JP2019179495A - Sensor processing system, distance measuring system, mobile object, sensor processing method, and program - Google Patents

Abstract

To provide a sensor processing system, distance measuring system, mobile object, sensor processing method, and program capable of increasing the accuracy of obstacle detection.SOLUTION: A sensor processing system 1 comprises: an acquisition part 101 that acquires a result detected by a distance measuring sensor 20; and an output part 106. The output part 106 outputs a detection result obtained by removing a removal object from a result detected by the distance measuring sensor 20 based on map information including data for the height of the removal object to be removed from the results detected by the distance measuring sensor 20.SELECTED DRAWING: Figure 1

Description

  The present disclosure relates generally to a sensor processing system, a ranging system, a moving body, a sensor processing method, and a program, and more particularly to a sensor processing system, a ranging system, a moving body, a sensor processing method, and a program for detecting an obstacle.

  Conventionally, an autonomous mobile robot that autonomously moves to a destination is known (for example, Patent Document 1).

  The autonomous mobile robot of Patent Document 1 includes sensing means capable of detecting the presence or absence of obstacles present in the surroundings, and storage means for storing a map within a moving range. The autonomous mobile robot of Patent Document 1 moves to the destination indicated on the map stored in the storage means while confirming that there is no obstacle in the moving direction from the detection result of the sensing means.

JP 2006-321001 A

  In order to move an autonomous mobile robot (for example, a moving body such as a vehicle) to a destination while avoiding an obstacle in a system (sensor processing system) that determines the presence or absence of an obstacle based on the detection result of the sensing means, It is important to improve the accuracy of detecting obstacles.

  The present disclosure has been made in view of the above problems, and an object thereof is to provide a sensor processing system, a ranging system, a moving body, a sensor processing method, and a program that can further improve the accuracy of obstacle detection.

  A sensor processing system according to an aspect of the present disclosure includes an acquisition unit that acquires a result detected by a distance measurement sensor, and an output unit. The output unit is a detection result obtained by removing the removal object from the result detected by the distance sensor based on map information including data on the height of the object to be excluded from the result detected by the distance sensor. Is output.

  A ranging system according to an aspect of the present disclosure includes the sensor processing system and the ranging sensor.

  A moving body according to an aspect of the present disclosure includes the distance measuring system and a main body.

  The sensor processing method according to an aspect of the present disclosure includes an acquisition step of acquiring a result detected by the distance measuring sensor, and an output step. The output step outputs a detection result obtained by removing an object to be removed from the result detected by the distance measuring sensor based on map information including height data of an object that is a detection target of the distance measuring sensor.

  A program according to an aspect of the present disclosure is a program for causing a computer system to execute an acquisition step of acquiring a result detected by a distance measuring sensor and an output step. The output step outputs a detection result obtained by removing an object to be removed from the result detected by the distance measuring sensor based on map information including height data of an object that is a detection target of the distance measuring sensor.

  According to the present disclosure, it is possible to further increase the accuracy of obstacle detection.

FIG. 1 is a block diagram illustrating a configuration of a ranging system according to an embodiment of the present disclosure. 2A and 2B are diagrams for explaining an application example of the distance measuring sensor of the distance measuring system. FIG. 3 is a diagram for explaining a range regarded as an obstacle in the above-described distance measuring system. FIG. 4 is a flowchart for explaining the operation of the detection device provided in the distance measuring system. FIG. 5 is a diagram for describing processing for comparing feature amounts included in tracking processing performed by the above-described detection device. FIG. 6 is a diagram for describing processing for comparing the positions of the center of gravity included in the tracking processing performed by the above-described detection device. FIG. 7 is a diagram illustrating a process for comparing feature amounts included in the object identification process performed by the above-described detection apparatus. FIG. 8 is a flowchart for explaining the process of excluding the ground included in the object detection process performed by the above-described detection apparatus.

  Embodiments and modifications described below are merely examples of the present disclosure, and the present disclosure is not limited to the embodiments and the modifications. Even if it is except this embodiment and a modification, if it is a range which does not deviate from the technical idea which concerns on this indication, various changes are possible according to a design etc.

(Embodiment)
The sensor processing system, ranging system, moving body, sensor processing method, and program according to this embodiment will be described with reference to FIGS.

(1) Outline The sensor processing system 1 and the distance measuring system 5 of the present embodiment automatically recognize the surrounding environment by a distance measuring sensor (for example, LIDAR (Light Detection and Ranging)) and GPS (Global Positioning System). It is provided in a moving body (for example, vehicle 6) that travels by driving.

  The sensor processing system 1 is provided in the distance measuring system 5 and detects an obstacle (object) existing around the vehicle 6. The obstacle is, for example, a person or another vehicle. The operation control unit 40 (see FIG. 1) included in the vehicle 6 determines whether to reduce or accelerate the traveling speed, avoid the obstacle, or the like based on the position, shape, size, etc. of the obstacle detected by the sensor processing system 1. To control the automatic driving.

