is not included in the base point cloud, based on the result of the comparison between the target point cloud data and the base point cloud data; setting point clouds at corners or edges in the fixed point cloud as exceptional point clouds that are not subject to downsampling; setting point clouds within clusters obtained by clustering the fixed point cloud that indicate thin objects narrower than a predetermined width or objects smaller than a predetermined size as the exceptional point group; and data compressing the target point cloud data by downsampling the point cloud of the fixed point cloud that is not set as the exceptional point group. [Effects of the Invention]

[0009] According to the point cloud data reduction device, point cloud data reduction method, and point cloud data reduction program of the present invention, it is possible to compress point cloud data at a high compression rate while suppressing loss of necessary information. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 is a block diagram showing a point cloud data reduction device according to an embodiment. [Figure 2A] FIG. 2A is a diagram conceptually illustrating an example of a base point cloud. [Figure 2B] FIG. 2B is a diagram conceptually illustrating an example of a target point group. [Figure 3] FIG. 3 is a diagram conceptually showing an exceptional point group set in the target point group shown in FIG. 2B by a point group discriminator included in a point group data reduction device according to an embodiment. [Figure 4A] FIG. 4A is a partial flowchart illustrating a point cloud data reduction process according to one embodiment. [Figure 4B] FIG. 4B is a partial flow chart that follows on from FIG. 4A. [Figure 4C] FIG. 4C is a partial flowchart that continues from FIG. 4B. [Figure 5] FIG. 5 is a block diagram showing an example of the configuration of a computer. DETAILED DESCRIPTION OF THE INVENTION

[0011] A point cloud data reduction device, a point cloud data reduction method, and a point cloud data reduction program according to an embodiment will be described below with reference to the accompanying drawings. Fig. 1 shows a point cloud data reduction device 100 according to an embodiment. The point cloud data reduction device 100 includes a point cloud data storage unit 1, a data selection unit 2, a base point cloud data storage unit 3, a target point cloud data storage unit 4, a point cloud discrimination unit 5, a downsampling unit 6, and a processed point cloud data writing unit 7.

[0012] The point cloud data storage unit 1 stores base point cloud data as a reference and target point cloud data. The point cloud data in this embodiment is assumed to be point cloud data acquired by measuring an office in three-dimensional space using a three-dimensional laser scanner (hereinafter referred to as a 3D laser scanner). The 3D laser scanner is sometimes referred to as LiDAR (Light Detection and Ranging). FIG. 2A conceptually shows a point cloud acquired by measuring an office using a 3D laser scanner at a certain date and time. The point cloud shown in FIG. 2A is assumed to be a base point cloud PC1.

[0013] As shown in Fig. 2A, the base point cloud PC1 includes point clouds representing pillars 11 to 13, a desk 14, a chair 15, a floor stand 16, a cabinet 17, a floor cabinet 18, and a potted plant 19. The point cloud data storage unit 1 stores base point cloud data representing the base point cloud PC1 shown in Fig. 2A.

[0014] FIG. 2B conceptually illustrates a point cloud acquired by a 3D laser scanner at a date and time later than the date and time at which the base point cloud PC1 was acquired. The point cloud illustrated in FIG. 2B is referred to as a target point cloud PC2. The target point cloud PC2 is a point cloud that is the target of data compression by the point cloud data reduction device 100. As illustrated in FIG. 2B, the target point cloud PC2 includes point clouds representing a trash can 21 and luggage 22 in addition to the point clouds included in the base point cloud PC1. In the target point cloud PC2, the potted plant 19 that was present in the base point cloud PC1 has disappeared. The point cloud data storage unit 1 stores target point cloud data representing the target point cloud PC2 illustrated in FIG. 2B.

[0015] The point cloud data storage unit 1 may further store base point cloud data indicating a base point cloud different from the base point cloud PC1 shown in Fig. 2A. A base point cloud different from the base point cloud PC1 is, for example, a point cloud indicating an office in a state where no furniture is arranged except for the point clouds indicating the desk 14, chair 15, floor stand 16, cabinet 17, floor cabinet 18, and potted plant 19 in the base point cloud PC1.

[0016] 1, a user operates a data selection unit 2 to select a base point cloud PC1 and input it into a base point cloud data storage unit 3, and to select a target point cloud PC2 and input it into a target point cloud data storage unit 4. The base point cloud data storage unit 3 and the target point cloud data storage unit 4 store base point cloud data indicating the input base point cloud PC1 and target point cloud data indicating the input target point cloud PC2, respectively.

