Obstacle intrusion detection method and system

Abstract

The invention discloses an obstacle intrusion detection method and system, and relates to the technical field of obstacle intrusion detection. According to the technical scheme, the obstacle avoidance detection area of the target object and the scanning distance measurement data of at least one laser radar on the target object are obtained, the target object is a detected object for detecting whether an intrusion obstacle exists in the obstacle avoidance detection area, and the target data set of the target object is obtained according to all the scanning distance measurement data; the target data set is a set constructed by the omnibearing point cloud data in the circumferential direction of the target object, so that whether an invasive obstacle exists in the obstacle avoidance detection area can be accurately judged according to the target data set, and when the obstacle exists in the obstacle avoidance detection area, a detection result representing the existence of the obstacle is output. The technical scheme has relatively high processing efficiency and accuracy when aiming at all-directional intrusion detection, so that the reliability of implementing all-directional intrusion detection on the detection object is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of obstacle intrusion detection, and in particular to an obstacle intrusion detection method and system. Background Art

[0002] There are many application requirements for obstacle intrusion detection. For example, in the sky rail collection and distribution system, the test line site is equipped with omnidirectional container transfer vehicles. During the automatic operation of the omnidirectional vehicle, it is necessary to consider maintaining a safe distance from surrounding obstacles, especially the safety and anti-collision issues between the sky rail track beam piers. The existing obstacle intrusion detection methods are relatively limited and have insufficient reliability when used in the above scenarios.

[0003] Therefore, how to improve the reliability of implementing all-round intrusion detection on the detection object is a technical problem that needs to be solved urgently. Summary of the invention

[0004] The invention provides an obstacle intrusion detection method and system, which can improve the reliability of implementing all-round intrusion detection on the detection object.

[0005] The embodiment of the present invention provides the following solution:

[0006] In a first aspect, an embodiment of the present invention provides an obstacle intrusion detection method, the method comprising:

[0007] Obtaining an obstacle avoidance detection area of ​​a target object and scanning ranging data of at least one laser radar on the target object, wherein the target object is a detected object to detect whether there is an intruding obstacle in the obstacle avoidance detection area;

[0008] Obtain a target data set of the target object according to all the scanning and ranging data, wherein the target data set is a set constructed by all-round point cloud data around the target object;

[0009] Determine whether there are intruding obstacles in the obstacle avoidance detection area based on the target data set;

[0010] When there is an obstacle in the obstacle avoidance detection area, a detection result indicating the existence of the obstacle is output.

[0011] In an optional embodiment, before obtaining a target data set of the target object according to all the scanning ranging data, the method further includes:

[0012] When there is a shielded area in the obstacle avoidance detection area, the scanning ranging data of the laser radar is eliminated according to the shielded area to obtain the scanning ranging data after the data elimination;

[0013] Performing data cleaning on the scanning ranging data after data elimination according to a preset ranging threshold value to obtain scanning ranging data after data cleaning;

[0014] The scanning distance measurement data that has completed data cleaning is input into a preset filtering processing model to obtain the scanning distance measurement data that has completed preprocessing.

[0015] In an optional embodiment, obtaining a target data set of the target object according to all the scanning ranging data includes:

[0016] According to the data coordinate system of the scanning ranging data of each laser radar and the central coordinate system of the target object, the relative offset position of each data coordinate system and the central coordinate system is obtained;

[0017] Implementing coordinate system conversion of the scanning ranging data corresponding to the laser radar according to each relative offset position, so as to convert the scanning ranging data of the data coordinate system into the central coordinate system;

[0018] According to the coordinate system conversion results of all scanning ranging data, the target data set of the target object is obtained.

[0019] In an optional embodiment, before determining whether there is an intruding obstacle in the obstacle avoidance detection area according to the target data set, the method further includes:

[0020] All data of the target data set are stored as a KD tree structure, and corresponding index labels are configured for each data in the KD tree structure.

[0021] In an optional embodiment, the obstacle avoidance detection area is a polygonal area composed of multiple boundary segments; judging whether there is an invading obstacle in the obstacle avoidance detection area according to the target data set includes:

[0022] According to the scanning points and obstacle avoidance detection areas of the target data set, a detection angle between each boundary segment of the obstacle avoidance detection area and the scanning point is obtained, wherein the vertex of the detection angle is the scanning point;

[0023] All detected angles of each scanning point are summed to obtain the summed angle result of each scanning point;

[0024] When the sum of the included angles is less than or equal to a preset angle threshold, the current scanning point is determined to be an intrusion point;

[0025] Determine whether there are obstacles to invasion at the location of the invasion point.

