Information processing device, information processing system, information processing method, and program

By segmenting and analyzing point cloud data based on distance and angle distributions, the method addresses the issue of fixed threshold failures, ensuring accurate data removal and maintaining essential information integrity.

JP7593417B2Active Publication Date: 2024-12-03NEC CORP
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
JP2022574940
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2021-01-14
Publication Date
2024-12-03
Estimated Expiration
2041-01-14

AI Technical Summary

Technical Problem

Existing methods for filtering point cloud data using fixed threshold values fail to accurately remove unnecessary data when the distance between the sensor and the measurement target is not constant, leading to the removal of originally necessary data.

Method used

The method involves acquiring first and second point cloud data at different measurement points, dividing the data into segments, calculating the distribution of distances or angles within each segment, and removing outliers based on these distributions to filter unnecessary data.

Benefits of technology

This approach effectively removes unnecessary data from point cloud data without eliminating essential information, stabilizing brightness values, and preventing the loss of important measurement data, especially in environments with varying distances or angles.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007593417000001
    Figure 0007593417000001
  • Figure 0007593417000002
    Figure 0007593417000002
  • Figure 0007593417000003
    Figure 0007593417000003
Patent Text Reader

Abstract

An information processing device (100) according to the present disclosure is provided with: a first acquisition unit (110) that acquires first point group data indicating distances between a sensor and a measurement object at first measurement spots along a path; a second acquisition unit (120) that acquires second point group data indicating distances between the sensor and the measurement object at second measurement spots different from the first measurement spots along the path; a division unit (130) that divides, into one or more segments, point group data including the first point group data and the second point group data; a distribution calculation unit (140) that calculates a distribution of the point group data in each of the segments; a removal unit (150) that removes an outlier from the point group data in the segments on the basis of the distribution.
Need to check novelty before this filing date? Find Prior Art

Description

[Technical Field]

[0001] The present disclosure relates to an information processing device, an information processing system, an information processing method, and a non-transitory computer-readable medium. [Background technology]

[0002] Sensors such as LiDAR can calculate the distance to each measurement point based on the time it takes for the laser to be emitted and received. By using such sensors while moving, it is possible to obtain the distance to a measurement object such as a tunnel and its shape.

[0003] For example, Patent Document 1 discloses a wireless communication system that collects point cloud data including measurement data at multiple measurement points measured by sensors installed on the roadside or in vehicles. The collected measurement data is transmitted to a server, where it can be used to detect vehicles and pedestrians and provide driving assistance. Furthermore, this wireless communication system has a problem in that if the measurement data also includes many measurement points of static objects, such as fixed objects like buildings and trees, or the ground, the communication traffic transmitted from the transmitting device to the server increases. To solve this problem, Patent Document 1 also proposes removing measurement data from measurement points that form trajectories longer than a predetermined threshold length and measurement data from measurement points that fall within a predetermined angle range from the point cloud data. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] International Publication No. 2019 / 131075 Summary of the Invention [Problem to be solved by the invention]

[0005] However, when filtering using such a fixed threshold, there is a problem that even necessary data may be removed if the distance between the object to be measured and the sensor is not constant, such as when the sensor is not moving parallel to the object.

[0006] The present disclosure has been made to solve such problems, and aims to provide an information processing device, an information processing system, an information processing method, and a non-transitory computer-readable medium that remove unnecessary data from point cloud data measured by moving a sensor without removing data that is actually necessary. [Means for solving the problem]

[0007] An information processing device according to a first aspect of the present disclosure includes: a first acquisition unit that acquires first point cloud data indicating a distance between the sensor and the measurement object at a first measurement point along the route; a second acquisition unit that acquires second point cloud data indicating a distance between a sensor and the measurement object at a second measurement point along the route that is different from the first measurement point; a dividing unit that divides point cloud data including the first point cloud data and the second point cloud data into one or more segments; a distribution calculation unit that calculates a distribution of the point cloud data within the segment; a removal unit that removes outliers from the point cloud data within the segment based on the distribution; Equipped with.

[0008] An information processing system according to a second aspect of the present disclosure includes: a sensor that irradiates a laser beam from the sensor onto a measurement object while moving the sensor along a path, and receives reflected light from the sensor to acquire point cloud data that measures the distance from the sensor to the measurement object; a first acquisition unit that acquires first point cloud data indicating a distance between the sensor and the measurement object at a first measurement point along the route; a second acquisition unit that acquires second point cloud data indicating a distance between the sensor and the measurement object at a second measurement point along the route that is different from the first measurement point; a dividing unit that divides point cloud data including the first point cloud data and the second point cloud data into a plurality of segments; a distribution calculation unit that calculates a distribution of the point cloud data within the segment; and a removal unit that removes outliers from the point cloud data within the segment based on the distribution.

