Lidar-based intrusion monitoring method, apparatus, and electronic device

Abstract

This invention discloses an intrusion detection method, device, and electronic device based on lidar, comprising: real-time acquisition of an initial specified number of lidar data frames to establish a scene background; calculation and conversion into a lidar computational background; fusion of the scene background and the computational background to generate an intrusion detection boundary background; calculation of the intrusion saliency coefficient of the intrusion detection boundary background; preprocessing, angle alignment, and differential processing of the real-time acquired lidar data to obtain differentially processed data; binarizing the projection values ​​to obtain a set of all possible intrusion points; performing connected component analysis to statistically analyze the possible candidate intrusion point set to obtain a set of connected components; obtaining the largest connected component in the set of connected components and counting the number of valid points within the connected component, and determining whether an intrusion has occurred based on the number of valid points. This solves the problem of intrusion events caused by time constraints and blind spots resulting from reliance on manual or visual inspections in high-security locations.

Description

Technical Field

[0001] This application relates to the field of intrusion detection technology, and in particular to an intrusion detection method, device and electronic device based on lidar. Background Technology

[0002] Intrusion detection refers to the use of technology to detect and identify unauthorized personnel entering a specific area and to take appropriate measures when intrusion is detected. This monitoring technology is widely used in the security field to protect critical areas from unauthorized access.

[0003] Traditional water conservancy intrusion monitoring relies on manual patrols or visual methods for high-risk locations such as sluice gates, pumping stations, reservoir inlets, spillways, and dangerous water areas. This method has blind spots in complex terrain, at night, or in severe weather, making it easy to miss incidents and unable to monitor them 24 / 7. This can lead to unsuspecting intrusions, resulting in serious consequences and even endangering people's lives. Summary of the Invention

[0004] The purpose of this application is to provide an intrusion detection method, device, and electronic device based on lidar, in order to solve the technical problems in related technologies, such as the timeliness and blind spots caused by reliance on manual patrols or visual inspections in high-security locations, which in turn affect the safety of personnel.

[0005] According to a first aspect of the embodiments of this application, an intrusion detection method based on lidar is provided, comprising:

[0006] Real-time acquisition of an initial specified number of LiDAR frames to establish the scene background;

[0007] The angle range for intrusion detection, the boundary distance of the intrusion, and the boundary restriction distance are set and converted into the calculation background of the lidar through mathematical calculation;

[0008] The scene background and the computational background are fused to generate an intrusion detection boundary background;

[0009] Calculate the projection of the normal vectors of all points in the background of the intrusion detection boundary, and use it as a coefficient of the intrusion salience;

[0010] The real-time acquired lidar data is preprocessed, and the preprocessed data is angularly aligned with the intrusion detection boundary background. Then, the data is differentially processed with the intrusion detection boundary background to obtain the differential data.

[0011] The projection value of the differenced data onto the normal vector is calculated using the coefficients, and the projection value is binarized to obtain the set of all possible intrusion points.

[0012] Noise removal is performed on all possible intrusion point sets;

[0013] After noise removal, perform connected component analysis on possible intrusion points, count the possible candidate intrusion point set, and obtain the connected component set.

[0014] Obtain the largest connected component in the set of connected components, and count the number of valid points within the connected component. If the number of valid points is greater than the threshold for the number of intrusion points, it indicates an intrusion; otherwise, it indicates no intrusion.

[0015] According to a second aspect of the embodiments of this application, an intrusion detection device based on lidar is provided, comprising:

[0016] The scene background creation module is used to collect an initial specified number of frames of LiDAR data in real time to create the scene background;

[0017] The background calculation module is used to set the angle range of intrusion detection, the boundary distance of intrusion, and the boundary restriction distance, and convert them into the calculation background of lidar through mathematical calculation;

[0018] The detection boundary generation module is used to fuse the scene background with the computational background to generate an intrusion detection boundary background;

[0019] The first calculation module is used to calculate the projection of the normal vectors of all points in the background of the intrusion detection boundary, as a coefficient of the salience of the intrusion;

[0020] The second calculation module is used to preprocess the real-time acquired lidar data, align the preprocessed data with the intrusion detection boundary background at an angle, and then perform differential processing on the preprocessed data with the intrusion detection boundary background to obtain the differential data.

[0021] The third calculation module is used to calculate the projection value of the differenced data onto the normal vector using the coefficients, and to perform binarization processing on the projection value to obtain the set of all possible intrusion points.

