Object detection device and object detection method

Abstract

To provide an object detection device and an object detection method capable of coping with a complicated structure and detecting an object by automatically detecting a background area.SOLUTION: An object detection device 10 includes: a cube processing unit 11a installed on the ground to scan a scanning range by intermittently irradiating light, receive point group information from a LiDAR 2 for receiving reflection light of the light to acquire the point group information based on the reflection light, and detect a background area being an area of a stationary object within a prescribed space S on the basis of point group information inputted multiple times during a prescribed time; and a detection unit 11b for detecting an object in the prescribed space S on the basis of point group information included in an area other than the background area detected by the cube processing unit 11a. Then, the cube processing unit 11a detects the background area on the basis of a detection state of a point group in each of a plurality of cubic areas obtained by dividing the prescribed space S by an octree.SELECTED DRAWING: Figure 2

Description

[Technical Field] 【0001】 The present invention relates to an object detection device and an object detection method for detecting objects on a road, for example. [Background technology] 【0002】 For example, as an object detection device for detecting objects on a road, a method is known that uses clusters generated by clustering point clouds detected by LiDAR (Light Detection and Ranging) to detect objects. 【0003】 For example, Patent Document 1 describes how to detect objects with high accuracy in the correct units by determining if clusters are over-coupled based on image targets detected from ambient light images and clusters obtained by clustering point clouds, splitting over-coupled clusters if it is determined that clusters are over-split, and combining two or more clusters present in the portion of the point cloud corresponding to the image target if it is determined that clusters are over-split. [Prior art documents] [Patent Documents] 【0004】 [Patent Document 1] Japanese Patent Publication No. 2021-131385 [Overview of the project] [Problems that the invention aims to solve] 【0005】 As described in Patent Document 1, it is possible to remove the influence of stationary objects (background areas) such as houses and walls by detecting overcoupling, but this has the problem of complicating the process. 【0006】 Therefore, one method is to detect only moving objects by detecting the background region in advance and excluding the point cloud contained within that background region. However, if, for example, an area with trees is defined as the background region in a planar manner, pedestrians walking under the branches of the trees may not be detected. 【0007】 Furthermore, the stationary objects that make up the background area are not limited to relatively simple shapes such as walls. For example, vegetation is inherently complex in shape, and its form changes due to growth and the presence or absence of leaves depending on the season. There are also environmental changes such as road construction and snowdrifts. Traditionally, such environmental changes were checked manually or the background area was remeasured, but this is very time-consuming, so automation is desirable. 【0008】 Therefore, the present invention aims to provide an object detection device and object detection method that can handle complex structures and automatically detect background areas and objects. [Means for solving the problem] 【0009】 The invention described in claim 1, made to solve the above problems, is an object detection device comprising: an input unit that receives point cloud information from a sensor installed on the ground, which intermittently irradiates light to scan a predetermined space, receives reflected light from the light and acquires point cloud information based on the reflected light; a detection unit that detects a background region which is the region of stationary objects in the predetermined space based on the point cloud information input to the input unit multiple times within a predetermined time; and an object detection unit that detects objects in the predetermined space based on the point cloud information included in the region other than the background region detected by the detection unit, wherein the detection unit detects the background region based on the detection state of the point cloud in each of the multiple cubic regions into which the predetermined space is divided by an octree. [Effects of the Invention] 【0010】 According to the present invention, since the background region is detected based on the detection state of the point group in each of a plurality of cubic spaces obtained by dividing a predetermined space by an octree, even if the shape of the background region is complex, it can be approximated by a combination of cubes. Further, by inputting point group information, it becomes possible to automatically detect the background region. 