[0026] In an optional embodiment, obtaining a detection angle between each boundary segment of the obstacle avoidance detection area and the scanning point according to the scanning point and the obstacle avoidance detection area of ​​the target data set includes:

[0027] Determine multiple target boundary points on the obstacle avoidance detection area according to a preset interval value;

[0028] The closest data point to each target boundary point in the target data set is determined as the scanning point for intrusion detection;

[0029] According to the angle formed by the scanning point of intrusion detection and each boundary segment, the detection angle of the scanning point of intrusion detection is obtained.

[0030] In an optional embodiment, before determining that there is an intrusion obstacle at the location of the intrusion point, the method further includes:

[0031] Performing intrusion detection on a scanning point adjacent to the intrusion point to obtain intrusion detection results of the adjacent scanning points;

[0032] When the intrusion detection result indicates that the adjacent scanning points are also intrusion points, it is determined that there are intrusion obstacles at the locations where the intrusion points are located.

[0033] In a second aspect, an embodiment of the present invention further provides an obstacle intrusion detection system, the system comprising:

[0034] At least one laser radar installed on the target object, the laser radar is used to obtain scanning ranging data of the target object, wherein the target object is the object to be detected whether there is an intruding obstacle in the obstacle avoidance detection area;

[0035] The data processing terminal is connected to the laser radar and is used to obtain the obstacle avoidance detection area and the target data set constructed by obtaining the omnidirectional point cloud data of the target object based on all the scanning ranging data; determine whether there is an invading obstacle in the obstacle avoidance detection area based on the target data set; when there is an obstacle in the obstacle avoidance detection area, output the detection result indicating the existence of the obstacle.

[0036] In an optional embodiment, the system further includes:

[0037] The external terminal is connected to the data processing terminal, and is used to configure the associated parameters of the data processing terminal for implementing intrusion detection, and / or to issue corresponding alarms or displays according to the detection results, wherein the associated parameters include obstacle avoidance detection areas.

[0038] In an optional embodiment, the laser radar includes a first laser radar and a second laser radar, and the first laser radar and the second laser radar are arranged diagonally.

[0039] Compared with the prior art, the obstacle intrusion detection method and system of the present invention have the following advantages:

[0040] The technical solution of the present invention obtains the obstacle avoidance detection area of ​​the target object and the scanning ranging data of at least one laser radar on the target object. The target object is the object to be detected to detect whether there is an invading obstacle in the obstacle avoidance detection area. The target data set of the target object is obtained according to all the scanning ranging data. Since the target data set is a set constructed by the omnidirectional point cloud data around the target object, it can be accurately judged whether there is an invading obstacle in the obstacle avoidance detection area according to the target data set. When there is an obstacle in the obstacle avoidance detection area, a detection result representing the existence of the obstacle is output. The technical solution implements omnidirectional intrusion detection of the target object in the obstacle avoidance detection area through the scanning ranging data collected by the laser radar, and can quickly and accurately determine the obstacle intrusion in the obstacle avoidance detection area. It has high processing efficiency and accuracy for omnidirectional intrusion detection, thereby improving the reliability of omnidirectional intrusion detection of the detection object. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of this specification or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0042] Figure 1 A flowchart of an obstacle intrusion detection method provided by an embodiment of the present invention;

[0043] Figure 2 A schematic top view of implementing intrusion detection on a target object provided by an embodiment of the present invention;

[0044] Figure 3 A step diagram of implementing an intrusion detection method provided by an embodiment of the present invention;

[0045] Figure 4 A schematic diagram of the structure of an intrusion detection system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0046] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention belong to the scope of protection of the embodiments of the present invention.

[0047] The obstacle intrusion detection method proposed in the embodiment of the present invention can be applied to the intrusion detection of the target object in its obstacle avoidance detection area in any scenario. The target object can be a moving object or a stationary object, and no specific limitation is made here. The embodiment of the present invention will take an omnidirectional container transfer vehicle as an example to specifically explain how to improve the reliability of implementing all-round intrusion detection of the target object.

[0048] See also Figure 1 , Figure 1 The flowchart of an obstacle intrusion detection method provided by an embodiment of the present invention is a flowchart of an obstacle intrusion detection method provided by an embodiment of the present invention. The intrusion detection method can be applied to a controller or a computer for implementation. It only needs to be able to run the intrusion detection method. No specific limitation is made here. The detection method includes:

[0049] S11, obtaining an obstacle avoidance detection area of ​​a target object and scanning ranging data of at least one laser radar on the target object, wherein the target object is a detected object for detecting whether there is an intruding obstacle in the obstacle avoidance detection area.