(2) Configuration Hereinafter, the configuration of the distance measuring system 5 will be described.

  The ranging system 5 is provided in the vehicle 6 and includes a detection device 10 as the sensor processing system 1, a ranging sensor 20, and a storage unit 30.

  The detection apparatus 10 includes a microcomputer having a processor and a memory, for example. Then, the microcomputer functions as the detection device 10 when the processor executes the program stored in the memory. Here, the program executed by the processor is recorded in advance in the memory of the microcomputer, but may be provided by being recorded on a recording medium such as a memory card, or may be provided through an electric communication line such as the Internet. .

  The distance measuring sensor 20 is, for example, a LIDAR, is provided in the main body 61 of the vehicle 6, and outputs a laser R1 every 100 msec. The distance measuring sensor 20 outputs the laser R1 while rotating the laser 6 about the vertical direction of the vehicle 6 (see FIG. 2A). Further, the distance measuring sensor 20 outputs the laser R1 while moving the laser R1 around the left and right directions as an axis in front and rear of the vehicle (see FIG. 2B). The distance measuring sensor 20 receives the reflected wave reflected by the object with respect to the output laser R1. Thereby, since the ranging sensor 20 receives the reflected wave every time the laser R1 is irradiated in the horizontal direction and the vertical direction, the ranging system 5 can acquire a plurality of reflection positions as a point group.

  The storage unit 30 includes a device selected from a ROM (Read Only Memory), a RAM (Random Access Memory), an EEPROM (Electrically Erasable Programmable Read Only Memory), or the like.

  The storage unit 30 stores map information. The map information is used for searching for a route when the vehicle 6 travels and for checking the current position. The map information includes height information of an object (for example, the ground, a curb, etc.) that is not regarded as an obstacle when the sensor processing system 1 (detection device 10) detects the obstacle. For example, the map information includes, for each specific point, the height of the ground (road) 1 m away from the point (difference from a predetermined reference point), the height of the ground 10 m away, and the height of the ground 20 m away Etc. As a result, when the vehicle 6 exists at a specific point, the vehicle 6 knows the height of the ground (road) 1 m ahead of the point, the height of the ground 10 m ahead, and the height of the ground 20 m ahead. it can.

  The storage unit 30 stores the detection result of the detection device 10 as a history. Furthermore, the memory | storage part 30 memorize | stores model data, such as a person and a vehicle, beforehand.

  As illustrated in FIG. 1, the detection apparatus 10 includes an acquisition unit 101, an object detection unit 102, a feature amount extraction unit 103, a tracking processing unit 104, an object identification unit 105, and an output unit 106.

  The acquisition unit 101 acquires the detection result of the distance measuring sensor 20. Specifically, the acquisition unit 101 acquires the detection result of the distance measuring sensor 20 as a point group in the space around the vehicle 6 (horizontal direction, vertical direction, front-rear direction). For example, the acquisition unit 101 uses, as point group data, three-dimensional coordinate points of each point, that is, horizontal coordinate values, vertical coordinate values based on the distance measuring sensor 20, and front and rear from the distance measuring sensor 20. A coordinate value in the direction (a distance from the distance measuring sensor 20) is acquired.

  The acquisition unit 101 acquires map information including the current position of the vehicle 6 based on the measurement result of the vehicle 6 by GPS from the storage unit 30. For example, the acquisition unit 101 acquires map information including an area within a radius of 50 m with the current position of the vehicle 6 as the center.

  The object detection unit 102 excludes an object (exclusion target) that is not regarded as an obstacle from the detection result acquired by the acquisition unit 101. Specifically, first, the object detection unit 102 divides the point cloud data into a plurality of grids in the vertical direction. That is, data of a point group within a predetermined angle range with the vertical direction as the rotation axis is defined as one grid. The object detection unit 102 removes data (observation data) of one or more points whose heights are equal to or less than the first value α1 with respect to a plurality of grids, with the height α to be removed as the first value α1. To do. Thereafter, the object detection unit 102 changes the height α to be removed according to the presence or absence of an object in the vertical direction, and removes one or more observation data having a height α or less from a plurality of grids.

  The object detection unit 102 excludes one or more observation data having a height β or more with respect to a plurality of grids. Thereby, the object detection part 102 can extract only the object of the range whose height is larger than (alpha) and smaller than (beta) about several grids (refer FIG. 3). That is, the object detection unit 102 can regard an object in a range where the height is larger than α and smaller than β as an obstacle.

  The object detection unit 102 extracts a grid belonging to an area in which the vehicle 6 can travel from a grid having a height greater than α and smaller than β.