【Brief Description of Drawings】 【0011】 [Figure 1] This is an installation example of an object detection device according to an embodiment of the present invention. [Figure 2] This is a block diagram of the object detection device shown in FIG. 1. [Figure 3] This is a flowchart of the operation of the object detection device shown in FIG. 2. [Figure 4] This is an explanatory diagram of a polygonal closed region. [Figure 5] This is an explanatory diagram of an example of dividing a predetermined space. [Figure 6] This is an explanatory diagram of an octree structure. [Figure 7] This is a diagram showing the relationship between the installation position of the sensor and the observation point in the x-y plane. [Figure 8] This is an explanatory diagram of a case where the vertex with the shortest distance does not become the shortest distance from the reference point of the sensor to the cube region. [Figure 9] This is an explanatory diagram of the change in the size of the cube region according to the distance from the sensor. [Figure 10] This is an explanatory diagram of the background determination threshold. [Figure 11] This is an explanatory diagram regarding the integration of child nodes. 【Embodiments for Carrying Out the Invention】 【0012】 Hereinafter, an embodiment of the present invention will be described with reference to FIGS. 1 to 11. FIG. 1 is an installation example of an object detection device according to an embodiment of the present invention. 【0013】 The object detection device 10 shown in Figure 1 is installed on a support column P fixed to the ground, for example, near an intersection. The object detection device 10 detects moving objects such as pedestrians and vehicles passing through the intersection. In other words, the object detection device 10 is installed on the ground. This means that it is fixed to the ground, whether directly or indirectly; that is, it is installed on an object that is fixed to the ground, such as a building, rather than on a moving object such as a vehicle. 【0014】 The object detection device 10 comprises a main body 1 and a LiDAR 2. The main body 1 is provided as a control box or the like at the lower part of the support column P. As shown in the block diagram of Figure 2, the main body 1 comprises a control unit 11 and a communication unit 12. 【0015】 The control unit 11 is composed of, for example, a computer. The control unit 11 is responsible for the overall control of the object detection device 10. In Figure 1, the control unit 11 functionally comprises a cube processing unit 11a and a detection unit 11b. 【0016】 The cube processing unit 11a divides the point cloud information acquired by LiDAR2 into multiple cubic regions, as described later, and determines whether each cubic region should be a background region or a detection target region. The background region is a region of stationary objects other than moving objects such as vehicles and pedestrians, and is excluded from object detection by the detection unit 11b. The detection unit 11b detects objects based on the background regions defined by the cube processing unit 11a. 【0017】 The communication unit 12 outputs information about the object detected by the control unit 11 to the outside. The communication unit 12 can be configured as, for example, a wireless communication device (it may also be a wired communication device). Information about the object includes the type of object, direction of movement, and speed of movement. Specific examples of output destinations include moving objects such as vehicles passing near the support column P, and traffic control centers. 【0018】 LiDAR2 is a sensor that scans a predetermined space by intermittently irradiating it with laser light, receives the reflected light, and acquires point cloud information (3D point cloud information) based on the reflected light. LiDAR2 discretely measures the distance to objects in the space within the irradiation range based on the reflected light of the irradiated laser light, and recognizes the position and shape of the objects as a 3D point cloud. In this embodiment, as shown in Figure 1, LiDAR2 is installed on the top of the support column P and scans the road surface R (road surface R and the space above it) as the predetermined space. 【0019】 Next, the operation of the object detection device with the above configuration will be explained with reference to the flowchart in Figure 3. First, the control unit 11 acquires observation points from the LiDAR 2 (step S1). In other words, the control unit 11 acquires point cloud information acquired by the LiDAR 2. That is, the control unit 11 functions as an input unit to which point cloud information is input from the LiDAR 2. 【0020】 Next, the cube processing unit 11a determines whether the observation point is within a polygonal closed region (step S2). A polygonal closed region will be explained with reference to Figure 4. The left side of Figure 4 shows a predetermined space S that includes the polygonal closed region CL. In this embodiment, as will be described later, a predetermined space S in the shape of a cube that includes the space that LiDAR2 can scan is defined, and this predetermined space S is divided into multiple cube regions by an octree, and this is repeated recursively. However, in a realistic usage environment for LiDAR2, especially when detecting distant vehicles on a road, there will be a large amount of excess space in the width and height directions relative to the depth direction. This excess space is defined in advance as a background region and is excluded from the background region detection process described later. 【0021】 Specifically, a planar polygon is defined in the width (x)-depth (y) plane by multiple vertex coordinates (x, y) (Figure 4, upper right). Next, the planar polygon is moved in the height (z) direction z min ~z maxThe region is extended to define a three-dimensional closed region (Figure 4, lower right). For observation points, if they are inside the closed region, they are subject to the cubic region division and background region determination described later; if they are outside the closed region, they are excluded from subsequent processing as part of the background region. The polygonal closed region CL shown in Figure 4 is set in the control unit 11 in advance before executing the flowchart in Figure 3. 