[0017] The point cloud discriminator 5 compares the base point cloud data with the target point cloud data to determine the fixed point clouds included in the target point cloud PC2. A fixed point cloud is a point cloud representing fixed objects that are included in both the base point cloud PC1 and the target point cloud PC2 at the same position. In the target point cloud PC2 shown in FIG. 2B, the point cloud discriminator 5 discriminates the point clouds representing pillars 11-13, desk 14, chair 15, floor lamp 16, cabinet 17, and floor cabinet 18 as fixed point clouds. The point cloud discriminator 5 discriminates the point clouds representing trash can 21 and luggage 22, which are only present in the target point cloud PC2, as non-fixed point clouds that may represent moving objects. The point cloud discriminator 5 may also discriminate the point cloud representing a missing potted plant 19 as non-fixed point clouds.

[0018] The point cloud discrimination unit 5 sets the point cloud of corners or edges in the fixed point cloud as an exceptional point cloud that is an exception to the downsampling target described below. Corners or edges can be detected using a known method. Both corners and edges may be detected using a single method, or corners and edges may be detected using separate methods. As an example, a Harris corner detector can be used to detect corners. The technology described in Non-Patent Document 2, which applies the Harris corner detector to three-dimensional point clouds, can also be used. The technology described in Non-Patent Document 3 can be used to extract edges from three-dimensional point clouds.

[0019] The processed point cloud data (described later) obtained by compressing the target point cloud data representing the target point cloud PC2 may be used, for example, as map data for estimating the self-position of a service robot, an example of a mobile machine, when the robot moves around an office. The point clouds at corners or edges in the fixed point cloud are important point clouds when the service robot estimates its self-position. Therefore, the point cloud discriminator 5 sets the point clouds at corners or edges in the fixed point cloud as exceptional point clouds.

[0020] An example of a service robot is a cleaning robot. The service robot is equipped with a sensor for estimating its own position. The sensor is one of a camera, a gyro sensor, a geomagnetic sensor, an acceleration sensor, an inertial measurement unit, a tire pulse sensor, a LiDAR, and a radar sensor (RADAR: Radio Detection and Ranging), or a combination of any two or more of these sensors.

[0021] For service robots used outdoors, a GNSS sensor is used to detect its position by receiving GNSS signals transmitted from three or more satellites for the Global Navigation Satellite System (GNSS). Typically, the GNSS is the Global Positioning System (GPS), and the GNSS sensor is a GPS sensor.

[0022] Furthermore, the point cloud discriminator 5 extracts, from the fixed point cloud, point clouds that represent narrow objects narrower than a predetermined width and point clouds that represent objects smaller than a predetermined size, and sets these extracted point clouds as exceptional point clouds. In the target point cloud PC2 shown in Figure 2B, the point cloud discriminator 5 sets the point cloud that represents the floor stand 16 as the exceptional point cloud.

[0023] Point clouds representing thin objects can be extracted using known methods such as curvature analysis, which is an analysis of local geometric features, or normal vector variance, which is also an analysis of local geometric features, or skeleton extraction. Curvature analysis calculates the local curvature of a point cloud and identifies points with high curvature as candidates for thin objects. Normal vector variance calculations calculate the normal vectors of each point and evaluate their variance. Skeleton extraction uses a skeletonization algorithm to identify thin objects by extracting a skeleton from point cloud data.

[0024] Point clouds representing small objects can be extracted using known methods such as clustering called DBSCAN (Density-Based Spatial Clustering of Applications with Noise), voxel grid filtering, and region growing algorithms.

[0025] In the clustering method, point cloud data is clustered, the size of each cluster is evaluated, and small clusters are determined to be point clouds that represent small objects. In the voxel grid filtering method, point cloud data is divided into a voxel grid, the number of points in each voxel is evaluated, and voxels with a small number of points are determined to be point clouds that represent small objects. In the region growing algorithm method, clusters are formed by adding points adjacent to a specific seed point, the size of the formed clusters is evaluated, and small clusters are determined to be point clouds that represent small objects.

[0026] The point cloud discriminator 5 may extract a point cloud representing a thin object by combining a plurality of methods, or may extract a point cloud representing a small object by combining a plurality of methods. There are no limitations on how the point cloud discriminator 5 extracts a point cloud representing a thin object or a point cloud representing a small object.

[0027] As described above, the point cloud discriminator 5 sets, as exceptional point clouds, point clouds at corners or edges in the fixed point cloud included in the target point cloud PC2 shown in FIG. 2B, and point clouds in clusters of the fixed point cloud that represent thin objects narrower than a predetermined width or objects smaller than a predetermined size. Corner or edge point clouds are important point clouds when the service robot estimates its own position. Thin or small objects have a small number of points, so they should not be downsampled to ensure that their size is maintained in the processed point cloud data.