[0022] A noise removal module is used to remove noise from the set of all possible intrusion points;

[0023] The connected component analysis module is used to perform connected component analysis on possible intrusion points after noise removal, count the possible candidate intrusion point set, and obtain the connected component set.

[0024] The intrusion detection module is used to obtain the largest connected component in the set of connected components and count the number of valid points in the connected component. If the number of valid points is greater than the intrusion point threshold, it indicates an intrusion; otherwise, it indicates no intrusion.

[0025] According to a third aspect of the embodiments of this application, an electronic device is provided, comprising:

[0026] One or more processors;

[0027] Memory, used to store one or more programs;

[0028] When the one or more programs are executed by the one or more processors, the one or more processors perform the method as described in the first aspect.

[0029] According to a fourth aspect of the embodiments of this application, a computer-readable storage medium is provided that stores computer instructions thereon, which, when executed by a processor, implement the steps of the method as described in the first aspect.

[0030] The technical solutions provided by the embodiments of this application may include the following beneficial effects:

[0031] As can be seen from the above embodiments, this application establishes a scene background and calculates the background based on the angle range of intrusion detection, the boundary distance of the intrusion, and the limitation distance of the boundary. The scene background and the calculated background are fused to generate an intrusion detection background. By performing a series of processes such as preprocessing, differential processing, binarization processing, and connected component analysis on the real-time acquired lidar data, the status of whether an intrusion has occurred is finally obtained. This overcomes the technical problems of timeliness and blind spots caused by traditional monitoring that rely on manual inspection or visual inspection, thereby achieving the effect of real-time detection and a large detection range.

[0032] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0033] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0034] Figure 1 This is a flowchart illustrating an intrusion detection method based on lidar according to an exemplary embodiment.

[0035] Figure 2 This is a schematic diagram illustrating a calculation process for an intrusion detection angle greater than or equal to 180° in the horizontal direction, according to an exemplary embodiment.

[0036] Figure 3 This is a schematic diagram illustrating a calculation process for an intrusion detection angle of less than 180° in the horizontal direction, according to an exemplary embodiment.

[0037] Figure 4 This is a schematic diagram illustrating the radial distance calculation process of a lidar according to an exemplary embodiment.

[0038] Figure 5 This is a schematic diagram of the structure of an intrusion detection device based on lidar according to an exemplary embodiment.

[0039] Figure 6 This is a schematic diagram of the structure of an electronic device according to an exemplary embodiment. Detailed Implementation

[0040] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application.

[0041] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used herein are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0042] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0043] Figure 1 This is a flowchart illustrating an intrusion detection method based on lidar according to an exemplary embodiment, such as... Figure 1 As shown, the method may include the following steps:

[0044] S1: Real-time acquisition of an initial specified number of LiDAR frames to establish the scene background;

[0045] Specifically, a specified frame of LiDAR data is continuously collected, and the points at each angle in the horizontal direction of the LiDAR beams at different vertical angles are processed sequentially. If the distance of a point at a given angle is greater than a distance threshold, the distance of the point at that angle is accumulated, and the effective point count of the point at that angle is incremented by 1; otherwise, the point is discarded, and the distance and effective point count of the point at that angle are not accumulated. After the specified frame of LiDAR data is processed, the total distance of the points at each angle is divided by the effective point count of the points at each angle to obtain the average distance of the points at each angle, thereby establishing the scene background.

[0046] The main purpose of this step is to establish the background of the scene scanned by the LiDAR. Multiple frames of data are continuously collected, processed, and averaged to remove errors, making the established scene background more accurate. Generally, the refresh rate of the LiDAR is 10Hz, meaning it scans 10 times per second. The number of consecutively collected frames can be consistent with the LiDAR's refresh rate; 10 frames are sufficient. Since the LiDAR may not detect any objects (infinitely far away) during scanning, resulting in no distance measurement (a distance value of 0), and the LiDAR may be damaged or obstructed (e.g., covered by a shield), the measured distance may be infinitely close, a distance threshold is set. This threshold can be determined based on the actual situation on site. Considering the possibility of side intrusion, the distance threshold can be set to 0.1-0.5 meters.

[0047] S2: Set the angle range for intrusion detection, the boundary distance of the intrusion, and the boundary restriction distance, and convert them into the calculation background for lidar through mathematical calculations; this step includes the following sub-steps:

[0048] S21: Set the angle range for intrusion detection, the boundary distance of the intrusion, and the boundary restriction distance;

[0049] The main purpose of this step is to provide data for the subsequent calculation of the horizontal distance by the lidar.