【0022】 Note that while Figure 3 shows a quadrilateral as the polygon, it is not limited to quadrilaterals. However, a quadrilateral shape is preferable for the purpose of dividing the region into cubes. Also, the above-mentioned planar polygons and z min ~z max This can be determined based on the results of measurements taken manually or by other means at the installation location of the object detection device 10. In other words, the cube processing unit 11a sets a region within a predetermined space S that is limited in both the horizontal and vertical directions, and detects the background region within that limited region. 【0023】 Next, the cube processing unit 11a determines whether the cube region has not yet been created (step S3). If the cube region has not yet been created (step S3; YES), the size of the cube region is determined (step S4). On the other hand, if the cube region has not yet been created (step S3; NO), the process proceeds to step S6, which will be described later. 【0024】 Here, the spatial partitioning in this embodiment will be explained with reference to Figures 5 and 6. In Figure 5, for the sake of simplicity, the explanation is given using a predetermined cubic space S in which the process in step S2 has not been performed, but in reality, the following process is performed within the polygonal closed region described above. 【0025】 Figure 5 is an explanatory diagram illustrating an example of the division of a predetermined space S in this embodiment. The predetermined space S shown in Figure 5 includes the space that LiDAR2 can scan. In this embodiment, the predetermined space S is defined as a cube, and the cube is divided using an octree. By recursively dividing it using an octree, the predetermined space S can be divided into small cubic regions, such as 10 cm square. 【0026】 FIG. 6 is a diagram showing an octree structure. As is well known, an octree is a tree structure in which one parent node has eight child nodes. And the octree structure can quickly reference any node (cubic region) by a well-known method such as Morton order. This octree structure is formed in the memory of the control unit 11 or the like, and information indicating whether the node (cubic region) is a background region and the values of various counters described later are stored in each node. 【0027】 In step S3, it is determined whether a cubic region has already been generated by the processes executed in steps S4 and S5. 【0028】 Next, the process of step S4 will be described with reference to FIGS. 7 and 8. Since the spatial resolution of LiDAR2 decreases as the distance increases, applying the size of the cubic region suitable for the vicinity of LiDAR2 as shown in FIG. 7 to the far side of LiDAR2 results in excessive region division. 【0029】 FIG. 7 is a diagram showing the relationship between the installation position of LiDAR2 and the observation points in the x - y plane. In FIG. 7, the symbol L is the installation position of the LiDAR, and the symbols P 11 , P 12 are the observation points at the distance l1 from the installation position L of LiDAR2, and the symbols P 21 , P 22 are the observation points at the distance l2 (l1 < l2) from the installation position L of LiDAR2, respectively. And the observation points P 11 , P 12 and the observation points P 21 , P 22 are adjacent observation points (continuous pulsed light) to each other. Here, assuming that the division is made with the sizes shown by the symbols Ca1 and Cb1 as the cubic regions, at the distance l1, the observation points P 11 , P 12 may be observed in the cubic regions Ca1 and Cb1, but at the distance l2, the observation points P 21 , P 22 will not be observed in the cubic regions Ca2 and Cb2 of the same size. That is, at the distance l2, the cubic region is excessively divided. 【0030】 Therefore, in this embodiment, the number of recursive partitions by the octree is changed depending on the distance from LiDAR2 to the cubic region. Specifically, for the target cubic region, the distance from the reference point (installation position) of LiDAR2 to the vertex that is the shortest distance away is calculated, and the distance between observation points in the width direction at that position is calculated. The distance between observation points can be calculated from, for example, the irradiation interval of the laser beam or the operating speed (frequency) of the scanning means such as mirrors. 【0031】 In Figure 7, the reference point is the installation position L of LiDAR2, the vertex with the shortest distance is the vertex Pv that is closest to the installation position L among the vertices that form the boundary of the cubic regions Ca1 and Cb1, and the distance between observation points in the width direction corresponds to the interval lx between observation points in the x direction at vertex Pv. 【0032】 In the case of Figure 7, the angle between the line connecting the installation position L and vertex Pv and the x-axis is almost a right angle (almost the shortest distance), so the distance between the installation position L and vertex Pv can be used as is. However, as shown in Figure 8, there are cases where the vertex that is the shortest distance is not the shortest distance from the sensor's reference point to the cubic region. In this case, the intersection point P3 that is the shortest distance from the vertex and the sensor's reference point to the cubic region should be found and used to calculate the distance between observation points. 