[0009] An information processing method according to a third aspect of the present disclosure includes: acquiring first point cloud data indicating a distance between the sensor and the object to be measured at a first measurement point along the path; acquiring second point cloud data indicating a distance between a sensor and the measurement object at a second measurement point along the route that is different from the first measurement point; Dividing point cloud data including the first point cloud data and the second point cloud data into one or more segments; Calculating a distribution of the point cloud data within the segment; Based on the distribution, outliers are removed from the point cloud data within the segment.

[0010] A non-transitory computer-readable medium storing a program according to a fourth aspect of the present disclosure includes: acquiring first point cloud data indicating a distance between the sensor and the object at a first measurement point along the route; acquiring second point cloud data indicating a distance between a sensor and the object to be measured at a second measurement point along the route that is different from the first measurement point; Dividing point cloud data including the first point cloud data and the second point cloud data into one or more segments; A process of calculating a distribution of the point cloud data within the segment; and removing outliers from the point cloud data within the segment based on the distribution. [Effects of the Invention]

[0011] The present disclosure makes it possible to provide an information processing device or the like that removes unnecessary data from point cloud data measured by moving a sensor without removing data that is actually necessary. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a block diagram showing a configuration of an information processing device according to a first embodiment. [Figure 2] 1 is a flowchart showing an information processing method according to the first embodiment. [Figure 3] FIG. 10 is a diagram showing an example of measuring the inner wall of a tunnel using a sensor mounted on a moving object. [Figure 4] FIG. 10 is a diagram illustrating a first reason why the luminance value is unstable. [Figure 5] FIG. 10 is a diagram showing the distribution of luminance values ​​before filtering is performed. [Figure 6] FIG. 10 is a diagram showing the distribution of luminance values ​​after filtering. [Figure 7] FIG. 10 is a diagram showing an example in which an object to be measured by a sensor has a hole in part. [Figure 8] FIG. 10 is a diagram showing an example in which measurement data that should not be filtered is filtered. [Figure 9] An example is shown in which part of the object to be measured by the sensor has a slope. [Figure 10] FIG. 10 is a diagram showing an example in which measurement data that should not be filtered is filtered. [Figure 11] FIG. 1 illustrates a distance distribution filtering method according to some embodiments. [Figure 12] FIG. 1 illustrates a distance distribution filtering method according to some embodiments. [Figure 13] FIG. 1 illustrates a filtering method based on the distribution of incident angles according to some embodiments. [Figure 14] FIG. 10 is a diagram illustrating an example of calculating the incident angle of a laser. [Figure 15] FIG. 10 is a block diagram showing the configuration of an information processing device according to a second embodiment. [Figure 16] FIG. 10 is a block diagram showing the configuration of an information processing system according to a third embodiment. [Figure 17] FIG. 1 is a block diagram illustrating an example of a hardware configuration of an information processing device according to some embodiments. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. In each drawing, the same or corresponding elements are designated by the same reference numerals, and for clarity of explanation, duplicate explanations will be omitted as necessary.

[0014] <Embodiment 1> FIG. 1 is a block diagram showing the configuration of an information processing apparatus according to the first embodiment. The information processing device 100 is realized by a computer having a processor, a memory, and the like. The information processing device 100 can be used to acquire point cloud data from a sensor that measures a measurement object and analyze the point cloud data. The sensor can be a radar sensor (e.g., LiDAR) that measures the distance to a measurement point. Specifically, the information processing device 100 includes a first acquisition unit 110, a second acquisition unit 120, a division unit 130, a distribution calculation unit 140, and a removal unit 150.

[0015] The first acquisition unit 110 acquires first point cloud data indicating the distance between the sensor and the object to be measured at a first measurement point along the route, and the second acquisition unit 120 acquires second point cloud data indicating the distance between the sensor and the object to be measured at a second measurement point along the route that is different from the first measurement point.

[0016] Typically, a sensor is mounted on a moving object (e.g., a vehicle) and moves to a first measurement point and a second measurement point, irradiates the object to be measured with a laser, and receives the reflected light to obtain point cloud data indicating the distance between the sensor and the object to be measured. However, the sensor of the present invention is not limited to this. For example, a person may carry the sensor and measure while moving. Also, the sensor at the first measurement point and the sensor at the second measurement point may be different. For example, if a tunnel is long, measurements can be made using multiple sensors (e.g., at predetermined distances).