[0050] S22: Starting from the initial angle of intrusion detection, scan sequentially according to the horizontal angle resolution. Combining the boundary limitation distance of the lidar and the boundary distance of the lidar intrusion, the radial distance of the lidar is projected in the horizontal direction according to the trigonometric function theorem.

[0051] Specifically, Figure 2 The diagram shows the calculation process for an intrusion detection angle greater than or equal to 180° in the horizontal direction. Figure 3 The diagram shows the calculation process for intrusion detection angles less than 180° in the horizontal direction. Figure 2 , Figure 31 indicates the distance limit of the boundary; data outside the boundary is not processed. 2 indicates the boundary distance of the intrusion; data within the boundary indicates an intrusion, and data outside the boundary indicates no intrusion. 3 indicates the starting and ending angle range of the intrusion monitoring. 4 is used to divide the calculation range. Based on the angle with the boundary, two regions are formed. The small angle region is constrained by the distance limit of the boundary, and the large angle region is constrained by the boundary distance of the intrusion. Figure 2 5, 6, 10 and Figure 3 In the equations 5 and 6, the radial distance of the lidar during horizontal scanning is projected horizontally. The calculation is performed by progressively increasing the distance in resolution units, with the example illustrating each process. 7 represents the angle relative to the starting angle of intrusion detection during the scanning process; 8 represents the angle relative to the ending angle of intrusion detection during the scanning process; and 9 represents the angle range of the intrusion detection. Starting from the starting angle of intrusion detection, the horizontal angle is calculated progressively according to the resolution, with distances 5, 6, and 10 calculated. If the angle range of the intrusion detection is greater than 180°, it is calculated in 3 segments; if the angle range is not greater than 180°, it is calculated in 2 segments. The horizontal angle can be obtained by progressively increasing the resolution according to the angle range of the intrusion detection. Combining the distance of the lidar to the boundary limit line and the distance of the lidar to the intrusion boundary line, the lengths of 5, 6, or 10, i.e., the projected distance of the lidar's radial distance in the horizontal direction, can be calculated using trigonometric functions.

[0052] S23: The angle in the vertical direction can be obtained by gradually increasing the vertical beam resolution based on the angle of the initial beam. Combined with the radial distance of the laser radar projected in the horizontal direction, the radial distance of the laser radar, i.e. the distance between the laser radar and the target object, can be calculated using trigonometric functions.

[0053] Specifically, such as Figure 4 The diagram shows the radial distance calculation process of a lidar. 1 represents the projected radial distance of the lidar in the horizontal direction, 2 represents the angle between the current beam and the horizontal direction, and 3 represents the radial distance of the lidar. The angle in the vertical direction is obtained by progressively increasing the vertical beam resolution based on the initial beam angle. Combining the projected radial distance of the lidar in the horizontal direction, the radial distance of the lidar can be calculated using trigonometric functions.

[0054] S24: Starting from the first beam, repeat S22-S23 to obtain the radial distance of the beams in different vertical directions from the lidar at each angle in the horizontal direction.

[0055] The main purpose of this step is to calculate the radial distance of the laser radar at each point in the horizontal direction for the laser radar beams in different vertical directions at each angle.

[0056] S25: Calculate the angle corresponding to the lidar data based on the vertical beam resolution and the horizontal angle resolution. The position of each angle can be mapped to a two-dimensional array. The horizontal coordinate of the two-dimensional array is the index of the vertical beam resolution, and the vertical coordinate is the index of the horizontal angle resolution. Assign the radial distance of the lidar to the position of the two-dimensional array corresponding to the lidar angle to obtain the calculation background of the lidar.

[0057] Specifically, for ease of calculation, the angle corresponding to the lidar data can be calculated based on the vertical beam resolution and the horizontal angle resolution. The position of each angle can be mapped to a two-dimensional array. The horizontal coordinate of the two-dimensional array is the index of the vertical beam resolution, and the vertical coordinate is the index of the horizontal angle resolution. The radial distance of the lidar is assigned to the position of the two-dimensional array corresponding to the lidar angle, thereby obtaining the calculation background of the lidar.

[0058] The main purpose of this step is to obtain the computational background, providing a data foundation for generating the intrusion detection boundary background later.