【0033】 Next, half the distance between observation points in the width direction is set as the threshold value for the size of the cubic region. This threshold value is set to half the distance between observation points in the width direction in order to suppress the positional shift between the observation points and the center of the cube due to changes in the size of the cubic region. 【0034】 Next, the number of divisions for the cube region is determined such that the size of the cube region is the largest possible size below a threshold. By determining the number of divisions for the cube region in this way, the size of the cube region increases as you move away from the sensor (LiDAR2), as shown in Figure 9. In other words, the upper limit of the number of divisions by the octree is determined based on the spacing of the observation points in the width direction by the sensor. 【0035】 Figure 9 is a graph showing the distance between the sensor and the cubic region on the horizontal axis and the size of the cubic region on the vertical axis. The dashed line shows the change in the width direction by half the distance between the shortest observation points. The solid line shows the length of one side of the cubic region. d represents the minimum size of the cubic region. As will be explained later, d is a value set to prevent the size of the cubic region from becoming too small, and as shown in Figure 9, it is defined by the length of one side of the cubic region. 【0036】 If the threshold size of the cubic region falls below a predetermined minimum size, the system uses the predetermined number of divisions required to reach that minimum size. This means that a minimum size for the cubic region is set in advance, and the region is not divided below that size. In the division method described above, the cubic region becomes smaller as it approaches LiDAR2, but dividing it too small relative to the object to be detected is also excessive. Therefore, by setting a minimum size in advance, excessive region division is suppressed. 【0037】 Returning to the explanation of Figure 3, the cube processing unit 11a generates a cubic region (step S5). In step S5, the cubic region is generated using the method described above. 【0038】 Next, the cube processing unit 11a identifies the corresponding cube region (step S6). In step S6, it identifies which of the recursively divided, terminal node cube regions, as shown in Figure 5, each point included in the point cloud information acquired in step S1 belongs to. In other words, it identifies the terminal node cube region whose coordinates are included within its range. 【0039】 Next, the cube processing unit 11a adds (+1) the number of observations of the cube region (end node) identified in step S2 (step S7). The result of the addition is stored, for example, in the memory of the control unit 11. 【0040】 Next, the cube processing unit 11a determines whether a predetermined time has elapsed (step S8). If the predetermined time has not elapsed (step S8; NO), the process returns to step S1. Step S8 determines the period for adding the number of observations, and the predetermined time should be set appropriately to allow multiple observation points (point cloud information) to be acquired depending on the point cloud acquisition interval, etc. 【0041】 On the other hand, if a predetermined time has elapsed (step S8; YES), the cube processing unit 11a determines whether the number of observations added in step S3 is equal to or greater than a threshold (step S9). If the number of observations is equal to or greater than a threshold (observation count threshold) (step S9; YES), the cube processing unit 11a defines (detects) the currently focused cube region as the background region (step S10). 【0042】 Next, the cube processing unit 11a increments the background determination count (step S11). The background determination count is the number of times that an area was defined as a background region in step S10. 【0043】 Next, the cube processing unit 11a determines whether the number of background detections is equal to or greater than a threshold (step S12). If the number of background detections is equal to or greater than the threshold (background detection threshold) (step S12; YES), the cube processing unit 11a updates the classification of the cube region defined as the background region in step S10 (step S13). On the other hand, if the number of background detections is less than the threshold (step S12; NO), the process proceeds to step S14, which will be described later, without executing step S13. 【0044】 Specifically, the cube processing unit 11a functions as a detection unit that detects background regions, which are areas of stationary objects, within a predetermined space based on point cloud information input to the input unit multiple times within a predetermined time period. It detects background regions based on the point cloud detection status in each of the multiple cube-shaped regions into which the predetermined space is divided by an octree. Furthermore, the cube processing unit 11a determines that a cube-shaped region is a background region if the number of point cloud detections within a predetermined time period in that region is equal to or greater than a predetermined threshold. 