[0017] The dividing unit 130 divides the point cloud data, including the first point cloud data and the second point cloud data, into one or more segments. A segment is a small area into which point cloud data obtained from multiple measurement points is divided. A segment includes one or more points from the point cloud data. If the measurement object is a three-dimensional object, the segment may be a three-dimensional shape such as a sphere, a rectangular parallelepiped, or a cube. If (at least) a portion of the measurement object is planar, the segment may be an assumed plane such as a circle, rectangle, or square, and points near this plane may be extracted. The size of the segment can be set arbitrarily taking into account the size of a portion of interest in the measurement object (e.g., a crack, dent, hole, etc.). The sizes of the segments may be the same or different from each other.

[0018] The distribution calculation unit 140 calculates the distribution of point cloud data within a segment. The distribution calculation unit 140 can calculate, for example, the distribution of distances from a sensor to a measurement object, or the distribution of laser incident angles from the sensor to the measurement object. The distribution calculation unit 140 can calculate the distribution of distances or incident angles for all segments. Furthermore, the distribution calculation unit 140 may calculate the distribution of point cloud data within one or more segments, rather than for all segments.

[0019] The removal unit 150 removes outliers from the point cloud data within the segment based on the calculated distribution. For example, the removal unit 150 can exclude values ​​that are different from the mean value of the distribution by a predetermined threshold (for example, standard deviations σ, 2σ, 3σ, etc.) as outliers.

[0020] FIG. 2 is a flowchart showing the information processing method according to the first embodiment. The information processing method includes acquiring first point cloud data indicating a distance between a sensor and a measurement object at a first measurement point along a route (step S101). Acquiring second point cloud data indicating a distance between a sensor and the measurement object at a second measurement point along the route that is different from the first measurement point (step S102). Dividing point cloud data including the first point cloud data and the second point cloud data into one or more segments (step S103). Calculating a distribution of the point cloud data within a segment (step S104). Removing outliers from the point cloud data within the segment based on the distribution (step S105).

[0021] The information processing device and method according to the first embodiment described above can remove unnecessary data from point cloud data measured by moving a sensor without removing data that is actually necessary.

[0022] <Embodiment 2> FIG. 3 is a diagram illustrating an example of detecting an abnormality in a tunnel using a sensor according to the second embodiment. In FIG. 3, a moving object 2 (e.g., a vehicle) equipped with a LiDAR sensor 5 travels through the tunnel and irradiates a laser beam from the sensor onto the inner wall of the tunnel (a measurement object 6). The sensor then receives the reflected laser beam from the measurement object. The sensor may irradiate the laser beam almost continuously at a very short interval. Alternatively, the moving object 2 may move to any position and irradiate the laser beam at each position. The sensor measures the distance to each measurement point based on the time between irradiating multiple laser beams in the circumferential direction and receiving the reflected beam. Alternatively, the sensor may acquire the brightness value of the reflected beam, thereby estimating the material of the measurement object (e.g., metal, fabric, etc.). In this manner, the sensor can acquire point cloud data including measurement data at multiple measurement points measured by the sensor.

[0023] In this example, small cracks in the interior walls of a tunnel are detected based on the acquired point cloud data. Generally, cracks are so small that they cannot be identified based on differences in distance measured by the sensor, i.e., the shape of the crack. Meanwhile, the brightness value, which indicates the intensity of light reflected from a crack, generally tends to be weak. Therefore, cracks are detected based on the brightness value of the reflected light. However, when point cloud data is acquired by scanning the interior walls of a tunnel while moving the sensor, there is a problem in that the brightness values ​​contained in the point cloud data are unstable. This makes it difficult to detect cracks.

[0024] There are two main reasons why the luminance value is unstable. Figure 4 illustrates the first reason why the luminance value is unstable. The first reason is that when the sensor 5 is moved, lasers 8a and 8b are irradiated onto the same region P of the object 6 from different measurement points A and B. The sensor 5 receives reflected light from different distances (trajectories) through the same region P, resulting in data acquisition of points with different luminance values. In other words, the reflected luminance value acquired by the sensor can be affected by the distance between the sensor 5 and the object 6. The greater the distance between the sensor and the object 6, the smaller the reflected luminance value. Furthermore, as shown in Figure 4, the sensor irradiates a beam in a 360° radius. Therefore, when measurement is performed while the sensor is moving, lasers from different positions on the sensor's movement path irradiate the same region of the object, resulting in data acquisition of different luminance values. In other words, the sensor receives reflected light from different distances from the same region of the object, resulting in data acquisition of different luminance values. For this reason, if the inner wall of a tunnel is scanned while the sensor is moving to acquire point cloud data, it is thought that the brightness values ​​contained in the point cloud data will not be stable.