  The object detection unit 102 classifies the grids belonging to the travelable area into one or more groups (clusters). Specifically, the object detection unit 102 sets one or more grids having the same distance from the distance measuring sensor 20 and continuous in the horizontal direction as one cluster. Here, “the distance is the same” includes not only completely the same but also a case where the difference is within a predetermined range. In the present embodiment, each extracted cluster is regarded as a detected object.

  The feature amount extraction unit 103 extracts feature amounts for one or more clusters from which the object detection unit 102 has been extracted. Specifically, the feature amount extraction unit 103 extracts a plurality of feature points such as corners and edges of the object represented by the plurality of points based on a plurality of observation data included in the cluster, and the plurality of features. In accordance with the arrangement of the points, the feature points are digitized to generate a histogram. For example, the feature points are digitized according to whether the feature points are arranged in a straight line, a curve, or distributed, and a histogram is generated.

  The tracking processing unit 104 determines the presence / absence of the detected movement of the object, and obtains the speed when there is movement. The tracking processing unit 104 includes the feature amounts of all the clusters included in the result (first result) detected by the distance measuring sensor 20 at the present time, and all the values included in the result (second result) detected by the distance measuring sensor 20. Compare with cluster features. The tracking processing unit 104 compares the centroid positions of all the clusters included in the first result with the centroid positions of all the clusters included in the second result. The tracking processing unit 104 performs matching processing between all the clusters included in the first result and all the clusters included in the second result, using the comparison result of the feature amount and the comparison result of the centroid position. Here, the comparison of the center of gravity position is a comparison of the distance from the distance measuring sensor 20 to the center of gravity of the object.

  When the tracking processing unit 104 determines that the first result and the second result include the same type of cluster (cluster corresponding to the same obstacle) by the matching process, the tracking processing unit 104 moves and moves the cluster. And the moving speed is obtained. When there is a cluster that is not included in the second result but included in the first result, the tracking processing unit 104 determines that the cluster is a newly detected cluster.

  The object identifying unit 105 identifies what obstacle is associated with each cluster included in the result detected by the distance measuring sensor 20. The object identification unit 105 obtains a feature amount of each model data stored in the storage unit 30. As described above, with respect to the feature amount of each model data, a plurality of feature points such as corners and edges of the object represented by the model data are extracted, and the feature points are digitized according to the arrangement of the plurality of feature points. The object identification unit 105 compares the histogram of the feature amount of each model data with the histogram created by the feature amount extraction unit 103. Based on the comparison result, the object identification unit 105 identifies the object represented by the comparison target cluster as the object represented by the model data having the highest similarity with the histogram created by the feature amount extraction unit 103.

  As a result of the object detection unit 102 excluding the object to be excluded from the detection result of the distance measuring sensor 20, the output unit 106 indicates whether or not each cluster has been obtained by the tracking processing unit 104, the movement direction when there is a movement, Outputs the travel distance and travel speed. Further, the output unit 106 outputs information (object name) of the object identified by the object identifying unit 105 to the operation control unit 40. The result of excluding the object to be excluded from the detection result of the distance measuring sensor 20 by the object detection unit 102 is, for example, the position information of each cluster detected by the object detection unit 102.

  The operation control unit 40 uses the position information of each cluster, the presence / absence of movement of each cluster, the movement direction, movement distance and movement speed, if any, and the object name of each cluster. Control of automatic travel is performed by judging deceleration or acceleration, avoidance of obstacles, or the like.

(3) Operation (3.1) Overall Operation Here, the operation of the ranging system 5 will be described with reference to FIG. Note that the detection device 10 is premised on acquiring map information of a region including the position of the vehicle 6 from the storage unit 30 in advance.

  The acquisition unit 101 acquires the detection result of the distance measuring sensor 20 as a point group in a space (three-dimensional) around the vehicle 6 (step S1).

  The object detection unit 102 performs an object detection process (step S2). The object detection unit 102 first performs ground removal processing. Details of the ground removal process will be described later. The object detection unit 102 excludes one or more observation data having a height β or more from the plurality of grids subjected to the ground exclusion process. The object detection unit 102 extracts a grid belonging to an area in which the vehicle 6 can travel from a grid having a height greater than α and smaller than β. The object detection unit 102 classifies the grid belonging to the travelable area into one or more clusters.

  The feature amount extraction unit 103 performs feature amount extraction processing (step S3). The feature quantity extraction unit 103 extracts a plurality of feature points such as corners and edges of the object represented by the plurality of points based on a plurality of observation data included in the cluster, and a plurality of feature points are extracted. In accordance with the arrangement of the feature points, for example, depending on whether the feature points are arranged in a straight line, a curve, or dispersed, the feature points are digitized to generate a histogram. Thereby, it is possible to identify whether the object represented by the cluster has a rounded shape or a flat shape without roundness.

  The tracking processing unit 104 performs tracking processing (step S4).