[0050] Specifically, the obstacle avoidance detection area can be set based on the actual obstacle avoidance needs of the target object. It can be a circular area or a rectangular area on a plane; it can also be a three-dimensional area in space, and the three-dimensional area is composed of multiple plane areas stacked in layers. The laser radar can be installed on the top or side of the target object, and it can perform scanning and ranging around the target object. In order to improve the reliability of intrusion detection of the target object, the laser radar needs to perform omnidirectional scanning and ranging of objects around the target object. The number of laser radars can be determined based on their scanning range. For example, when the scanning and ranging range of the laser radar is 360°, only one laser radar can be configured; when the scanning and ranging range of the laser radar is 180°, four laser radars need to be configured to implement omnidirectional intrusion detection. Taking the target object installed in a rectangular structure as an example, the four laser radars are installed at four right angles to the target object.

[0051] It can be understood that laser radar emits a laser beam and receives its reflected signal, measures the round-trip time of the laser beam, and then accurately calculates the distance between the laser radar and the ground or the object being measured. Figure 2 , Figure 2Two laser radars are installed on the target object (car body), and the two laser radars are installed diagonally. The laser radar in the figure includes a first laser radar and a second laser radar, the first laser radar is marked as 1#, and the second laser radar is marked as 2#. The scanning ranging range of each laser radar is 270°. In order to achieve all-round obstacle avoidance detection of the target object, the first laser radar and the second laser radar are installed diagonally. The dotted line of the laser radar facing the outside in the figure indicates the scanning direction of the laser beam. The diagonal installation structure can avoid the scanning blind area during the scanning ranging process. The polygonal area composed of N1, N2, N3 and Nm in the figure is the obstacle avoidance detection area. The intrusion detection implemented in the embodiment of the present invention is the detection implemented for obstacles entering this area.

[0052] In practical applications, the implementation of laser radar ranging is affected by many factors, and there may be errors in the acquired scanning ranging data, which in turn affects the accuracy of subsequent detection results. Based on this, in a specific implementation, after obtaining the scanning ranging data, the data is also preprocessed, specifically including:

[0053] When there is a shielded area in the obstacle avoidance detection area, the scanning and ranging data of the laser radar is eliminated according to the shielded area to obtain the scanning and ranging data with data elimination. The shielded area is the area that needs to be ignored when implementing intrusion detection on the target object. For example, if there is a protrusion on the target object, it is easy to misjudge it as an obstacle when implementing intrusion detection. The area where the protrusion is located needs to be set as a shielded area, and the corresponding scanning and ranging data is eliminated based on the shielded area to prevent such data from being misjudged as an obstacle in the subsequent processing process. When there is a shielded area in the obstacle avoidance detection area, the above-mentioned data elimination task needs to be performed; if there is no shielded area, go directly to the next step.

[0054] The scanning distance measurement data that has completed data elimination is cleaned according to the preset distance measurement threshold to obtain the scanning distance measurement data that has completed data cleaning. Data cleaning is to filter out the erroneous data in the scanning distance measurement data. The scanning distance measurement data can be traversed. If the scanning distance measurement data is greater than the distance measurement threshold, it means that the scanning distance measurement data may be caused by other factors. The data is filtered out. Other factors may be signal interference in the distance measurement process, or errors caused by data calculation.

[0055] It should be noted that data cleaning can be deletion or data replacement. By data cleaning, out-of-range values and invalid values in the scan ranging data can be filtered out. Taking data replacement as an example, the ranging threshold is set to include the maximum ranging threshold Dmax and the minimum ranging threshold Dmin. The distance limiting process is carried out through the maximum ranging threshold and the minimum ranging threshold. For example, if r[i] is the scan ranging data of the i-th laser beam, and r[i]>Dmax or r[i]<Dmin, it means that the ranging result of the i-th laser beam exceeds the range Dmax of the lidar or is less than the minimum detection range Dmin. Then r[i] = r[i - 1], and the scan ranging data not in the above interval is replaced with the adjacent scan ranging data. After completing the data cleaning, proceed to the next step.