[0025] The second reason is as follows: When the sensor is moved, lasers 8a and 8b are irradiated onto the same region P of the object 6 from different measurement points A and B. The sensor then receives reflected light at different angles of incidence from the same region P, resulting in data with different luminance values. That is, the reflected luminance values ​​acquired by the sensor can be affected by the incident angle of the laser from the sensor to the object. Generally, the smaller the incident angle of the laser from the sensor to the object, the smaller the reflected luminance value. Furthermore, the sensor irradiates a beam 360° around the sensor. Therefore, when measurements are performed while the sensor is moving, lasers from different positions on the sensor's movement path irradiate the same region of the object, resulting in data with different luminance values. That is, the sensor receives reflected light at different angles of incidence from the same region of the object, resulting in different luminance values. Therefore, if point cloud data is acquired by scanning the interior wall of a tunnel while the sensor is moving, the luminance values ​​included in the point cloud data are likely to be unstable.

[0026] From the above, one possible method for stabilizing brightness values ​​is to remove reflected light whose distance is greater than or equal to a threshold length or whose incident angle is within a threshold angle from point cloud data (also known as filtering using a fixed threshold). The effect of filtering using a fixed threshold will be explained with reference to Figures 5 and 6. Figure 5 shows the distribution of brightness values ​​before filtering using a fixed length threshold (e.g., a distance of 1 m). Figure 6 shows the distribution of brightness values ​​after filtering using a fixed length threshold (e.g., a distance of 1 m). In both Figures 5 and 6, the horizontal axis represents the reflected brightness value, and the vertical axis represents the number of points with the corresponding reflected brightness value. As shown in Figure 6, the distribution of brightness values ​​is more concentrated in the center than the distribution in Figure 5, and the spread of the distribution (variation in brightness values) is suppressed.

[0027] However, when filtering using a fixed threshold is used, the following problem may arise. FIG. 7 shows an example in which a part of an object to be measured by a sensor has a hole (6b in FIG. 7). While the sensor 5 moves near the objects 6a to 6c, the object is irradiated with a laser at various arbitrary points and measured. In this case, if filtering is performed for the objects 6a and 6c using a fixed threshold distance, data with distribution spread suppressed to some extent can be obtained, as described above. On the other hand, the object 6b is farther away from the sensor movement path than the objects 6a and 6c. In this case, if filtering is performed using a fixed threshold distance, measurement data related to the hole 6b in the object may be removed, as shown in FIG. 8. In this way, measurement data that should not be filtered may also be filtered.

[0028] Similarly, FIG. 9 shows an example in which a part of an object to be measured by a sensor has a slope. As the sensor moves near the object, a laser is irradiated onto the object at various arbitrary points, and measurements are taken. The object includes object 6d, which is positioned approximately parallel to the sensor's movement path, and object 6e, which has a slope that gradually moves away from the sensor's movement path. In this case, filtering using a fixed threshold incident angle may result in the removal of measurement data related to the slope of object 6e, as shown in FIG. 10. In this way, measurement data that should not have been filtered may also be filtered.

[0029] Therefore, in this embodiment, point cloud data obtained from multiple measurement points is divided into small segments, and the distribution of distances from measurement points to each segment is calculated and outliers are removed. This distribution-based filtering method will be described below with reference to FIGS. 11 and 12.

[0030] The LiDAR is moved from the measurement start point along a route, and at each measurement point (measurement points A, B, and C in FIG. 11), the LiDAR irradiates the object to be measured with a laser, and then the LiDAR receives the reflected light.

[0031] For measurement point A, point cloud data with measurement point A as the origin and coordinate data for measurement point A with the measurement start point as the origin are obtained. For measurement point B, point cloud data with measurement point B as the origin and coordinate data for measurement point B with the measurement start point as the origin are obtained. For measurement point C, point cloud data with measurement point C as the origin and coordinate data for measurement point C with the measurement start point as the origin are obtained. By using the coordinate data for each of measurement points A to C with the measurement start point as the origin, the coordinate data of the measurement data with measurement points A to C as the origin (referred to as coordinate data A' to C') can be converted into coordinate data with the measurement start point as the origin (referred to as coordinate data A' to C'). By combining the coordinate data A' to C' obtained in this way, point cloud data with the measurement start point as the origin is obtained, as shown in the lower part of Figure 11.

[0032] The distance from the measurement point is calculated for each point in the point cloud data. In this case, the point cloud data can contain information that identifies the measurement point from which each point was obtained using the laser. For example, the point cloud data contains the acquisition time, and by comparing the predetermined route and the time, it is possible to identify the measurement position (measurement points A, B, and C in Figure 11) from which each point was obtained.