  First, the tracking processing unit 104 detects the feature amounts of one or more clusters included in the result (first result) detected by the distance measurement sensor 20 this time and the result (second result) detected by the distance measurement sensor 20 last time. The feature amounts of one or more clusters included in the result) are compared. Specifically, as shown in FIG. 5, the histogram H1 generated based on the feature amount of the object B1 represented by the cluster included in the first result and the feature amount of the object B11 represented by the cluster included in the second result. Is compared with the histogram H11 generated based on the above, and the degree of coincidence of the feature amounts is obtained.

  Next, the tracking processing unit 104 compares the centroid positions of all the one or more clusters included in the first result with the centroid positions of all the one or more clusters included in the second result. Specifically, the tracking processing unit 104 includes a first distance from the distance measuring sensor 20 to the center of gravity of the cluster included in the first result, and a second distance from the distance measuring sensor 20 to the center of gravity of the cluster included in the second result. Find the distance. The tracking processing unit 104 compares the first distance and the second distance, and sets the comparison result so that the value increases as the difference between the first distance and the second distance decreases. For example, as shown in FIG. 6, it is assumed that the object (cluster) B21 is detected as the previous detection result, and the objects (clusters) B2, B3, and B4 are detected as the current detection result. In this case, the tracking processing unit 104 obtains the center of gravity of each of the objects B21, B2 to B4. The tracking processing unit 104 obtains the distance from the distance measuring sensor 20 to the center of gravity from the observation data of the objects B21, B2 to B4. The tracking processing unit 104 sets a value corresponding to the comparison result between the distance to the object B21 and the distance to the objects B2 to B4 in each of the objects B21 and B2 to B4. A comparison result with each of B2 to B4 is set for the object B21.

  Next, the tracking processing unit 104 performs matching processing between all the clusters included in the first result and all the clusters included in the second result, using the comparison result of the feature amount and the comparison result of the centroid position. . The tracking processing unit 104 performs a matching process using, for example, a mathematical expression “similarity = ω × d1 + (1−ω) × d2”. Here, ω is a weighting coefficient, d1 is a comparison result (matching degree) of feature amounts, and d2 is a comparison result (evaluation value) of centroid positions. Among the clusters included in the first result and the clusters included in the second result, clusters having high similarity are regarded as the same object. If the degree of similarity is equal to or less than a predetermined value, the corresponding cluster is regarded as a newly detected cluster this time or a cluster detected last time but not detected this time.

  The object identification unit 105 performs object identification processing (step S5). The object identification unit 105 performs the following processing on all the clusters detected this time.

  First, the object identification unit 105 obtains a feature amount of each model data stored in the storage unit 30. The object identification unit 105 extracts a plurality of feature points such as corners and edges of the object represented by the model data for the feature amount of each model data, digitizes the feature points according to the arrangement of the plurality of feature points, and displays a histogram. Generate.

  The object identification unit 105 compares each of the histograms of all the clusters generated by the feature amount extraction unit 103 with the histograms of all the model data, and obtains the degree of coincidence with all the model data. The object identification unit 105 calculates a similarity by multiplying each obtained degree of coincidence by a weighting coefficient. For example, the object identification unit 105 compares the histogram H31 of the cluster (object) B31 detected this time with the histogram H32 of the model data (see FIG. 7). The object identification unit 105 calculates the similarity of the cluster (object) B31 detected this time with respect to the object represented by the histogram H32 by multiplying the comparison result (matching degree) between the histogram H31 and the histogram H32 by the weighting coefficient.

  The object identification unit 105 identifies each of the histograms of all the clusters generated by the feature amount extraction unit 103 as an object represented by the model data having the highest similarity among all the model data.

  The output unit 106 performs output processing (step S6). As a result of the object detection unit 102 excluding the object to be excluded from the detection result of the distance measuring sensor 20, the output unit 106 indicates whether or not each cluster has been obtained by the tracking processing unit 104, the movement direction when there is a movement, Outputs the travel distance and travel speed. Further, the output unit 106 outputs information (object name) of the object identified by the object identifying unit 105 to the operation control unit 40.

  The operation control unit 40 controls automatic traveling by determining deceleration or acceleration of the line speed or avoiding an obstacle based on the information output from the output unit 106.

(3.2) Ground Exclusion Processing Next, the ground exclusion processing included in the object detection processing will be described with reference to FIG.

  The object detection unit 102 sets the removal target height α to the first value α1 (step S11). The object detection unit 102 removes one or more observation data whose height is not more than the first value α1 with respect to a plurality of grids (step S12).

  The object detection part 102 performs step S13-step S16, and excludes the observation data regarded as an exclusion object for each of the plurality of grids.