[0059] S3: Fuse the scene background with the computational background to generate an intrusion detection boundary background;

[0060] Specifically, the distances between each corresponding point in the scene background and the calculated background are compared. If the distance in the scene background is greater than the distance in the calculated background, the distance between the corresponding points in the intrusion detection boundary background is set to the distance between the corresponding points in the calculated background. Otherwise, if the distance in the scene background is 0, the distance between the corresponding points in the intrusion detection boundary background is set to the distance between the corresponding points in the calculated background. Otherwise, if the distance in the scene background is not greater than the distance in the calculated background, the distance between the corresponding points in the intrusion detection boundary background is set to the distance between the corresponding points in the scene background. This completes the generation of the intrusion detection boundary background.

[0061] The main purpose of this step is to generate an intrusion detection boundary background. The scene background is the actual background, while the calculated background is calculated based on the intrusion detection angle range, the intrusion boundary distance, and the boundary restriction distance; it is essentially a virtual background. Since other objects, such as stone blocks, doors, and trees, may exist within the background scanned by the LiDAR, directly using the calculated background would treat these existing objects as intruders, leading to false positives. If the scene background is used directly to detect intruders, any changes to the intruder need to be detected, and without the constraint of the intrusion boundary for comparison, each point's data needs to be judged individually, thus increasing the detection burden. By fusing the scene background and the calculated background, they can complement each other, focusing only on whether the intruder exceeds the intrusion detection boundary; events outside the boundary do not need to be detected.

[0062] S4: Calculate the projection of the normal vectors of all points in the background of the intrusion detection boundary, as a coefficient of the intrusion saliency; this step includes the following sub-steps:

[0063] S41: Calculate the vector coordinates of each point in the intrusion detection boundary background based on trigonometric functions and the distance to each point in the intrusion detection boundary background;

[0064] Specifically, given the vertical and horizontal angles of each point on the lidar, the vector coordinates of each point on the intrusion detection boundary background can be calculated using trigonometric functions and the distance to each point on the intrusion detection boundary background.

[0065] The main purpose of this step is to calculate the vector coordinates of each point in the intrusion detection boundary background, providing data support for subsequent calculations.

[0066] S42: For each data point from the LiDAR, search for valid points around each point according to a predetermined grid. If the distance between a point and its surrounding valid points is less than the mutation threshold, record the position of the valid point and increment the number of valid points by 1; otherwise, do not record the position of the point. After searching each grid, count the number of valid points in the same row or column. If the number of valid points in the same row and column is greater than 0, then the number of valid points is the number of valid points found; otherwise, the number of valid points is 0.

[0067] Specifically, for each point of the lidar, valid points around each point are searched according to a predetermined grid. If the distance between the point and the surrounding valid points is less than the mutation threshold, the position of the valid point is recorded and the number of valid points is incremented by 1; otherwise, the position of the point is not recorded. After each grid search is completed, the number of valid points in the same row or column is counted. If the number of valid points in the same row and column is greater than 0, the number of valid points is the number of valid points found; otherwise, the number of valid points is 0.

[0068] The main purpose of this step is to calculate the effective number of points on the surface formed by each data point and its neighboring points, providing a data foundation for calculating the normal vector of the surface later. By finding the normal vector of the surface formed by each data point and its neighboring points, the vertical resolution is higher than the horizontal resolution. To maintain consistency, the number of neighboring points is determined according to the vertical resolution: 2 points in the vertical direction and the number in the horizontal direction is the vertical resolution divided by the horizontal resolution. Taking a 16-line LiDAR as an example, with a vertical resolution of 2° and a horizontal resolution of 0.18, the grid size is 2*10. Theoretically, the distance between adjacent points should not change significantly. Excessive change indicates anomalies or discontinuities. The length per degree is calculated using the formula 2πr / 360, where r is the boundary distance of the intrusion. The distance between two adjacent points can be obtained according to the vertical resolution; this distance is the mutation threshold.

[0069] S43: If the number of valid points is greater than the number threshold, and all valid points are neither in the same row nor in the same column, save the vector coordinates of the valid points in sequence, establish a vector set, calculate the covariance matrix of the vector set, and perform eigenvalue decomposition on the covariance matrix to obtain eigenvectors and eigenvalues. Obtain the vector value corresponding to the smallest eigenvalue, and perform a dot product operation on the vector value and the vector value of the corresponding point to obtain the projection of the direction line of the corresponding point of the intrusion detection boundary background onto the normal vector, that is, obtain the coefficient of the intrusion significance.