[0056] The scan ranging data after completing data cleaning is input into a preset filtering processing model to obtain the scan ranging data after completing preprocessing. The filtering processing model can be selected based on actual needs. For example, median filtering or smoothing filtering can be used. To improve the applicability in different scenarios, the embodiment of the present invention uses smoothing filtering processing and implements smoothing through a weighted median filtering algorithm. Still taking r[i] as the scan ranging data of the i-th laser beam, the formula of the weighted median filtering algorithm is:

[0057] r[i] = a1*r[i - 2] + a2*r[i - 1] + a3*r[i] + a4*r[i + 1] + a5*r[i + 2], where a1, a2, a3, a4, and a5 are all weight coefficients, a1 + a2 + a3 + a4 + a5 = 1, r[i - 2] is the scan ranging data collected in the two cycles before the scan ranging data r[i], r[i - 1] is the scan ranging data collected in the cycle before the scan ranging data i, r[i + 1] is the scan ranging data collected in the cycle after the scan ranging data r[i], and r[i + 2] is the scan ranging data collected in the two cycles after the scan ranging data i. Through the above method, the obtained scan ranging data can be made more accurate. After obtaining the obstacle avoidance detection area and the scan ranging data, proceed to step S12.

[0058] S12. Obtain the target data set of the target object according to all the scan ranging data, where the target data set is a set constructed from the omnidirectional point cloud data in the circumferential direction of the target object.

[0059] Specifically, the scan ranging data represents the ranging value of the lidar, and the data in the target data set is the point cloud data distributed in space. Each scan ranging data can be represented as a scan point based on the coordinate position of each lidar. Still taking Figure 2Take two laser radars as an example, each laser radar performs ranging at an interval of 1°, and the interval angle of the acquired scanning ranging data is 1°. Each laser radar can obtain 270 scanning ranging data, and the ranging end position represented by each scanning ranging data is the scanning point. The target data set is obtained based on the point cloud data represented by all scanning points. It should be noted that the above scanning ranging data interval angle is only for illustration. In actual applications, the interval angle between adjacent scanning ranging data can be freely configured as long as it can meet the accuracy requirements of intrusion detection.

[0060] Exemplarily, obtaining a target data set of a target object according to all scanning ranging data includes:

[0061] S12-1. According to the data coordinate system of the scanning distance measurement data of each laser radar and the central coordinate system of the target object, the relative offset position of each data coordinate system and the central coordinate system is obtained. Figure 2 In the figure, the data coordinate system of the 1# laser radar is recorded as lx1-ly1, the data coordinate system of the 2# laser radar is recorded as lx2-ly2, and the center coordinate system of the target object is recorded as XOY. Based on the positional relationship between the three coordinate systems, the relative offset position of each data coordinate system and the center coordinate system can be obtained.

[0062] S12-2. Implement the coordinate system conversion of the scanning ranging data of the corresponding laser radar according to each relative offset position to convert the scanning ranging data of the data coordinate system to the central coordinate system. Implementing data conversion includes coordinate system conversion between the data coordinate system and the central coordinate system and data conversion. The coordinate system conversion is based on the formula lx[i]=r[i]*cos(θ(i)), ly[i]=r[i]*sin(θ(i)); the angle conversion from the data coordinate system to the central coordinate system can be implemented through the above formula, lx[i] is the horizontal coordinate value of r[i] converted to the central coordinate system, and ly[i] is the vertical coordinate value of r[i] converted to the central coordinate system. Data conversion includes data angle conversion and displacement conversion. The specific mathematical formula is: x[i]=R(α)*[lx[i]+lx; ly[i]+ly], where lx, ly are the offsets of the laser radar relative to the center of the vehicle body, and R(α) is the rotation offset α, that is, the relative angle between the laser radar direction and the vehicle head direction. The rotation matrix is ​​R(α)=[cos(α),-sin(α); sin(α),cos(α)]; Based on the above method, the scanning ranging data can be converted to the central coordinate system.

[0063] S12-3, according to the coordinate system conversion results of all the scanning and ranging data, a target data set of the target object is obtained. Each data of the target data set is represented by a corresponding data point of the central coordinate system, that is, a scanning point, and all scanning points are determined as omnidirectional point cloud data of the circumference of the target object.

[0064] It should be noted that all data in the target data set can be stored in sequence based on the timestamps of their collection; or each data in the target data set can be stored in sequence based on the azimuth of the central coordinate system to facilitate subsequent extraction and processing.

[0065] In practical applications, due to the large amount of data in the target data set, the subsequent search of traditional data storage methods is cumbersome. Based on this, in a specific implementation, after obtaining the target data set, all data of the target data set are stored as a KD tree structure, and each data in the KD tree structure is configured with a corresponding index label, and the index label can be the azimuth of the scanning point in the central coordinate system. It can be understood that the main benefit of storing data in the KD tree is that it can significantly improve the efficiency of data retrieval. The KD tree recursively divides the k-dimensional space and converts the data into a simple binary tree structure, reducing the amount of calculation. This structure allows rapid positioning of the subspace that may contain the required data, thereby avoiding unnecessary distance calculations, making data retrieval more efficient, and entering step S13 after obtaining the target data set of the target object.