  The object detection unit 102 determines whether or not an object exists in the vertical direction of the grid for the processing target grid (step S13). Specifically, the object detection unit 102 determines whether there is a predetermined number or more of observation data in the vertical direction of the processing target grid. When the object detection unit 102 determines that there is a predetermined number or more of observation data in the vertical direction of the grid to be processed, the object detection unit 102 determines that an object exists in the vertical direction of the grid. When the object detection unit 102 determines that a predetermined number or more of observation data does not exist in the vertical direction of the processing target grid, the object detection unit 102 determines that no object exists in the vertical direction of the grid.

  If the object detection unit 102 determines that an object is present in the vertical direction of the processing target grid (“Yes” in step S13), the object detection unit 102 sets the removal target height α to the second value α2 (step S14). Here, the second value α2 is a value larger than the first value α1.

  If the object detection unit 102 determines that there is no object in the vertical direction of the grid to be processed (“No” in step S13), the object detection unit 102 sets the removal target height α to the third value α3 (step S15). Here, the third value α3 is a value (a value represented by the height information) that is larger than the second value α2 and obtained from the height information of the excluded object included in the map information. That is, a relationship of “first value α1 <second value α2 <third value α3 (= height of the excluded object)” is established.

  The object detection unit 102 excludes one or more observation data that are equal to or less than the value α set in step S14 or step S15 for the grid to be processed (step S16).

  The object detection unit 102 determines whether or not ground exclusion has been executed for all grids, that is, whether or not ground exclusion for all grids has been completed (step S17).

  When the object detection unit 102 determines that ground removal has not been performed for all grids (“No” in step S17), the process returns to step S13, and steps S13 to S16 are performed for the next grid to be processed. Execute.

  If the object detection unit 102 determines that ground removal has been executed for all grids (“Yes” in step S17), the process ends.

  The object detection unit 102 can exclude objects (roads, curbs, etc.) that are not regarded as obstacles when the detection device 10 detects an obstacle by executing the ground exclusion process.

(4) Advantages In this embodiment, the detection device 10 (sensor processing system 1) uses the map information from the detection result (point cloud data) of the distance measuring sensor 20 when the vehicle 6 is traveling. Excludes observation data (roads, curbs, etc.). A sensor processing system that does not exclude observation data on the ground (roads, curbs, etc.) may recognize roads, curbs, etc. as obstacles. For example, when the distance measuring system recognizes a road as an obstacle, it always recognizes that there is an obstacle because the road always exists in the traveling direction of the vehicle. For this reason, there is a possibility of hindering automatic running such as always applying the brake, constantly decelerating, or running meandering to avoid an obstacle. Therefore, by excluding observation data of the ground (roads, curbs, etc.) from the detection results (point cloud data) of the distance measuring sensor 20, the possibility of misidentifying roads, curbs, etc. as obstacles can be reduced. . As a result, automatic traveling can be performed by appropriate control.

  In the present embodiment, the object detection unit 102 changes the value of the height α to be excluded depending on the presence or absence of an object when excluding the ground. For example, if the height α of the exclusion target is a value obtained from the height information of the exclusion target included in the map information regardless of the presence or absence of the object, if the object exists, observation data representing a part of the object May be deleted. If the observation data representing a part of the object is deleted, the feature quantity of the object may be deleted, so that accurate identification cannot be performed when the object is identified. Therefore, when the object exists, the observation data representing a part of the object is obtained by setting the height α of the exclusion target to a value smaller than the value obtained from the height information of the exclusion target included in the map information. Can be less likely to be deleted. Thereby, compared with the case where the observation data representing a part of the object is deleted, accurate identification can be performed when the object is identified.

  In the present embodiment, the histogram generated by the feature quantity extraction unit 103 includes feature quantities having a shape (rounded shape, flat shape). Therefore, the tracking processing unit 104 can determine whether the object included in the first result (current detection result) and the object included in the second result (previous detection result) are the same based on the shape. Thus, the presence or absence of movement of the same object can be determined. Moreover, the shape of the same object does not change suddenly in the first result and the second result. Therefore, the tracking processing unit 104 uses the feature amount of the shape of the object when obtaining the degree of coincidence between the object included in the first result and the object included in the second result, thereby obtaining a more accurate degree of coincidence. be able to.

  Further, the object identification unit 105 compares the feature amount histogram of the model data with a histogram including the feature amount of the shape (rounded shape, flat shape). For example, when the storage unit 30 stores the model data of a truck as a vehicle, it is unlikely that the shape of the track stored as the model data and the shape of the actually detected track are significantly different. That is, there is no possibility that the shape of the object stored as model data and the shape of the actually detected object are significantly different. Therefore, the object identification unit 105 can more accurately identify the object included in the detection result of the distance measuring sensor 20 by using the histogram including the feature amount of the shape (rounded shape, flat shape). .

(Modification)
Below, modifications are listed. Note that the modifications described below can be applied in appropriate combination with the above embodiment.