[0070] The main purpose of this step is to calculate the projection of the direction line of the corresponding point of the intrusion detection boundary background onto the normal vector in order to remove the influence of small ground objects such as ground protrusions, mice, cats, etc.

[0071] S5: Preprocess the real-time acquired lidar data, align the preprocessed data with the intrusion detection boundary background at an angle, and then perform differential processing on the data with the intrusion detection boundary background to obtain the differential data;

[0072] Specifically, real-time acquisition of LiDAR data is performed. After completing a full scanning cycle, each laser beam has completed a horizontal full-angle scan. Different LiDAR specifications may have different angles; some may be 360°, others 120°, or other angles. The index of each point is calculated based on the vertical and horizontal scanning angle resolutions. The data is then arranged sequentially in a two-dimensional array, retaining only the data from each LiDAR beam within the scanning angle range. This aligns the data with the intrusion detection boundary background. The pre-processed data is then differentially processed with the intrusion detection boundary background to obtain the differential data. During the differential process, the pre-processed data is subtracted from the intrusion detection boundary background; points greater than 0 indicate intrusion, and points not greater than 0 indicate no intrusion.

[0073] The main purpose of this step is to obtain data after differentiating it from the background of the intrusion detection boundary, so as to provide a data basis for subsequent determination of whether an intrusion has occurred.

[0074] S6: Calculate the projection value of the differenced data onto the normal vector using the coefficients, and binarize the projection value to obtain the set of all possible intrusion points;

[0075] Specifically, the differentiated data is multiplied by the coefficients at corresponding positions to obtain the projected value. The projected value is then binarized: if the projected value is greater than or equal to the intrusion point threshold, it is set to 255; otherwise, it is set to 0, thus obtaining the set of all possible intrusion points. Since the lidar itself has measurement errors, and external factors such as the movement of leaves in the wind may occur during the measurement process, an intrusion point threshold is introduced to eliminate these external and measurement errors. Therefore, the intrusion point threshold can be determined according to the lidar's own ranging resolution.

[0076] The main purpose of this step is to identify potential intrusion points, providing a data foundation for subsequent calculations of intruders.

[0077] S7: Perform noise removal on all possible intrusion point sets;

[0078] Specifically, noise removal is performed using the opening operation of a morphological noise reduction algorithm. The main purpose of this step is to remove isolated and discontinuous points, thus eliminating outliers for subsequent connected component analysis.

[0079] S8: Perform connected component analysis on the possible intrusion points after noise removal, count the possible candidate intrusion point set, and obtain the connected component set;

[0080] Specifically, after noise removal, connected component analysis is performed on potential intrusion points to statistically analyze the set of possible candidate intrusion points, thus obtaining a set of connected components. The main purpose of this step is to obtain the set of connected components, providing a data foundation for subsequent calculations on whether an intrusion has occurred.

[0081] S9: Obtain the largest connected component in the set of connected components, and count the number of valid points in the connected component. If the number of valid points is greater than the threshold of the number of intrusion points, it indicates an intrusion; otherwise, it indicates no intrusion.

[0082] Specifically, the connected component set is sorted, and the largest point in the connected component set is obtained. The number of valid points within the connected component is then counted.

[0083] Formula (1): c = 2πr / 360;

[0084] The length c of the laser radar reaching the intrusion boundary in one degree is calculated according to formula (1), where r is the distance of the laser radar reaching the intrusion boundary. Assuming a normal person's height is 1.7 meters (ignoring the influence of the head) and width is 0.4 meters, it is equivalent to a rectangle. The number of points in the vertical direction is calculated as nc1=1.7 / c, and the number of points after conversion according to the vertical resolution of the laser radar is nc=nc1 / vertical resolution. Similarly, the number of points in the horizontal direction is calculated as nr1=0.4 / c, and the number of points after conversion according to the horizontal resolution of the laser radar is nr=nr1 / horizontal resolution. Then the number of points for an adult at the intrusion boundary distance is ny=nr*nc. Considering that the body shape of children is different from that of adults and the error caused by the head, the coefficient a can be appropriately reduced. a can be 0.5 or even lower to obtain the effective number of points threshold ny*a. The above calculation is based on land. The calculation method for the number of intruders in water is similar. Considering that the exposed area of ​​intruders in water is smaller than that on land, the coefficient can be appropriately reduced.

[0085] If the number of valid points is greater than the threshold for the number of intrusion points, it indicates an intrusion; otherwise, it indicates no intrusion.