[0066] S13. Determine whether there are any intruding obstacles in the obstacle avoidance detection area according to the target data set.

[0067] Specifically, the target data set represents the scanning points after the laser radar implements circumferential ranging of the target object, and the obstacle avoidance detection area represents the detection range required for the target object to implement intrusion detection. If the scanning point is located in the obstacle avoidance detection area, it means that there are intruding obstacles in the obstacle avoidance detection area; conversely, if the scanning point is located outside the obstacle avoidance detection area, it means that there are no intruding obstacles in the obstacle avoidance detection area. All scanning points in the target data set can be traversed, and each scanning point can be judged one by one whether it is in the obstacle avoidance detection area through a preset processing model. When the obstacle avoidance detection area is a polygonal structure, the periphery of the obstacle avoidance detection area includes multiple intersections of the polygonal structure. The processing model can sum up all distances based on the distance from the scanning point to each intersection to obtain the distance sum result of the scanning point. If the distance sum result is greater than the set threshold, it means that the scanning point is outside the obstacle avoidance detection area; conversely, when the distance sum result is less than or equal to the set threshold, it is determined that the scanning point is in the obstacle avoidance detection area.

[0068] Exemplarily, the obstacle avoidance detection area is a polygonal area composed of multiple boundary segments; judging whether there is an invading obstacle in the obstacle avoidance detection area according to the target data set includes:

[0069] S13-1. According to the scanning points and obstacle avoidance detection areas of the target data set, the detection angle between each boundary segment of the obstacle avoidance detection area and the scanning point is obtained, wherein the vertex of the detection angle is the scanning point. Figure 2Taking the intrusion detection shown in the figure as an example, the obstacle avoidance detection area is a quadrilateral structure, and the boundary segments include N1N2, N2N3, N3N m 、N1N m The scanning point is set to S, and the scanning point is at any position inside or outside the obstacle avoidance detection area. The detection angles include ∠N1SN2, ∠N2SN3, ∠N3SN m and ∠N1SN m It can be understood that the number of detection angles at each scanning point is the same as the number of boundary segments, and a corresponding number of detection angles can be obtained for different obstacle avoidance detection areas.

[0070] Since intrusion detection needs to maintain a high processing efficiency, traversing all scanning points in the target data set will consume a lot of processing time, which may cause intrusion detection response hysteresis and affect its applicable scenarios. Based on this, in a specific implementation, the detection angle between each boundary segment of the obstacle avoidance detection area and the scanning point is obtained, including:

[0071] Multiple target boundary points are determined on the obstacle avoidance detection area according to a preset interval value, and the interval value can be an interval distance or an interval fraction. When the interval value is the interval distance, the outer boundary of the obstacle avoidance detection area is divided based on the interval distance to obtain multiple target boundary points. When the interval value is the interval fraction, the outer boundary of the obstacle avoidance detection area is divided into multiple line segments based on the interval fraction, and the endpoints of the line segments are the target boundary points.

[0072] After obtaining multiple target boundary points, the data point closest to each target boundary point in the target data set is determined as the scanning point for intrusion detection. The distance calculation of the nearby scanning points can be performed based on the target boundary point, and the scanning point closest to the target boundary point is the scanning point for intrusion detection. Taking all the data in the target data set stored as a KD tree structure as an example, the target boundary points on the polygonal edge obstacle avoidance detection area composed of N1~Nm are traversed in turn: p_polygon_i_t=[p_polygon_x_i_t, p_polygon_y_i_t]=[x_Ni+(x_Ni+1-x_Ni)*t, y_Ni+(y_Ni+1-y_Ni)*t], where i is 1~m-1, and t is a small number with equal spacing between 0 and 1 (such as an interval of 0.02, the boundary segment is cut into 50 points). The nearest neighbor search is implemented through the KD tree to obtain the point cloud coordinates p_nearest(i)=[x_nearest_i; y_nearest_i] and its index idx_nearest_i closest to the boundary point p_polygon_i_t (i is 1~m-1). According to the angle formed by the scanning point of intrusion detection and each boundary segment, the detection angle of the scanning point of intrusion detection is obtained, and the vertex of the detection angle is the scanning point.