  In the map information, an area including the distance measuring sensor 20 may be divided into a plurality of sections, and the removal amount in the height direction of the removal target may be set for each section. Thereby, since the removal amount can be changed according to the terrain (the road and the height of the curb) in the vicinity of the vehicle 6, it is possible to more accurately exclude the object to be excluded. As a result, the accuracy of obstacle detection can be further increased.

  In the above embodiment, the output unit 106 may output the processing result of the object detection unit 102, the processing result of the tracking processing unit 104, and the processing result of the object identification unit 105 to a display unit that displays information. In this case, the display unit displays an object detected as an obstacle on a map centered on the vehicle 6 in a shape corresponding to the identification result of the object identification unit 105. For example, when the object detected as an obstacle is a track, the display unit displays the object in the shape of a track, and when the object detected as an obstacle is a person, the display unit displays the object as a person's shape. Is displayed.

  In the said embodiment, although LIDAR was demonstrated as an example as the ranging sensor 20, the ranging sensor 20 is not limited to this. The distance measuring sensor 20 may be a stereo camera or a sonar, or may be a combination of a LIDAR, a stereo camera and a sonar.

  In the above embodiment, the tracking processing unit 104 calculates the similarity between the object included in the first result (current detection result) and the object included in the second result (previous detection result). Although the configuration uses the shape and the distance from the distance measuring sensor 20, it is not limited to this configuration. The tracking processing unit 104 calculates the similarity between the object included in the first result and the object included in the second result, in addition to the shape of the object and the distance from the distance measuring sensor 20, the size of the object, Height or both may be used.

  For example, when comparing the size of an object in the object included in the first result and the object included in the second result, the size of the object included in the first result and the size of the object included in the second result The comparison result is set so that the value increases as the difference decreases. When comparing the heights of the object included in the first result and the object included in the second result, the difference between the height of the object included in the first result and the height of the object included in the second result is The comparison result is set so that the value increases as the value decreases. Whether the comparison is based on the size of the object or the comparison based on the height of the object, the size and height of the object will not change between the previous detection and the detection result of this time. There will be no major changes. Therefore, the reliability of the comparison result can be improved by setting the comparison result based on the object size and the comparison result based on the object height as described above.

  In calculating the similarity using the object shape, the distance from the distance measuring sensor 20, the size and height of the object, the result obtained by multiplying the first weighting factor by the matching degree of the feature amount and the second weighting factor The total value of the result obtained by multiplying the comparison result of the center of gravity position, the result obtained by multiplying the third weighting factor by the comparison result on the size of the object, and the result obtained by multiplying the fourth weighting factor by the comparison result on the height, Become. The total value of the first to fourth weighting factors is “1”.

  In calculating the similarity using the shape of the object, the distance from the distance measuring sensor 20, and the size of the object, the result obtained by multiplying the first weighting factor by the matching degree of the feature amount, This is the total value of the result obtained by multiplying the comparison result and the result obtained by multiplying the third weighting coefficient by the result of comparing the object with the size. The total value of the first to third weighting factors is the value “1”.

  In calculating the similarity using the shape of the object, the distance from the distance measuring sensor 20, and the height of the object, the result obtained by multiplying the first weighting factor by the matching degree of the feature amount, This is the total value of the result obtained by multiplying the comparison result and the result obtained by multiplying the fourth weighting factor by the comparison result for the size of the object. Note that the total value of the first, second, and fourth weighting factors is the value “1”.

  Further, when calculating the similarity between the object included in the first result and the object included in the second result, the tracking processing unit 104 determines the shape of the object, the distance from the distance measuring sensor 20, the size of the object, Of the height, at least one element may be used.

  When calculating the similarity between the object included in the first result and the object included in the second result, the accuracy of the similarity is obtained by adding other elements in addition to the shape of the object and the distance from the distance measuring sensor 20. Can be improved.

  In the above embodiment, when the object identification unit 105 calculates the similarity between the object included in the detection result of the distance sensor 20 and the model data, the shape of the body and the distance from the distance sensor 20 are also calculated. Alternatively, the size and / or height of the object may be used. Alternatively, when calculating the similarity between the object included in the first result and the object included in the second result, the object identification unit 105 determines the shape of the object, the distance from the distance measuring sensor 20, the size of the object, Of the height, at least one element may be used.

  In the above embodiment, the tracking processing unit 104 is configured to determine the center of gravity of the object when determining the distance from the distance measuring sensor 20 to the object, but is not limited to this configuration. The tracking processing unit 104 may obtain the midpoint of the edge of the object and set the distance from the distance measuring sensor 20 to the midpoint of the edge as the distance from the distance measuring sensor 20 to the object.

  In the above embodiment, the user may specify (set) an object (exclusion target) that the object detection unit 102 does not regard as an obstacle. In this case, for example, the distance measuring system 5 accepts exclusion information including the height of the exclusion target object by a user operation and stores it in the storage unit 30. For example, when the user wants to exclude the utility pole, the exclusion information including the height of the utility pole is input. Thereby, since the detection apparatus 10 of the distance measuring system 5 excludes an object having the same height as that included in the exclusion information, the utility pole is excluded.