[0086] As can be seen from the above embodiments, this application establishes a scene background and calculates the background based on the angle range of intrusion detection, the boundary distance of the intrusion, and the limitation distance of the boundary. The scene background and the calculated background are fused to generate an intrusion detection background. By performing a series of processes such as preprocessing, differential processing, binarization processing, and connected component analysis on the real-time acquired lidar data, the status of whether an intrusion has occurred is finally obtained. This overcomes the technical problems of timeliness and blind spots caused by traditional monitoring that rely on manual inspection or visual inspection, thereby achieving the effect of real-time detection and a large detection range.

[0087] Corresponding to the aforementioned embodiments of the lidar-based intrusion detection method, this application also provides embodiments of a lidar-based intrusion detection device.

[0088] Figure 5This is a block diagram illustrating an intrusion detection device based on lidar, according to an exemplary embodiment. (Refer to...) Figure 5 The device includes:

[0089] Scene background creation module 1 is used to collect an initial specified number of frames of LiDAR data in real time to create a scene background;

[0090] The calculation background establishment module 2 is used to set the angle range of intrusion detection, the boundary distance of intrusion, and the boundary restriction distance, and convert them into the calculation background of lidar through mathematical calculation.

[0091] The detection boundary generation module 3 is used to fuse the scene background with the computational background to generate an intrusion detection boundary background;

[0092] The first calculation module 4 is used to calculate the projection of the normal vectors of all points in the background of the intrusion detection boundary, as a coefficient of the salience of the intrusion.

[0093] The second calculation module 5 is used to preprocess the real-time acquired lidar data, align the preprocessed data with the intrusion detection boundary background at an angle, and then perform differential processing on the data with the intrusion detection boundary background to obtain the differential data.

[0094] The third calculation module 6 is used to calculate the projection value of the differenced data on the normal vector using the coefficients, and to perform binarization processing on the projection value to obtain the set of all possible intrusion points.

[0095] Noise removal module 7 is used to remove noise from the set of all possible intrusion points;

[0096] The connected component analysis module 8 is used to perform connected component analysis on possible intrusion points after noise removal, count the possible candidate intrusion point set, and obtain the connected component set.

[0097] Intrusion detection module 9 is used to obtain the largest connected component in the set of connected components and count the number of valid points in the connected component. If the number of valid points is greater than the intrusion point threshold, it indicates an intrusion; otherwise, it indicates no intrusion.

[0098] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0099] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this application according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0100] Accordingly, this application also provides an electronic device, including: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the above-described lidar-based intrusion detection method. Figure 6 The diagram shown is a hardware structure diagram of any data processing-capable device, including a lidar-based intrusion detection device provided in an embodiment of the present invention. Except for... Figure 6 In addition to the processor, memory, DMA controller, disk, and non-volatile memory shown, any data processing device in the embodiment may also include other hardware depending on the actual function of the data processing device, which will not be described in detail here.

[0101] Accordingly, this application also provides a computer-readable storage medium storing computer instructions thereon, which, when executed by a processor, implement the aforementioned lidar-based intrusion detection method. The computer-readable storage medium can be an internal storage unit of any data-processing device as described in any of the foregoing embodiments, such as a hard disk or memory. The computer-readable storage medium can also be an external storage device, such as a plug-in hard disk, smart media card (SMC), SD card, flash card, etc., equipped on the device. Furthermore, the computer-readable storage medium can include both internal storage units of any data-processing device and external storage devices. The computer-readable storage medium is used to store the computer program and other programs and data required by the data-processing device, and can also be used to temporarily store data that has been output or will be output.

[0102] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only.