[0073] S13-2. Sum all the detection angles of each scanning point to obtain the sum of the angles of each scanning point. The scanning point for intrusion detection is denoted as p_nearest(k). The coordinates of the two corner points Ni~Nj on the outer boundary of the obstacle avoidance detection area are obtained through the configuration file, where i and j are adjacent corner points, and the value range of i and j is 1~m. The rectangular coordinates of the corner points correspond to:

[0074] p_corner(i)=[x_corner_i;y_corner_i];

[0075] p_corner(j)=[x_corner_i;y_corner_j];

[0076] According to the vector angle calculation formula, calculate the vector angle δ_ki composed of the nearest point p_nearest(k) and p_corner(i) and p_corner(j), that is, the detection angle; loop to increase i and j in sequence until i = m, j = 1, and accumulate and calculate to get That is, the sum of the angles at each scanning point.

[0077] S13-3. When the sum of the angles is less than or equal to the preset angle threshold, the current scanning point is determined to be an intrusion point. The angle threshold is set to μ. To improve the accuracy of intrusion detection, the sum of the angles can be represented as γ(k)-2π|. If γ(k)-2π|≤μ, the current scanning point is determined to be in the obstacle avoidance detection area and is determined to be an intrusion point. If the above inequality does not hold, k is incremented and the judgment of the next scanning point is continued. The value range of k is the number of scanning points for intrusion detection.

[0078] In practical applications, if only a single scanning point is used as the basis for intrusion detection, it may cause misjudgment of intrusion detection and affect the accuracy of intrusion detection. Based on this, in a specific implementation, before determining that there is an intrusion obstacle at the location of the intrusion point, the method also includes:

[0079] Perform intrusion detection on the scanning points adjacent to the intrusion point to obtain the intrusion detection results of the adjacent scanning points. Due to the volume of the obstacle, there is usually not only a single scanning point intrusion. To prevent single-point misjudgment, n scanning points can be sampled at two points before and after the idx_nearest_i index coefficient. n is a configuration file parameter, which takes a smaller integer. For example, n=2 is set, and the same intrusion detection as above is implemented to obtain the intrusion detection results of n scanning points. When the intrusion detection result indicates that the adjacent scanning points are also intrusion points, it means that there is no single-point misjudgment, and it is determined that there is an intrusion obstacle at the location of the intrusion point. When it is determined that there is an intrusion obstacle, step S14 is entered.

[0080] S14: When there is an obstacle in the obstacle avoidance detection area, output a detection result indicating the existence of the obstacle.

[0081] Specifically, the detection results can be characterized based on actual needs, which can be alarm signals or intrusion detection log records. For example, when it is determined that there is an obstacle in the obstacle avoidance detection area, the intrusion detection log is output, and the detection log records the current intrusion time, the coordinate data of the intrusion scanning point, and the original data of the complete point cloud of the two lidars.

[0082] It should be noted that the detection results proposed in the embodiments of the present invention can also be characterized accordingly based on the intrusion level. For example, if only a single scanning point exists in the obstacle avoidance detection area, it means that the obstacle is small, or the obstacle partially invades the obstacle avoidance detection area, and a detection result of the first-level intrusion level is output; if 2-5 scanning points exist in the obstacle avoidance detection area, it means that a small obstacle invades the obstacle avoidance detection area, and a detection result of the second-level intrusion level is output; if more than 5 scanning points exist in the obstacle avoidance detection area, it means that a large obstacle invades the obstacle avoidance detection area, and a detection result of the third-level intrusion level is output.

[0083] The following embodiments of the present invention will generally describe the intrusion detection scheme for target objects. Figure 3 , Figure 3 This is a diagram of the overall steps for implementing the intrusion detection method. Specifically, it includes:

[0084] S301, system power-on self-test, mainly through the script program to complete the communication status of each external laser radar software automatic detection.

[0085] S302, loading and configuring the safety area data, that is, configuring the obstacle avoidance detection area of ​​the target object, by loading and reading the configuration safety area and setting the threshold parameters in the file system, so as to configure the intrusion determination range and the related determination threshold parameters.

[0086] S303: Obtain point cloud data collected by multiple laser radars.

[0087] S304, pre-processing the point cloud data collected by the laser radar, including area shielding, data cleaning, filtering processing, etc.

[0088] S305 , coordinate transformation is performed on the preprocessed data to obtain a target data set of the target object.

[0089] S306 , traverse the nearest neighbor search under the polygon node to obtain the scanning points for implementing intrusion detection in the target data set.

[0090] S307: Determine whether the scanning point for intrusion detection is inside or outside the obstacle avoidance detection area.