[0028] Figure 3 conceptually shows the exceptional point clouds that the point cloud discriminator 5 has set in the target point cloud PC2 shown in Figure 2B. In Figure 3, as indicated by black circles at the corners and edges of the fixed point cloud, the point cloud discriminator 5 sets the point clouds at the corners or edges of the fixed point cloud as the exceptional point cloud. As indicated by the dashed line, the point cloud discriminator 5 sets the point cloud representing the floor stand 16 from the fixed point cloud as the exceptional point cloud. As indicated by the dashed line, the point cloud discriminator 5 sets the point clouds representing the trash can 21 and luggage 22 as the non-fixed point cloud.

[0029] Returning to Figure 1, target point cloud data indicating target point cloud PC2, which has been classified into fixed point clouds and non-fixed point clouds and in which some of the fixed point clouds have been determined to be exceptional point clouds, is supplied to a downsampling unit 6 along with the determination information. The downsampling unit 6 downsamples the points in the input target point cloud data excluding the non-fixed point clouds and exceptional point clouds, i.e., the points in the fixed point cloud other than the exceptional point cloud. The downsampling unit 6 supplies the downsampled target point cloud data to a processed point cloud data writing unit 7 as processed point cloud data.

[0030] The processed point cloud data writing unit 7 writes the processed point cloud data to the point cloud data storage unit 1. It is not essential to write the processed point cloud data to the point cloud data storage unit 1. The processed point cloud data may be transmitted to an external device of the point cloud data reduction device 100.

[0031] The proportion of corner or edge points in a fixed point cloud is significantly smaller than the proportion of points other than corners and edges in the fixed point cloud. The point cloud data reduction device 100 downsamples the points other than corners and edges in the fixed point cloud, which make up a significantly larger proportion, making it possible to compress the point cloud data at a high compression rate. Moreover, the point cloud data reduction device 100 does not downsample non-fixed point clouds or exceptional point clouds, thereby preventing the loss of necessary information. When a service robot uses the processed point cloud data as map data for estimating its own position, the information for estimating its own position is maintained even when the data is compressed at a high compression rate, so the movement of the service robot is not hindered.

[0032] If the 3D laser scanner acquires further point cloud data of the office after the date and time at which the target point cloud PC2 shown in Figure 2B was acquired, the target point cloud data indicating the target point cloud PC2 shown in Figure 2B can be used as the base point cloud data, and the latest point cloud data can be used as the target point cloud data.

[0033] The operation of the point cloud data reduction device 100 and the point cloud data reduction method executed by the point cloud data reduction device 100 will be described in more detail using the flowcharts shown in Figures 4A to 4C. In Figure 4A, when processing starts, the base point cloud data storage unit 3 reads and stores base point cloud data in step S1. The target point cloud data storage unit 4 reads and stores target point cloud data in step S2. The order of steps S1 and S2 may be reversed, or they may be performed simultaneously.

[0034] In step S3, surrounded by a dashed line, the point cloud data reduction device 100 sequentially determines whether or not each point included in the target point cloud data is a fixed point. Specifically, in step S31, the point cloud discriminator 5 compares the target point cloud with the base point cloud, and in step S32, determines whether or not the selected point exists in the base point cloud. If the selected point exists in the base point cloud (YES), in step S33, the point cloud discriminator 5 assigns a fixed point attribute to the selected point.

[0035] Next, in step S34, the point cloud discriminator 5 determines whether the selected point is an important point for self-position estimation, such as a corner or edge of an object. If the selected point is an important point (YES), the point cloud discriminator 5 assigns an exceptional attribute to the selected point in step S35. If the point selected in step S32 does not exist in the base point cloud (NO), the point cloud discriminator 5 assigns an exceptional attribute to the selected point in step S35. The point cloud discriminator 5 repeats the same process until the determination of all points included in the target point cloud data is completed.

[0036] In step S4, the point cloud discriminator 5 generates a fixed point cloud consisting only of fixed points, and in step S5, generates a non-fixed point cloud consisting of points other than fixed points. In step S6, the point cloud discriminator 5 clusters the fixed point cloud. By clustering the fixed point cloud, the fixed point cloud is grouped based on the similarity between adjacent points, and the fixed point cloud is divided into multiple clusters.