[0103] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. An intrusion detection method based on lidar, characterized in that, include: Real-time acquisition of an initial specified number of LiDAR frames to establish the scene background; The angle range for intrusion detection, the boundary distance of the intrusion, and the boundary restriction distance are set and converted into the calculation background of the lidar through mathematical calculation; The scene background and the computational background are fused to generate an intrusion detection boundary background; The projection of the direction line of the corresponding point of the intrusion detection boundary background onto the normal vector is calculated, which yields the coefficient of the intrusion salience. The real-time acquired lidar data is preprocessed, and the preprocessed data is angularly aligned with the intrusion detection boundary background. Then, the data is differentially processed with the intrusion detection boundary background to obtain the differential data. The projection value of the differenced data onto the normal vector is calculated using the coefficients, and the projection value is binarized to obtain the set of all possible intrusion points. Noise removal is performed on all possible intrusion point sets; After noise removal, perform connected component analysis on possible intrusion points, count the possible candidate intrusion point set, and obtain the connected component set. Obtain the largest connected component in the set of connected components, and count the number of valid points in the connected component. If the number of valid points is greater than the threshold of intrusion points, it indicates an intrusion; otherwise, it indicates no intrusion. This involves defining the angle range for intrusion detection, the boundary distance of the intrusion, and the boundary restriction distance, which are then mathematically converted into the calculation background for lidar, including: S21: Set the angle range for intrusion detection, the boundary distance of the intrusion, and the boundary restriction distance; S22: Starting from the initial angle of intrusion detection, the scanning proceeds sequentially according to the resolution of the horizontal angle. Combining the boundary limitation distance of the lidar and the boundary distance of the lidar intrusion, two regions are formed based on the angle between the scan line and the boundary. The small angle region is calculated using the boundary limitation distance, and the large angle region is calculated using the boundary distance of the intrusion. The radial distance of the lidar is projected in the horizontal direction according to the trigonometric function theorem. S23: The angle in the vertical direction can be obtained by gradually increasing the resolution of the starting beam angle in the vertical direction. Combined with the radial distance of the laser radar projected in the horizontal direction, the radial distance of the horizontal projection distance in the laser radar can be calculated using trigonometric functions. That is, the distance between the laser radar and the target object. S24: Starting from the first beam, repeat S22-S23 to obtain the radial distance of the beams in different vertical directions from the laser radar at each angle in the horizontal direction. S25: Calculate the angle corresponding to the lidar data based on the vertical beam resolution and the horizontal angle resolution. The position of each angle can be mapped to a two-dimensional array. The horizontal coordinate of the two-dimensional array is the index of the vertical beam resolution, and the vertical coordinate is the index of the horizontal angle resolution. Assign the radial distance of the lidar to the position of the two-dimensional array corresponding to the lidar angle to obtain the calculation background of the lidar.

2. The method according to claim 1, characterized in that, Real-time acquisition of an initial specified number of LiDAR frames to establish the scene background, including: Continuously collect specified frames of LiDAR data, and process the points at each angle in the horizontal direction for the LiDAR beams at different vertical angles in sequence. If the distance of a point at a given angle is greater than a distance threshold, the distance of the point at that angle is accumulated, and the effective point count of the point at that angle is incremented by 1; otherwise, the point is discarded, and the distance and effective point count of the point at that angle are not accumulated. After the specified frame of LiDAR data is processed, the total distance of the points at each angle is divided by the effective point count of the points at each angle to obtain the average distance of the points at each angle, thereby establishing the scene background.

3. The method according to claim 1, characterized in that, The scene background and the computational background are fused to generate an intrusion detection boundary background, including: The distances between corresponding points in the scene background and the calculated background are compared. If the distance in the scene background is greater than the distance in the calculated background, the distance between the corresponding points in the intrusion detection boundary background is set to the distance between the corresponding points in the calculated background. Otherwise, if the distance in the scene background is 0, the distance between the corresponding points in the intrusion detection boundary background is set to the distance between the corresponding points in the calculated background. Otherwise, if the distance in the scene background is not greater than the distance in the calculated background, the distance between the corresponding points in the intrusion detection boundary background is set to the distance between the corresponding points in the scene background. This completes the generation of the intrusion detection boundary background.

4. The method according to claim 1, characterized in that, Calculate the projection of the direction line of the corresponding point of the intrusion detection boundary background onto the normal vector, which yields the coefficient of intrusion salience, including: Based on trigonometric functions and the distance to each point in the intrusion detection boundary background, the vector value of the vector coordinates of each point in the intrusion detection boundary background is calculated. For each data point from the LiDAR, search for valid points around each point according to a predetermined grid. If the distance between a point and its surrounding valid points is less than a threshold, record the position of the valid point and increment the number of valid points by 1; otherwise, do not record the position of the point. After searching each grid, count the number of valid points in the same row or column. If the number of valid points in the same row and column is greater than 0, then the number of valid points is the number of valid points found; otherwise, the number of valid points is 0. If the number of valid points is greater than the threshold, and all valid points are neither in the same row nor in the same column, the vector coordinates of the valid points are saved sequentially, a vector set is established, the covariance matrix of the vector set is calculated, and the covariance matrix is ​​decomposed into eigenvalues ​​to obtain eigenvectors and eigenvalues. The vector value corresponding to the smallest eigenvalue is obtained, and the vector value corresponding to the smallest eigenvalue and the vector coordinate of the corresponding point are multiplied by a dot product to obtain the projection of the direction line of the corresponding point of the intrusion detection boundary background onto the normal vector, which is the coefficient of the intrusion saliency.