[0033] Next, the point cloud data for the measurement object is divided into small segments as shown in FIG. 12. If the point cloud data is three-dimensional, these segments may be spheres, rectangular parallelepipeds, or cubes (sometimes called voxels). The sizes of the segments may be the same or different. In the example shown in FIG. 12, the point cloud data is divided into a number of spheres as segments. The spheres are all the same size (i.e., spheres with a radius r). For example, a specific point of interest can be determined in the point cloud data, and a sphere with a radius r centered around that point can be used as one segment. The point of interest can be set arbitrarily from among the points in the point cloud data. In another example, a specific point of interest can be determined in the point cloud data, and a predetermined number of points (e.g., 30) surrounding the point of interest can be used as one segment.

[0034] The points of interest may be all points in the point cloud data, or one point randomly extracted from the point cloud data and points a predetermined distance away from the extracted point may be extracted sequentially.

[0035] In the example of FIG. 12, the point cloud data is divided so that all points are included in each segment, but it may be divided so that one or more unnecessary points are not included in any segment.

[0036] For each segment, the distribution of distances from the measurement point to other measurement points within the object is calculated, and outliers are removed. The graph at the bottom of Figure 12 shows the distance distribution for a single segment. In this graph, the horizontal axis represents distance, and the vertical axis represents the number of points. From the distribution obtained in this way, the mean value and variance are calculated, and measurement data that exceeds a predetermined threshold value from the mean value (for example, the positive square root of the variance (an integer multiple of the standard deviation σ, i.e., σ, 2σ, 3σ, etc.) can be removed from the obtained point cloud data.

[0037] By calculating the distribution for each segment and eliminating outliers, optimal point cloud data with unnecessary measurement data removed can be obtained, even when the sensor is moved. By eliminating long-distance and short-distance data from the point cloud data, data with reduced variability, as shown in Figure 6, can be obtained. Furthermore, in the example of Figure 7, for example, the inner walls 6a and 6b of the object to be measured and the hole in the inner wall 6b are each divided into separate segments, and the distribution of distances between the measurement point and each point is calculated for each segment. This prevents the measurement data related to the hole in the inner wall 6b from being uniformly removed, as occurs with filtering using a fixed threshold. In this way, the size of the segments can be set taking into account possible points of interest (e.g., dents, holes, etc.) within the object to be measured. Furthermore, cracks in the tunnel's inner walls can be detected from the brightness values ​​based on the optimal point cloud data with unnecessary measurement data removed.

[0038] 13, distributions are taken for all segments and outliers are excluded, but it is not necessary to take distributions for all segments. Also, in the above example, measurement data is obtained at three measurement points, but measurement data may be obtained at four or more measurement points.

[0039] Next, an example of obtaining the distribution of incident angles will be described with reference to FIGS. As with the example of calculating the distance distribution, by combining the measurement data acquired at each measurement point, point cloud data with the measurement start point as the origin is obtained, as shown in Figure 13. Next, the laser incident angle from the measurement point is calculated for each point. Here, a method for calculating the laser incident angle will be explained with reference to Figure 14.

[0040] FIG. 14 is a diagram illustrating an example of calculating the incident angle of the laser. The point (focus point) for which the angle of incidence is to be calculated is determined from the point cloud data. The laser incident plane is estimated from the point cloud data contained in a sphere of radius r from the focus point. If the estimated surface is curved, the plane where the curved surface and the focus point are tangent can be estimated as the incident plane. Finally, the angle between the estimated surface and the vector connecting the measurement point and the focus point is calculated. In this way, the angle of incidence can be calculated for a large number of points.

[0041] Returning to FIG. 13, we will continue the explanation of obtaining the distribution of incident angles. For each point, the incident angle from the measurement point is calculated. In this case, the point cloud data can contain information that identifies which measurement point each point was obtained from. For example, the point cloud data includes the time of acquisition, and by checking the predetermined route and time, it is possible to recognize which measurement position (measurement points A, B, and C in FIG. 11) each point was obtained from.

[0042] Next, as described above, the point cloud data for the measurement object is divided into small segments. For each segment, the distribution of incident angles from one measurement point to another is calculated, and outliers are removed. As described above, from the distribution obtained in this manner, data exceeding a predetermined threshold (e.g., standard deviation σ, 2σ, 3σ, etc.) from the average value can be removed. By calculating the distribution for each segment and removing outliers, optimal point cloud data from which unnecessary data has been removed can be obtained, even when the sensor is moved. For example, by removing data with significantly large and small incident angles, point cloud data with reduced variation, such as that shown in FIG. 6, can be obtained. Furthermore, for example, in the example of FIG. 9, the inner wall 6d and the slope 6e are each divided into separate segments, and the distribution of incident angles between the measurement point and each point is calculated for each segment. This prevents the measurement data for the sloped inner wall 6e from being uniformly removed, as occurs with filtering using a fixed threshold.