[0037] In FIG. 4B, in step S7 surrounded by a dashed line, the point cloud discriminator 5 executes a process of selecting, for each cluster in turn, whether or not to set an exceptional point cloud based on the shape determination of the cluster.

[0038] Specifically, the point cloud discriminator 5 analyzes the shape of the selected cluster in step S71, and determines whether the selected cluster is a thin or small cluster in step S72. If the selected cluster is a thin or small cluster (YES), the point cloud discriminator 5 repeats the process of assigning exceptional attributes to points in the selected cluster in step S73, and assigns exceptional attributes to all points in the cluster.

[0039] If the cluster selected in step S72 is not a thin or small cluster (NO), no exceptional attribute is assigned. The point cloud discriminator 5 repeats the same process until the process for all clusters is completed.

[0040] In FIG. 4C , the downsampling unit 6 performs downsampling processing on the fixed point cloud in step S8, which is surrounded by a dashed line. Specifically, in step S81, the downsampling unit 6 determines whether or not an exceptional attribute is assigned to the selected point of the fixed point cloud. If the selected point does not have an exceptional attribute (NO), the downsampling unit 6 generates a random number in step S82. In step S83, the downsampling unit 6 determines whether or not the random number is equal to or greater than a threshold. If the random number is equal to or greater than the threshold (YES), the downsampling unit 6 deletes the selected point from the fixed point cloud in step S84.

[0041] If an exceptional attribute is assigned to the selected point in step S81 (YES), or if the random number is not equal to or greater than the threshold in step S83, the downsampling unit 6 does not delete the selected point from the fixed point group. The downsampling unit 6 repeats the same process for all points in the fixed point group until the process of selecting whether to delete a point based on the comparison result between the random number and the threshold is completed.

[0042] In step S9, the downsampling unit 6 combines the fixed point cloud and the non-fixed point cloud to generate processed point cloud data. In step S10, the processed point cloud data writing unit 7 reads the processed point cloud data into the point cloud data storage unit 1, and the process ends.

[0043] The method by which the downsampling unit 6 selects points to be deleted from the fixed point cloud is not limited to the method using random numbers. The downsampling unit 6 may determine whether the distance between the selected point and the center of gravity of a voxel is equal to or greater than a threshold, and delete the selected point from the fixed point cloud if the distance is equal to or greater than the threshold. The downsampling unit 6 may determine whether the selected point is the center of gravity of a voxel, and delete the selected point from the fixed point cloud if it is not the center of gravity of a voxel.

[0044] When the downsampling unit 6 uses random numbers to determine points to be deleted from the fixed point cloud and deletes the points, there is an advantage that the deleted points are not fixed. When the downsampling unit 6 deletes points whose distance from the voxel centroid is equal to or greater than a threshold, the target point cloud can be downsampled while maintaining multiple points, including points located at the centroid, which are important points. When the downsampling unit 6 deletes points other than the voxel centroid, the target point cloud can be downsampled at a higher compression rate while maintaining only points located at the centroid, which are important points.

[0045] The point cloud data reduction device 100 shown in Fig. 1 may be configured as hardware, or may be functionally configured by a computer executing a point cloud data reduction program. Fig. 5 shows the configuration of a general-purpose computer. The computer includes a central processing unit (hereinafter referred to as CPU) 31, a main memory 32, a non-transitory storage medium 33, an input device 34, and an output device 35. The point cloud data reduction device 100 can be realized by the CPU 31 executing the point cloud data reduction program stored in the non-transitory storage medium 33. The point cloud data reduction program can also be distributed via a network.

[0046] The point cloud data reduction program stored in the non-transitory storage medium 33 is loaded into the main memory 32, and the CPU 31 executes the point cloud data reduction program. The point cloud data storage unit 1 may be the non-transitory storage medium 33, or may be another storage medium not shown. The data selection unit 2 may be configured with an input device 34, or an input device 34 and an output device 35. The input device 34 is a keyboard or a mouse, etc., and the output device 35 is, for example, a display.

[0047] The CPU 31 displays base point cloud candidates and target point cloud candidates on the display, and the user selects the base point cloud and target point cloud using the input device 34. This makes it possible to easily select the base point cloud data to be stored in the base point cloud data storage unit 3 and the target point cloud data to be stored in the target point cloud data storage unit 4.

[0048] The base point cloud data storage unit 3 and the target point cloud data storage unit 4 may be main memory 32 or may be memory built into the CPU 31. The CPU 31 functions as a point cloud discrimination unit 5, a downsampling unit 6, and a processed point cloud data writing unit 7. The flowcharts shown in FIGS. 4A to 4C also represent the processing that the point cloud data reduction program causes the CPU 31 to execute. In FIG. 5, a GPU (Graphics Processing Unit) may be used instead of the CPU 31.