5. The method according to claim 1, characterized in that, The real-time acquired lidar data is preprocessed, and the preprocessed data is angularly aligned with the intrusion detection boundary background. Then, differential processing is performed between the preprocessed data and the intrusion detection boundary background to obtain the differential data, including: Real-time acquisition of LiDAR data, after completing a full scanning cycle, is performed sequentially, retaining only the data for each line count of the LiDAR within the scanning angle range, thereby aligning the data with the intrusion detection boundary background at an angle. The pre-processed data is then differentially processed with the intrusion detection boundary background to obtain the differential data.

6. The method according to claim 1, characterized in that, The projection values ​​of the differencing data onto the normal vector are calculated using the coefficients, and the projection values ​​are binarized to obtain a set of all possible intrusion points, including: The differenced data is multiplied by the coefficients according to their corresponding positions to obtain the projected value. The projected value is then binarized. If the projected value is greater than or equal to the intrusion point threshold, it is set to 255; otherwise, it is set to 0, thus obtaining the set of all possible intrusion points.

7. An intrusion detection device based on lidar, characterized in that, include: The scene background creation module is used to collect an initial specified number of frames of LiDAR data in real time to create the scene background; The background calculation module is used to set the angle range of intrusion detection, the boundary distance of intrusion, and the boundary restriction distance, and convert them into the calculation background of lidar through mathematical calculation; The detection boundary generation module is used to fuse the scene background with the computational background to generate an intrusion detection boundary background; The first calculation module is used to calculate the projection of the direction line of the corresponding point of the intrusion detection boundary background onto the normal vector, that is, to obtain the coefficient of the intrusion salience. The second calculation module is used to preprocess the real-time acquired lidar data, align the preprocessed data with the intrusion detection boundary background at an angle, and then perform differential processing on the preprocessed data with the intrusion detection boundary background to obtain the differential data. The third calculation module is used to calculate the projection value of the differenced data onto the normal vector using the coefficients, and to perform binarization processing on the projection value to obtain the set of all possible intrusion points. A noise removal module is used to remove noise from the set of all possible intrusion points; The connected component analysis module is used to perform connected component analysis on possible intrusion points after noise removal, count the possible candidate intrusion point set, and obtain the connected component set. The intrusion detection module is used to obtain the largest connected component in the set of connected components and count the number of valid points in the connected component. If the number of valid points is greater than the intrusion point threshold, it indicates an intrusion; otherwise, it indicates no intrusion. This involves defining the angle range for intrusion detection, the boundary distance of the intrusion, and the boundary restriction distance, which are then mathematically converted into the calculation background for lidar, including: S21: Set the angle range for intrusion detection, the boundary distance of the intrusion, and the boundary restriction distance; S22: Starting from the initial angle of intrusion detection, the scanning proceeds sequentially according to the resolution of the horizontal angle. Combining the boundary limitation distance of the lidar and the boundary distance of the lidar intrusion, two regions are formed based on the angle between the scan line and the boundary. The small angle region is calculated using the boundary limitation distance, and the large angle region is calculated using the boundary distance of the intrusion. The radial distance of the lidar is projected in the horizontal direction according to the trigonometric function theorem. S23: The angle in the vertical direction can be obtained by gradually increasing the resolution of the starting beam angle in the vertical direction. Combined with the radial distance of the laser radar projected in the horizontal direction, the radial distance of the horizontal projection distance in the laser radar can be calculated using trigonometric functions. That is, the distance between the laser radar and the target object. S24: Starting from the first beam, repeat S22-S23 to obtain the radial distance of the beams in different vertical directions from the laser radar at each angle in the horizontal direction. S25: Calculate the angle corresponding to the lidar data based on the vertical beam resolution and the horizontal angle resolution. The position of each angle can be mapped to a two-dimensional array. The horizontal coordinate of the two-dimensional array is the index of the vertical beam resolution, and the vertical coordinate is the index of the horizontal angle resolution. Assign the radial distance of the lidar to the position of the two-dimensional array corresponding to the lidar angle to obtain the calculation background of the lidar.

8. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-6.

9. A computer-readable storage medium storing computer instructions thereon, characterized in that, When executed by the processor, this instruction implements the steps of the method as described in any one of claims 1-6.

Images