[0043] FIG. 15 is a block diagram showing the configuration of an information processing apparatus 100 according to the second embodiment. The information processing device 100 according to this embodiment can be realized by a computer that analyzes point cloud data acquired by a sensor mounted on a mobile object. The information processing device 100 is usually a computer located at a different location from the mobile object, but may also be a computer mounted on the mobile object. In the information processing device 100, the same components as those of the information processing device 100 of the first embodiment are designated by the same reference numerals, and descriptions thereof will be omitted as appropriate.

[0044] The distribution calculation unit 140 of the information processing device 100 further includes a distance distribution calculation unit 141, a plane estimation unit 142, and an incident angle distribution calculation unit 143. As described with reference to Figs. 11 and 12, the distance distribution calculation unit 141 calculates the distribution of distances for point cloud data within each segment. As described with reference to Fig. 14, the plane estimation unit 142 estimates a plane from distance data for multiple points in the vicinity of a point of interest. As described with reference to Fig. 13, the incident angle distribution calculation unit 143 calculates the distribution of incident angles for point cloud data within each segment.

[0045] Furthermore, the removal unit 150 according to this embodiment removes outliers from the point cloud data within the segment based on the distribution of distances. Additionally or alternatively, the removal unit 150 removes outliers from the point cloud data within the segment based on the distribution of incident angles.

[0046] The information processing device and method according to the second embodiment described above can remove unnecessary data from point cloud data measured by moving a sensor without removing essential data. Furthermore, by removing unnecessary data based on the obtained distance distribution and incident angle distribution, more appropriate point cloud data can be obtained.

[0047] <Embodiment 3> FIG. 16 is a block diagram illustrating a configuration of an information processing system according to the third embodiment. The information processing system (also called a measurement system) moves the sensor 5 along a path, irradiates a laser from the sensor 5 onto the object to be measured 6, and measures the distance from the sensor 5 to the object to be measured 6 by the sensor 5 receiving the reflected light.

[0048] The information processing system includes the information processing device 100 described above and a mobile sensor that measures a measurement target. The information processing device 100 may be mounted on a mobile object, for example, and configured integrally with the sensor 5. In this case, the information processing device 100 can remove unnecessary data from the point cloud data obtained by the sensor. Furthermore, the information processing device 100 has a wireless communication unit and can transmit the point cloud data from which unnecessary data has been removed to an external device via a wireless communication network. This can reduce communication traffic.

[0049] Alternatively, the information processing device 100 may be configured as a separate entity from the mobile object equipped with the sensor 5. In this case, the information processing device 100 may have a communication unit, receive point cloud data via a wired or wireless communication network, and store a program for executing the above-described information processing method.

[0050] FIG. 17 is a block diagram showing an example of the hardware configuration of the information processing device 100. Referring to FIG. 17, the information processing device 100 includes a network interface 1201, a processor 1202, and a memory 1203. The network interface 1201 is used to communicate with other network node devices that constitute a communication system. The network interface 1201 may be used to perform wireless communication. For example, the network interface 1201 may be used to perform wireless LAN communication defined in the IEEE 802.11 series or mobile communication defined in the 3GPP (3rd Generation Partnership Project). Alternatively, the network interface 1201 may include, for example, a network interface card (NIC) conforming to the IEEE 802.3 series.

[0051] The processor 1202 reads and executes software (computer programs) from the memory 1203 to perform the processing of the information processing device 100 described using flowcharts or sequences in the above-described embodiments. The processor 1202 may be, for example, a microprocessor, an MPU (Micro Processing Unit), or a CPU (Central Processing Unit). The processor 1202 may include multiple processors.

[0052] The memory 1203 is configured by a combination of volatile memory and non-volatile memory. The memory 1203 may include storage located remotely from the processor 1202. In this case, the processor 1202 may access the memory 1203 via an I / O interface (not shown).

[0053] 17, the memory 1203 is used to store a group of software modules. The processor 1202 reads and executes these software modules from the memory 1203, thereby performing the processing of the information processing device 100 described in the above embodiment.

[0054] As described with reference to FIG. 17, each of the processors included in the information processing device 100 or the like executes one or more programs including a group of instructions for causing a computer to perform the algorithm described with reference to the drawings.