  In the above embodiment, the tracking processing unit 104 is configured to track an object using the current detection result and the previous detection result, but is not limited to this configuration. The tracking processing unit 104 may track the object using the current detection result and the detection result a predetermined number of times before the current time. That is, the tracking processing unit 104 removes an object to be removed from the detection result (current result) detected by the distance measuring sensor 20 at the present time and the result detected by the distance measuring sensor 20 retroactively for a predetermined time from when the current result is detected. The configuration may be such that the object is tracked using the detection result (past result) excluding.

  In the said embodiment, although the vehicle 6 was demonstrated as an example as a moving body, a moving body is not limited to the vehicle 6. FIG. The moving body is not limited to the vehicle 6 but may be an aircraft, a ship, or a walking robot.

  The application destination of the distance measuring system 5 is not limited to a moving body. The ranging system 5 may be fixed to a building or the like. By applying the distance measuring system 5 to a building, it can be used as a security monitoring system.

  The above embodiment is only one of various embodiments of the present disclosure. The above embodiment can be variously modified according to the design and the like as long as the object of the present disclosure can be achieved. Moreover, the same function as the sensor processing system 1 may be embodied by a sensor processing method, a computer program, or a recording medium on which the program is recorded. The sensor processing method of the sensor processing system 1 according to one aspect includes an acquisition step and an output step. In the acquisition step, the result detected by the distance measuring sensor 20 is acquired. The output step outputs a detection result obtained by removing the removal object from the result detected by the distance measuring sensor 20 based on the map information including the height data of the object that is the detection target of the distance measuring sensor 20. The program according to one aspect is a program for causing a computer system to function as the above-described sensor processing method.

  The execution subject of the sensor processing system 1 or the sensor processing method in the present disclosure includes a computer system. The computer system has a processor and memory as hardware. When the processor executes a program recorded in the memory of the computer system, a function as an execution subject of the sensor processing system 1 or the sensor processing method according to the present disclosure is realized. The program may be recorded in advance in the memory of the computer system, but may be provided through a telecommunication line. The program may be provided by being recorded on a non-transitory recording medium such as a memory card, an optical disk, or a hard disk drive that can be read by a computer system. A processor of a computer system includes one or a plurality of electronic circuits including a semiconductor integrated circuit (IC) or a large scale integrated circuit (LSI). The plurality of electronic circuits may be integrated on one chip, or may be distributed on the plurality of chips. The plurality of chips may be integrated into one device, or may be distributed and provided in a plurality of devices.

  The sensor processing system 1 that is a computer system may be a system constituted by one or a plurality of computers. For example, at least a part of the functions of the sensor processing system 1 may be realized by cloud (cloud computing).

(Summary)
As described above, the sensor processing system (1) of the first aspect includes the acquisition unit (101) that acquires the result detected by the distance measuring sensor (20) and the output unit (106). The output unit (106) removes the object to be removed from the result detected by the distance sensor (20) based on the map information including the height data of the excluded object to be excluded from the result detected by the distance sensor (20). Outputs the detection result excluding.

  According to this configuration, the sensor processing system (1) excludes the removal object from the result detected by the distance measuring sensor (20). Thereby, since the sensor processing system (1) can exclude the object which is not regarded as an obstacle as an object to be excluded, the accuracy of obstacle detection can be further increased.

  The sensor processing system (1) according to the second aspect further includes an object detection unit (102) that generates a detection result obtained by removing the removal target object from the result detected by the distance measuring sensor (20) in the first aspect. Prepare. The object detection unit (102) reduces the amount of removal in the height direction of the removal target when there is an object that is not the removal target in the height direction of the removal target compared to the case where no object exists. To do.

  According to this configuration, it is possible to reduce the possibility of excluding a part of an object as an obstacle. Thereby, compared with the case where a part of the object as an obstacle is deleted, the sensor processing system (1) can be accurately identified when identifying the object.

  In the sensor processing system (1) of the third aspect, in the first or second aspect, the map information includes information on the current position of the distance measuring sensor (20).

  According to this configuration, the sensor processing system (1) can detect the presence or absence of an obstacle around the distance measuring sensor (20).

  In the sensor processing system (1) of the fourth aspect, in any one of the first to third aspects, in the map information, the area including the distance measuring sensor (20) is divided into a plurality of sections and removed for each section. The removal amount in the height direction of the object is set.

  According to this configuration, the removal amount can be changed according to the classification. Thereby, the sensor processing system (1) can perform detailed removal according to a division, or can perform simple removal.