【0045】 Steps S11 to S13 will now be explained. When the background area is determined using criteria like those in step S9, some objects, such as parked cars or road construction, are temporarily determined to be part of the background area, while others, such as buildings, are always determined to be part of the background area. Therefore, in step S13, the objects determined to be part of the background area are classified. 【0046】 For example, as shown in Figure 10, multiple background detection thresholds are set, and objects are classified according to these thresholds. In Figure 10, the classifications are as follows: 0. Outside the background, 1. Moving object, 2. Stationary object (short-term remaining objects such as parked vehicles), 3. Stationary object (medium-term remaining objects such as equipment and vehicles at construction sites), and 4. Stationary object (long-term remaining objects such as buildings). Thresholds 1 and T are set as the boundary thresholds for each classification. 12 , threshold T 23 , threshold T 34 Set the following. Note that although the background determination threshold is described as "parked vehicles, etc.," this is merely an example of background area classification and does not involve the identification of specific objects such as vehicles. Therefore, steps S11 to S13 are different from object detection processes such as clustering, which will be described later. 【0047】 If the background detection count counter for the cube region being judged is 0, the classification is 0 (outside the background). When the counter reaches 1, the classification becomes 1 (moving object). Thereafter, the counter reaches the threshold T. 12 ,T 23 ,T 34 At this point, the classifications transition from 1→2, 2→3, and 3→4, respectively. Threshold T 12 ,T 23 ,T 34 This is a variable and should be set appropriately according to the installation environment of the object detection device 10. Specifically, the cube processing unit 11a is equipped with a counter that counts the number of times a cube-shaped region is detected as a background region, and classifies the background region according to the count of that counter. 【0048】 Next, the cube processing unit 11a resets the number of observations and the elapsed time (step S14). The cube processing unit 11a resets the number of observations counted in step S9 and the elapsed time determined in step S8. 【0049】 On the other hand, in the determination in step S9, if the number of observations is not equal to or greater than the threshold (step S9; NO), the cube processing unit 11a defines the currently focused cube region as the detection target region (step S15). The detection target region is the region in which object detection is performed by clustering the point cloud, as will be described later. 【0050】 Next, the cube processing unit 11a determines whether the classification of the currently focused cube region is anything other than "0" (step S16). In step S16, it determines whether the classification is anything other than "0" among the classifications 0 to 4 shown in Figure 10, i.e., whether it is outside the background. If the classification is anything other than "0", the cube processing unit 11a determines whether it has been determined to be a target region for a longer period than the number of times it has been held (step S17). If it has been determined to be a target region for a longer period than the number of times it has been held (step S17; YES), the background determination count counter is reset (step S18). If it has not been determined to be a target region for a longer period than the number of times it has been held (step S17; NO), step S18 is not executed and the process proceeds to step S14. 【0051】 Steps S16 to S18 are explained below. When the background detection count counter for a given cubic region is 1 or more, and the cubic region is determined to be a detection target region, the background detection count counter is basically reset to 0. However, a set number of retention times for each classification is set, and up to that number of retention times, the background detection count counter and classification will not be reset even if the region is determined to be a detection target region. For example, in the case of parking meters where parked vehicles change frequently, in the short term it may be classified as a short-term residual object of classification "2", but in the long term it may be appropriate to treat it as a medium-term residual object of classification "3". Therefore, by setting a retention time, it is possible to prevent the classification from being reset to "0" when a vehicle moves. 【0052】 Steps S9 through S18 are performed individually for each cubic region. In other words, they are repeated for the number of cubic regions. For example, if there are 4096 cubic regions, the process is repeated 4096 times. 【0053】 Next, the cube processing unit 11a determines whether there is a parent node in which a certain number of child nodes are in the background area (step S19). The certain number in step S19 could be, for example, a majority. 【0054】 If a certain number of child nodes are part of a parent node that is part of the background region (step S19; YES), the cube processing unit 11a integrates the child nodes that have that parent node (step S20). Steps S19 and S20 will be explained with reference to Figure 11. The left side of Figure 11 shows a cube region c that has been determined to be part of the background region in a predetermined space S. In addition, in the left side of Figure 11, multiple cube regions c1 are child nodes that have the same parent node, and multiple cube regions c2 are child nodes that have the same parent node. 