[0055] <Other embodiments> Although the above-described embodiment has been described as a hardware configuration, the present disclosure is not limited to this. Any processing in the present disclosure can also be realized by causing a CPU to execute a computer program.

[0056] In the above example, the program can be stored and supplied to a computer using various types of non-transitory computer-readable media. Non-transitory computer-readable media include various types of tangible storage media. Examples of non-transitory computer-readable media include magnetic storage media (e.g., flexible disks, magnetic tapes, hard disk drives), magneto-optical storage media (e.g., magneto-optical disks), CD-ROMs (Read Only Memory), CD-Rs, CD-R / Ws, DVDs (Digital Versatile Discs), and semiconductor memories (e.g., mask ROMs, PROMs (Programmable ROMs), EPROMs (Erasable PROMs), flash ROMs, and RAMs (Random Access Memory)). The program may also be supplied to a computer by various types of transitory computer-readable media. Examples of transitory computer-readable media include electrical signals, optical signals, and electromagnetic waves. The transitory computer-readable media can be supplied to a computer via wired communication paths such as electrical wires and optical fibers, or via wireless communication paths.

[0057] The present disclosure is not limited to the above-described embodiments, and may be modified as appropriate without departing from the spirit and scope of the present disclosure. In addition, the present disclosure may be implemented by appropriately combining the respective embodiments.

[0058] Some or all of the above embodiments can be described as, but are not limited to, the following supplementary notes. (Appendix 1) a first acquisition unit that acquires first point cloud data indicating a distance between the sensor and the measurement object at a first measurement point along the route; a second acquisition unit that acquires second point cloud data indicating a distance between a sensor and the measurement object at a second measurement point along the route that is different from the first measurement point; a dividing unit that divides point cloud data including the first point cloud data and the second point cloud data into one or more segments; a distribution calculation unit that calculates a distribution of the point cloud data within the segment; a removal unit that removes outliers from the point cloud data within the segment based on the distribution; An information processing device comprising: (Appendix 2) 2. The information processing device according to claim 1, wherein the distribution calculation unit further comprises a distance distribution calculation unit that calculates a distribution of distances for the point cloud data within the segment. (Appendix 3) the distribution calculation unit includes a surface estimation unit that estimates a surface from the point cloud data within the segment; 3. The information processing device according to claim 1, further comprising: an incident angle distribution calculation unit that calculates a distribution of incident angles onto the estimated surface for the point cloud data within the segment. (Appendix 4) 2. The information processing device according to claim 1, wherein the segment is a sphere of a predetermined size or a voxel of a predetermined size. (Appendix 5) 5. The information processing device according to any one of claims 1 to 4, wherein the point cloud data includes luminance values. (Appendix 6) a sensor that irradiates a laser beam from the sensor onto a measurement object while moving the sensor along a path, and receives reflected light from the sensor to acquire point cloud data that measures the distance from the sensor to the measurement object; a first acquisition unit that acquires first point cloud data indicating a distance between the sensor and the measurement object at a first measurement point along the route; a second acquisition unit that acquires second point cloud data indicating a distance between the sensor and the measurement object at a second measurement point along the route that is different from the first measurement point; a dividing unit that divides point cloud data including the first point cloud data and the second point cloud data into a plurality of segments; a distribution calculation unit that calculates a distribution of the point cloud data within the segment; a removal unit that removes outliers from the point cloud data within the segment based on the distribution; An information processing system comprising: (Appendix 7) 7. The information processing system according to claim 6, further comprising: a mobile body that is equipped with the sensor. (Appendix 8) The information processing system described in Appendix 6 or 7, wherein the sensor, the first acquisition unit, and the second acquisition unit are connected to each other via a wireless communication network so that point cloud data can be communicated. (Appendix 9) acquiring first point cloud data indicating a distance between the sensor and the object to be measured at a first measurement point along the path; acquiring second point cloud data indicating a distance between a sensor and the measurement object at a second measurement point along the route that is different from the first measurement point; Dividing point cloud data including the first point cloud data and the second point cloud data into one or more segments; Calculating a distribution of the point cloud data within the segment; removing outliers from the point cloud data within the segment based on the distribution; Information processing methods. (Appendix 10) acquiring first point cloud data indicating a distance between the sensor and the object at a first measurement point along the route; acquiring second point cloud data indicating a distance between a sensor and the object to be measured at a second measurement point along the route that is different from the first measurement point; Dividing point cloud data including the first point cloud data and the second point cloud data into one or more segments; A process of calculating a distribution of the point cloud data within the segment; and a process of removing outliers from the point cloud data within the segment based on the distribution.