[0091] S308, intrusion determination and status output, when the scanning point is within the obstacle avoidance detection area, it is determined that there is an obstacle intrusion at the location of the scanning point; when the scanning point is outside the obstacle avoidance detection area, it is determined that there is no obstacle intrusion.

[0092] S309: Output the intrusion log record of the obstacle avoidance detection area.

[0093] Based on the same technical concept as the intrusion detection method, an embodiment of the present invention further provides an obstacle intrusion detection system, which includes a data processing terminal and at least one laser radar.

[0094] At least one laser radar is installed on the target object, and the laser radar is used to obtain scanning ranging data of the target object, wherein the target object is the object to be detected whether there is an intrusion obstacle in the obstacle avoidance detection area. The number of laser radars is set based on the intrusion detection requirements and is not specifically limited here.

[0095] The data processing terminal is connected to the laser radar and is used to obtain the obstacle avoidance detection area and the target data set constructed by obtaining the circumferential all-round point cloud data of the target object based on all the scanning ranging data; determine whether there is an invading obstacle in the obstacle avoidance detection area based on the target data set; when there is an obstacle in the obstacle avoidance detection area, output the detection result indicating the existence of the obstacle. The data processing terminal can be a controller or a computer device that can implement the above data processing.

[0096] See also Figure 4 , Figure 4 The diagram is a schematic diagram of the structure of the intrusion detection system. Specifically, it includes a first laser radar, a second laser radar and a data processing terminal. The first laser radar and the second laser radar are set diagonally, and the ranging scanning range of each laser radar is 270°. The purpose of all-round scanning can be achieved through the diagonal setting. The data processing terminal includes a calculation unit, and the relevant thresholds of the obstacle avoidance detection area and the associated intrusion detection can be written through the configuration file.

[0097] In actual applications, due to the differences in obstacle avoidance detection areas used for different target objects, or the differences in obstacle avoidance detection areas implemented in different scenarios, writing the same obstacle avoidance detection area based on the configuration file may lead to insufficient applicability of the intrusion detection system in different scenarios. Based on this, in a specific implementation, the system also includes an external terminal, which can be a host computer, an alarm, an input device, an I / O module, etc., and can be configured based on actual needs.

[0098] The external terminal is connected to the data processing terminal. The external terminal is used to configure the associated parameters of the data processing terminal for implementing intrusion detection, and / or to issue corresponding alarms or displays according to the detection results, wherein the associated parameters include obstacle avoidance detection areas. Different obstacle avoidance detection areas and related thresholds for associated intrusion detection can be configured through the external terminal to improve the applicability of the intrusion detection system in different scenarios. Intrusion alarms can also be issued through the external terminal, and intrusion detection logs can be displayed.

[0099] The technical solution provided in the embodiments of the present invention has at least the following technical effects or advantages:

[0100] 1. The technical solution of the present invention obtains the obstacle avoidance detection area of ​​the target object and the scanning ranging data of at least one laser radar on the target object. The target object is the object to be detected for detecting whether there is an intruding obstacle in the obstacle avoidance detection area. The target data set of the target object is obtained according to all the scanning ranging data. Since the target data set is a set constructed by the omnidirectional point cloud data around the target object, it can accurately determine whether there is an intruding obstacle in the obstacle avoidance detection area according to the target data set. When there is an obstacle in the obstacle avoidance detection area, a detection result representing the existence of the obstacle is output. The technical solution implements omnidirectional intrusion detection of the target object in the obstacle avoidance detection area through the scanning ranging data collected by the laser radar, and can quickly and accurately determine the obstacle intrusion in the obstacle avoidance detection area. It has high processing efficiency and accuracy for omnidirectional intrusion detection, thereby improving the reliability of omnidirectional intrusion detection of the detection object.

[0101] 2. When applied to intrusion detection of omnidirectional container transfer vehicles, the data point cloud of the vehicle's surrounding environment can be quickly screened to determine whether there is an obstacle intrusion, and to achieve safety warning protection during the vehicle's movement. The above algorithm can realize real-time online operation, with flexible regional configuration and strong robustness. It can also log abnormal intrusions for offline query processing, and can replace the traditional safety lidar-related regional collision avoidance function.

[0102] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0103] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (modules, systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0104] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0105] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0106] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0107] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. An obstacle intrusion detection method, It is characterized in that The method comprises: Obtaining an obstacle avoidance detection area of ​​a target object and scanning ranging data of at least one laser radar on the target object, wherein the target object is a detected object for detecting whether there is an intruding obstacle in the obstacle avoidance detection area; Obtaining a target data set of the target object according to all the scanning and ranging data, wherein the target data set is a set constructed by all-round point cloud data around the target object; Determine whether there is an intruding obstacle in the obstacle avoidance detection area according to the target data set; When the obstacle exists in the obstacle avoidance detection area, a detection result indicating the existence of the obstacle is output.