[0049] The computer may include a communication device that transmits the processed point cloud data to an external device such as a service robot via a network such as the Internet.

[0050] The processed point cloud data may include information distinguishing between fixed and non-fixed point clouds. If the processed point cloud data used by the service robot as map data includes information distinguishing between fixed and non-fixed point clouds, the following secondary effect can be obtained: The service robot can change its behavior depending on whether it is a fixed object represented by a fixed point cloud or a moving object represented by a non-fixed point cloud.

[0051] Take the case where the service robot is a cleaning robot. When a moving object moves after cleaning its surroundings, the location where the moving object was located will not be cleaned. Therefore, the cleaning robot can be provided with information that no non-fixed point cloud exists in the latest target point cloud, and can spend more time cleaning the location where the non-fixed point cloud existed in the previous target point cloud than other locations.

[0052] The present invention is not limited to the above-described embodiment, and various modifications are possible without departing from the gist of the present invention. [Explanation of symbols]

[0053] 1 Point cloud data storage unit 2 Data selection section 3 Base point cloud data storage unit 4. Target point cloud data storage unit 5 Point cloud discriminator 6 Downsampling section 7. Processed point cloud data writing section 31 Central Processing Unit 32 main memory 33 Non-transitory storage medium 34 Input Devices 35 Output Device 100 Point Cloud Data Reduction Device

Claims

1. a point cloud discrimination unit that compares object point cloud data indicating an object point cloud to be subjected to data compression with base point cloud data indicating a base point cloud as a reference, and classifies the object point cloud into a fixed point cloud, which is a point cloud that is included in common at the same positions in the base point cloud and the object point cloud, and a non-fixed point cloud, which is a point cloud that is not included in the base point cloud, and sets point clouds at corners or edges in the fixed point cloud as exceptional point clouds that are not subject to downsampling, and sets point clouds in clusters obtained by clustering the fixed point cloud that indicate thin objects that are narrower than a predetermined width or objects that are smaller than a predetermined size as exceptional point clouds; a downsampling unit that compresses the target point cloud data by downsampling a point cloud that is not set as the exceptional point cloud among the fixed point cloud; A point cloud data reduction device comprising:

2. The point cloud data reduction device according to claim 1 , wherein the downsampling unit determines whether or not to delete a point based on a result of comparing a generated random number with a threshold value, and downsamples the fixed point cloud.

3. The point cloud data reduction device according to claim 1 , wherein the downsampling unit downsamples the fixed point cloud by deleting points whose distance from the center of gravity of a voxel is equal to or greater than a threshold.

4. The point cloud data reduction device according to claim 1 , wherein the downsampling unit downsamples the fixed point cloud by deleting points that are not located at the center of gravity of voxels.

5. comparing target point cloud data indicating the target point cloud to be subjected to data compression with base point cloud data indicating the base point cloud as a reference; Based on a result of the comparison between the target point cloud data and the base point cloud data, the target point cloud is classified into a fixed point cloud, which is a point cloud that is commonly included at the same position in the base point cloud and the target point cloud, and a non-fixed point cloud, which is a point cloud that is not included in the base point cloud; setting a corner or edge point group in the fixed point group as an exceptional point group that is not subject to downsampling; Among the clusters obtained by clustering the fixed point cloud, a point cloud in a cluster indicating a thin object having a width narrower than a predetermined width or an object smaller than a predetermined size is set as the exceptional point cloud; The target point cloud data is compressed by downsampling the point cloud that is not set as the exceptional point cloud among the fixed point cloud. Point cloud data reduction methods.

6. On the computer, A step of comparing target point cloud data indicating a target point cloud to be subjected to data compression with base point cloud data indicating a base point cloud as a reference; based on a comparison result between the target point cloud data and the base point cloud data, classifying the target point cloud into a fixed point cloud, which is a point cloud that is commonly included at the same position in the base point cloud and the target point cloud, and a non-fixed point cloud, which is a point cloud that is not included in the base point cloud; setting a corner or edge point group in the fixed point group as an exceptional point group that is not subject to downsampling; a step of setting, as the exceptional point group, a point group in a cluster that indicates a thin object having a width narrower than a predetermined width or an object that is smaller than a predetermined size, among clusters obtained by clustering the fixed point group; compressing the target point cloud data by downsampling points that are not set as the exceptional point cloud from the fixed point cloud; A point cloud data reduction program that executes the above.

Images