[0059] Although the present invention has been described above with reference to the embodiments (and examples), the present invention is not limited to the above-described embodiments (and examples). Various modifications that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention. [Explanation of symbols]

[0060] 1. Information Processing Systems 2. Mobile 5 sensors 6 Measurement object 8 Laser 100 Information processing device 110 First acquisition part 120 Second acquisition part 130 Split section 140 Distribution calculation part 141 Distance distribution calculation section 142 Surface estimation part 143 Incident angle distribution calculation section 150 Removal section

Claims

1. a first acquisition unit that acquires first point cloud data indicating a distance between the sensor and the measurement object at a first measurement point along the route; a second acquisition unit that acquires second point cloud data indicating a distance between a sensor and the measurement object at a second measurement point along the route that is different from the first measurement point; a division unit that divides point cloud data including the first point cloud data and the second point cloud data into a plurality of segments; a distribution calculation unit that calculates a distribution of the point cloud data for each segment, the first point cloud data based on a distance from the first measurement point or an incident angle of a laser, and the second point cloud data based on a distance from the second measurement point or an incident angle of a laser; a removal unit that removes outliers from the point cloud data for each of the segments based on the distribution and a threshold value calculated from the distribution; An information processing device comprising:

2. The information processing apparatus according to claim 1 , wherein the distribution calculation unit further comprises a distance distribution calculation unit that calculates a distribution of distances from the first measurement point and the second measurement point for the point cloud data within the segment.

3. The distribution calculation unit includes a surface estimation unit that estimates a surface from the point cloud data in the segment; The information processing device according to claim 1 , further comprising an incidence angle distribution calculation unit that calculates a distribution of incidence angles from the first measurement point and the second measurement point to the estimated surface for the point cloud data within the segment.

4. The information processing apparatus according to claim 1 , wherein the segment is a sphere of a predetermined size or a voxel of a predetermined size.

5. The information processing device according to claim 1 , wherein the point cloud data includes a luminance value.

6. a sensor that irradiates a laser from the sensor to a measurement object while moving the sensor along a path and receives reflected light to acquire point cloud data that measures a distance from the sensor to the measurement object; a first acquisition unit that acquires first point cloud data indicating a distance between the sensor and the measurement object at a first measurement point along the route; a second acquisition unit that acquires second point cloud data indicating a distance between the sensor and the measurement object at a second measurement point along the route that is different from the first measurement point; a division unit that divides point cloud data including the first point cloud data and the second point cloud data into a plurality of segments; a distribution calculation unit that calculates a distribution of the point cloud data within the segment based on indexes from the first measurement point and the second measurement point; a removal unit that removes outliers from the point cloud data within the segment based on the distribution; An information processing system comprising:

7. The information processing system according to claim 6 , further comprising a moving object on which the sensor is mounted.

8. The information processing system according to claim 6 , wherein the sensor, the first acquisition unit, and the second acquisition unit are connected to each other via a wireless communication network so as to be able to communicate the point cloud data.

9. acquiring first point cloud data indicative of a distance between the sensor and the object at a first measurement point along the path; acquiring second point cloud data indicating a distance between a sensor and the measurement object at a second measurement point along the route that is different from the first measurement point; Dividing point cloud data including the first point cloud data and the second point cloud data into a plurality of segments; A distribution of the point cloud data for each segment is calculated based on a distance from the first measurement point or an incident angle of a laser for the first point cloud data, and based on a distance from the second measurement point or an incident angle of a laser for the second point cloud data; removing outliers from the point cloud data for each of the segments based on the distribution and a threshold calculated from the distribution; Information processing methods.

10. acquiring first point cloud data indicative of distances between the sensor and the object at first measurement points along the route; acquiring second point cloud data indicating a distance between a sensor and the measurement object at a second measurement point along the route that is different from the first measurement point; Dividing point cloud data including the first point cloud data and the second point cloud data into a plurality of segments; A process of calculating a distribution of the point cloud data for each segment, the first point cloud data being based on a distance from the first measurement point or an incident angle of a laser, and the second point cloud data being based on a distance from the second measurement point or an incident angle of a laser; and removing outliers from the point cloud data for each segment based on the distribution and a threshold value calculated from the distribution.

Citation Information

Patent Citations

  • Information processing device, imaging apparatus, device control system, information processing method and program

    JP2018085059A

  • Distributed data processing system, center server, edge server, mobile terminal, and method

    JP2018092472A

  • Road sign detection device, road sign detection method, program, and road surface detection device

    JP2018173749A

  • Transmission device, point group data collection system, and computer program

    WO2019131075A1