【0055】 In this case, if a certain number of child nodes or more are determined to be background regions, the entire cube region of the parent node is treated as the background region. In other words, child nodes other than those determined to be background regions are also treated as background regions. The right side of Figure 11 shows the parent nodes of cube regions c1 and c2 as background regions. In the right side of Figure 11, the code cb1 indicates the region representing the parent node of cube region c1, and the code cb2 indicates the region representing the parent node of cube region c2. That is, if a predetermined number or more of the child nodes connected to the parent node of the octvine structure are determined to be background regions, the cube processing unit 11a treats the entire region indicated by the parent node as the background region. 【0056】 Returning to the explanation of Figure 3, the detection unit 11b determines whether the cubic region identified in step S5 is the region to be detected (step S21). This step S21 is performed after steps S20 have been completed and the definition of the background region in the predetermined space S has been completed at least once. In other words, the observation points (point cloud information) acquired in step S1 are processed separately from the processing in steps S7 to S20, and the processing from step S21 onwards is also performed. 【0057】 If the identified cubic region is the region to be detected (step S21; YES), the detection unit 11b adds the observation point to the object detection processing target (step S22). 【0058】 In step S22, the observation points that have become targets for object detection are subjected to, for example, clustering processing by the detection unit 11b, and detection target identification processing is performed on the generated clusters. Detection target is a process that identifies the type of moving object, such as a pedestrian or vehicle, based on the shape and size of the cluster. Once a detection target is identified, the detection unit 11b performs, for example, tracking processing on the target and monitors its movement within the predetermined space S. In other words, the detection unit 11b functions as an object detection unit that detects objects in a predetermined space based on point cloud information included in areas other than the background area detected by the detection unit. 【0059】 As is clear from the above explanation, step S1 functions as an input step, steps S2 to S20 as detection steps, and steps S21 and S22 as object detection steps. 【0060】 According to this embodiment, the object detection device 10 includes a cube processing unit 11a that receives point cloud information from a LiDAR 2 installed on the ground, which intermittently irradiates light to scan a scanning range and receives reflected light to acquire point cloud information based on the reflected light, and detects a background region which is the region of a stationary object in a predetermined space S based on the point cloud information received multiple times within a predetermined time, and a detection unit 11b that detects an object in the predetermined space S based on the point cloud information included in the region other than the background region detected by the cube processing unit 11a. The cube processing unit 11a then detects the background region based on the detection state of the point cloud in each of the multiple cube regions into which the predetermined space S is divided by an octree. 【0061】 With the object detection device 10 configured as described above, even if the shape of the background region is complex, it can be approximated by combining cubes. Furthermore, by inputting point cloud information, the background region can be automatically detected, eliminating the need for manual settings. In addition, the background region can be automatically updated at regular intervals, enabling it to respond to changes in the surrounding environment such as vegetation, road construction, and snowdrifts. 【0062】 Furthermore, the cube processing unit 11a determines that a cube region is a background region if the number of point cloud detections within a predetermined time period in that region exceeds a predetermined threshold. By doing so, it is possible to assume the presence of a stationary object if point clouds are detected continuously for a certain period of time. 【0063】 Furthermore, if the cube processing unit 11a detects that a predetermined number or more of the child nodes connected to the parent node of the octvine structure are background regions, it treats the entire region indicated by that parent node as a background region. By doing this, for example, if the majority of the child nodes are determined to be background regions, the region of the parent node is considered a background region, eliminating the need to trace down to the child nodes when identifying background regions. Therefore, it becomes possible to access background regions at high speed. 【0064】 Furthermore, the cube processing unit 11a is equipped with a counter that counts the number of background detections for the cube region, and classifies the background region according to the count of the counter. In this way, it is possible to change the response when a stationary object is removed for each classification. For example, if a stationary object is observed in the short term, it is possible to respond quickly to the environmental change, and if a stationary object is observed in the long term, it is possible to delay the response for confirmation. 