  The sensor processing system (1) of the fifth aspect further includes a tracking processing unit (104) in any of the first to fourth aspects. The tracking processing unit (104) detects the first detection result as a detection result and the result of detection by the distance measurement sensor (20) by going back a predetermined time from when the distance measurement sensor (20) detects the first detection result. The second detection result obtained by removing the object to be removed from the object is compared, and the object existing in the detection target region of the distance measuring sensor (20) is tracked.

  According to this configuration, it is possible to track an object using a result detected at the present time and a result detected in the past. Therefore, for example, a moving body or the like that performs automatic driving can accurately determine avoidance of the obstacle according to the movement (movement) of the obstacle.

  The sensor processing system (1) of the sixth aspect further includes an object identification unit (105) in any of the first to fifth aspects. The object identification unit (105) identifies one or more objects included in the detection result using a plurality of model data respectively representing features of the plurality of objects.

  According to this configuration, the sensor processing system (1) can identify an object as an obstacle. For this reason, for example, a moving body or the like that performs automatic driving can accurately determine avoidance of the obstacle according to the identified obstacle.

  The distance measuring system (5) according to the seventh aspect includes the sensor processing system (1) according to any one of the first to sixth aspects and the distance measuring sensor (20).

  According to this configuration, since the distance measuring system (5) can exclude an object that is not regarded as an obstacle as an object to be excluded, the accuracy of obstacle detection can be further increased.

  The moving body (for example, vehicle 6) according to the eighth aspect includes the distance measuring system (5) according to the seventh aspect and a main body (61).

  According to this configuration, since the moving body can exclude an object that is not regarded as an obstacle as an object to be excluded, the accuracy of obstacle detection can be further increased.

  The sensor processing method according to the ninth aspect includes an acquisition step of acquiring a result detected by the distance measuring sensor (20), and an output step. The output step outputs a detection result obtained by removing the removal target from the result detected by the distance measuring sensor (20) based on the map information including the height data of the object that is the detection target of the distance measuring sensor (20). To do.

  According to this sensor processing method, since an object that is not regarded as an obstacle can be excluded as an object to be excluded, the accuracy of obstacle detection can be further improved.

  The program according to the tenth aspect is a program for causing a computer system to execute an acquisition step of acquiring a result detected by the distance measuring sensor (20) and an output step. The output step outputs a detection result obtained by removing the removal target from the result detected by the distance measuring sensor (20) based on the map information including the height data of the object that is the detection target of the distance measuring sensor (20). To do.

  According to this program, since an object that is not regarded as an obstacle can be excluded as an object to be excluded, the accuracy of obstacle detection can be further improved.

DESCRIPTION OF SYMBOLS 1 Sensor processing system 5 Ranging system 6 Vehicle 10 Detection apparatus 20 Ranging sensor 61 Main body part 101 Acquisition part 102 Object detection part 104 Tracking processing part 105 Object identification part 106 Output part

Claims (10)

An acquisition unit for acquiring a result detected by the distance measuring sensor;
An output unit that outputs a detection result obtained by removing the removal object from the result detected by the distance sensor based on map information including data on the height of the object to be excluded from the result detected by the distance sensor. A sensor processing system comprising: An object detection unit that generates a detection result obtained by removing the removal target from the result detected by the distance measuring sensor;
When there is an object that is not a removal target in the height direction of the removal target object, the object detection unit calculates a removal amount in the height direction of the removal target object compared to a case where the object does not exist The sensor processing system according to claim 1, wherein the sensor processing system is reduced. The sensor processing system according to claim 1, wherein the map information includes information on a current position of the distance measuring sensor. The area including the distance measuring sensor is divided into a plurality of sections in the map information, and the removal amount in the height direction of the removal target is set for each section. The sensor processing system according to any one of claims. A first detection result as the detection result and a second detection in which a removal target is removed from a result detected by the distance measuring sensor retroactively for a predetermined time from when the distance measuring sensor detects the first detection result. The sensor processing according to any one of claims 1 to 4, further comprising a tracking processing unit that compares the result and tracks an object existing in a detection target region of the distance measuring sensor. system. The object identification unit for identifying one or more objects included in the detection result by using a plurality of model data respectively representing features of a plurality of objects, further comprising: The sensor processing system according to one item. The sensor processing system according to any one of claims 1 to 6,
A ranging system comprising the ranging sensor. A ranging system according to claim 7;
A moving body comprising: a main body portion. An acquisition step of acquiring a result detected by the distance measuring sensor;
An output step of outputting a detection result obtained by removing the removal target object from the result detected by the distance measurement sensor based on map information including the height data of the object that is the detection target of the distance measurement sensor. A sensor processing method characterized by the above. Computer system,
An acquisition step of acquiring a result detected by the distance measuring sensor;
An output step of outputting a detection result obtained by removing a removal target from the result detected by the distance sensor based on map information including height data of an object which is a detection target of the distance sensor; Program to let you.

Images