2. The obstacle intrusion detection method according to claim 1, It is characterized in that Before obtaining the target data set of the target object according to all the scanning ranging data, the method further includes: When a shielding area exists in the obstacle avoidance detection area, data is removed from the scanning ranging data of the laser radar according to the shielding area to obtain scanning ranging data after data removal; Performing data cleaning on the scanning ranging data after data elimination according to a preset ranging threshold value to obtain scanning ranging data after data cleaning; The scanning distance measurement data that has completed data cleaning is input into a preset filtering processing model to obtain the scanning distance measurement data that has completed preprocessing.

3. The obstacle intrusion detection method according to claim 1, It is characterized in that The step of obtaining a target data set of the target object according to all the scanning ranging data includes: According to the data coordinate system where the scanning ranging data of each laser radar is located and the central coordinate system of the target object, a relative offset position of each data coordinate system and the central coordinate system is obtained; Implementing coordinate system conversion of the scanning ranging data corresponding to the laser radar according to each relative offset position, so as to convert the scanning ranging data of the data coordinate system into the central coordinate system; A target data set of the target object is obtained according to the coordinate system conversion results of all the scanning ranging data.

4. The obstacle intrusion detection method according to claim 1, It is characterized in that Before determining whether there is an intruding obstacle in the obstacle avoidance detection area according to the target data set, the method further includes: All data of the target data set are stored as a KD tree structure, and a corresponding index label is configured for each data in the KD tree structure.

5. The obstacle intrusion detection method according to claim 1, It is characterized in that The obstacle avoidance detection area is a polygonal area composed of multiple boundary segments; The determining, according to the target data set, whether there is an intruding obstacle in the obstacle avoidance detection area includes: According to the scanning point of the target data set and the obstacle avoidance detection area, obtaining a detection angle between each boundary segment of the obstacle avoidance detection area and the scanning point, wherein the vertex of the detection angle is the scanning point; Sum all detected angles of each scanning point to obtain a summed angle result of each scanning point; When the sum of the angles is less than or equal to a preset angle threshold, determining the current scanning point as an intrusion point; It is determined that there is an invading obstacle at the location where the intrusion point is located.

6. The obstacle intrusion detection method according to claim 5, It is characterized in that The obtaining, according to the scanning points of the target data set and the obstacle avoidance detection area, a detection angle between each boundary segment of the obstacle avoidance detection area and the scanning point, comprises: Determining a plurality of target boundary points on the obstacle avoidance detection area according to a preset interval value; Determine the closest data point to each target boundary point in the target data set as a scanning point for intrusion detection; The detection angle of the intrusion detection scanning point is obtained according to the angle formed by the intrusion detection scanning point and each boundary segment.

7. The obstacle intrusion detection method according to claim 5, It is characterized in that Before determining that there is an invading obstacle at the location of the intrusion point, the method further includes: Performing intrusion detection on a scanning point adjacent to the intrusion point to obtain an intrusion detection result of the adjacent scanning point; When the intrusion detection result indicates that the adjacent scanning point is also an intrusion point, it is determined that there is an intrusion obstacle at the location where the intrusion point is located.

8. An obstacle intrusion detection system, It is characterized in that The system comprises: At least one laser radar installed on the target object, the laser radar is used to obtain scanning ranging data of the target object, wherein the target object is an object to be detected whether there is an intruding obstacle in the obstacle avoidance detection area; A data processing terminal is connected to the laser radar, and is used to obtain the obstacle avoidance detection area and a target data set constructed by obtaining the circumferential omnidirectional point cloud data of the target object based on all scanning ranging data; determine whether there is an invading obstacle in the obstacle avoidance detection area based on the target data set; and when the obstacle exists in the obstacle avoidance detection area, output a detection result indicating the existence of the obstacle.

9. The obstacle intrusion detection method according to claim 8, It is characterized in that The system further comprises: An external terminal is connected to the data processing terminal, and the external terminal is used to configure the associated parameters of the data processing terminal for implementing intrusion detection, and / or to perform corresponding alarms or displays based on the detection results, wherein the associated parameters include the obstacle avoidance detection area.

10. The obstacle intrusion detection method according to claim 8, It is characterized in that The laser radar includes a first laser radar and a second laser radar, and the first laser radar and the second laser radar are arranged diagonally.

Images