【0065】 Furthermore, the cube processing unit 11a sets a polygonal closed region within a predetermined space S, with a limited range in the horizontal and height directions, and detects the background region within this polygonal closed region. In this way, in a realistic usage environment for LiDAR2, it is possible to pre-define the background as the space that is surplus in the width and height directions relative to the depth direction. Therefore, the number of cube regions to be generated can be reduced. 【0066】 Furthermore, the cube processing unit 11a divides the cube region so that the size of the cube region increases as the distance from the sensor within a predetermined space S increases. This prevents excessive region division and allows for appropriate division according to the spatial resolution of the LiDAR2. 【0067】 Furthermore, since the upper limit on the number of octree divisions is determined based on the spacing of observation points in the width direction by LiDAR2, excessive region division can be suppressed. 【0068】 It should be noted that the present invention is not limited to the embodiments described above. That is, those skilled in the art can implement the invention in various ways, without departing from the core principles, in accordance with prior art knowledge. As long as such modifications still incorporate the configuration of the object detection device and object detection method of the present invention, they are of course included within the scope of the present invention. [Explanation of Symbols] 【0069】 1. Main body 2 LiDAR (sensor) 10. Object detection device 11 Control Unit 11a Cube Processing Unit 11b Detection unit S predetermined area C cubic area

Claims

[Claim 1] An input unit receives point cloud information from a sensor installed on the ground that intermittently irradiates light to scan a predetermined space, receives the reflected light, and acquires point cloud information based on the reflected light. A detection unit detects a background region which is the region of a stationary object within a predetermined space based on the point cloud information that has been input to the input unit multiple times within a predetermined time period, The system includes an object detection unit that detects objects in a predetermined space based on the point cloud information included in a region other than the background region detected by the detection unit, The detection unit detects the background region based on the detection state of the point cloud in each of the multiple cubic regions obtained by dividing the predetermined space by an octree. An object detection device characterized by the following features. [Claim 2] The object detection device according to claim 1, characterized in that the detection unit determines the cubic region to be the background region when the number of detections of the point cloud within the predetermined time in the cubic region is equal to or greater than a predetermined threshold. [Claim 3] The object detection device according to claim 1, characterized in that if the detection unit detects that a predetermined number or more of the child nodes connected to the parent node of the octvine structure are the background region, then the entire region indicated by the parent node is considered the background region. [Claim 4] The detection unit includes a counter that counts the number of times the cubic region is detected as the background region, The detection unit classifies the background region according to the count of the counter. The object detection device according to feature 1. [Claim 5] The detection unit sets a region within the predetermined space in which the range in the horizontal plane and the height direction is limited, and detects the background region within the limited region. The object detection device according to feature 1. [Claim 6] The object detection device according to claim 1, characterized in that the detection unit is divided such that the size of the cubic region increases as the distance from the sensor in the predetermined space increases. [Claim 7] The object detection device according to claim 6, characterized in that the upper limit of the number of divisions by the octree is determined based on the spacing in the width direction of the observation points by the sensor. [Claim 8] An object detection method performed by an object detection device installed on the ground, which includes a sensor that intermittently irradiates light to scan a predetermined space, receives the reflected light, and acquires point cloud information based on the reflected light, An input step in which the point cloud information is input from the aforementioned sensor, A detection step that detects a background region which is the region of a stationary object within a predetermined space based on the point cloud information that has been input multiple times to the input step during a predetermined time period, The system includes an object detection step which detects an object in a predetermined space based on the point cloud information contained in a region other than the background region detected in the detection step, The detection step involves detecting the background region based on the detection state of the point cloud in each of the multiple cubic regions obtained by dividing the predetermined space by an octree. A method for detecting an object characterized by the following features.

Images