Laser radar-based human intrusion detection method and security system

Abstract

The present application relates to a kind of human intrusion detection methods based on laser radar, whether the judgment of detection object is consistent with human characteristics from the dimension of human leg feature, simultaneously whether the judgment of detection object is consistent with human characteristics from the dimension of human gait feature, the detection result of human leg feature and human gait feature is judged whether the detection object is human, more specifically to the detection of human intrusion, can detect the human intrusion existing in the security range, avoid the false alarm caused by the moving object except human into the security detection range, to improve the alarm accuracy of security system.Therefore, the present application uses laser radar to collect the information of detection object, is not interfered by environmental factors such as ambient light, robust, and compared with camera monitoring, it has privacy security.

Description

Technical Field

[0001] The present invention relates to the technical field of human intrusion detection, and in particular to a human intrusion detection method and security system based on laser radar. Background Art

[0002] Intelligent security systems are an effective means of ensuring environmental safety, particularly in environments such as homes and private warehouses where they are crucial for preventing intrusion. There are currently two common approaches to detecting moving objects in an environment: one uses a digital camera to capture the surroundings, which is then uploaded to the cloud for image analysis to identify the intruder; the other uses lidar to detect the location of the intruder in the environment, and then analyzes this location to determine whether it is a moving object.

[0003] However, in the first solution, since the recognition rate of digital cameras is extremely dependent on ambient light, excessive or weak ambient light will make the captured image full of noise, easily losing key pixels, and causing image analysis to be unable to correctly identify the content of the image. Therefore, the robustness of this security system is poor; and when the captured data is uploaded to the cloud, there is a risk of data being stolen by others, which is not conducive to privacy security.

[0004] The second solution can only detect whether the intruder in the environment is a moving object, but cannot detect whether the moving object is a human. Therefore, when there are moving objects other than humans in the detection environment, such as stray cats and dogs, birds and insects, the security system will generate unnecessary false alarms. Summary of the Invention

[0005] The purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art and to provide a human intrusion detection method and security system based on lidar, which can identify moving humans within the detection range and can be adapted to detection in various environments with good privacy and security.

[0006] The present invention is achieved through the following technical solutions: a human intrusion detection method based on laser radar, comprising the steps of: obtaining the distance ρ and the orientation α of each collection point where the laser beam of the laser radar falls on the detection object relative to the laser emission point, to obtain a collection point set; calculating the similarity between the contour formed by the collection points in the collection point set and the human leg contour model, and when the similarity is greater than a similarity threshold E, determining that the collection point set is a suspicious feature point set; obtaining a feature vertex from the suspicious feature point set of a scanning cycle, obtaining the absolute distance dis between the feature vertex and each collection point in the suspicious feature point set of the previous scanning cycle, and if the suspicious feature point set of the previous scanning cycle does not contain a collection point X whose absolute distance dis is less than the suspicious threshold β, i, then mark the feature vertex as a new suspicious human leg; if the suspicious feature point set of the previous scanning cycle has a collection point X whose absolute distance dis is less than the suspicious threshold β i , the feature vertex is marked as the suspicious human leg that is the same as the feature vertex in the previous scanning cycle; wherein the feature vertex is the collection point with the smallest distance ρ in the collection point set; suspicious objects are constructed, each of the suspicious objects corresponds to a first feature sequence and a second feature sequence, and the first feature sequence and the second feature sequence each contain multiple feature vertices marked as the same suspicious human leg; suspicious gait features are calculated based on the first feature sequence and the second feature sequence of the same suspicious object, and whether the detected object is human is determined based on the comparison result of the suspicious gait features and the human gait features.

[0007] Compared to the existing technology, the present invention provides a human intrusion detection method based on laser radar. It determines whether the detection object meets the human characteristics from the dimension of human leg characteristics, and also determines whether the detection object meets the human characteristics from the dimension of human gait characteristics. The detection results of human leg characteristics and human gait characteristics are combined to determine whether the detection object is human. This method can detect human intrusion more specifically, detect the presence of human intrusion within the security range, and avoid false alarms caused by moving objects other than humans entering the security detection range, thereby improving the alarm accuracy of the security system. In addition, the present invention uses laser radar to collect information about the detection object, which is not affected by environmental factors such as ambient light, has strong robustness, and is more private and secure than video surveillance.

[0008] Furthermore, before calculating the similarity between the contour formed by the points in the collection point set and the human leg contour model, the method further includes the following steps: filtering out abnormally positioned points in the collection point set using a sliding window clustering method; obtaining the contour formed by the collection point set, and moving the contour in a polar coordinate system so that the characteristic vertices of the collection point set are aligned with the model vertices of the human leg contour model and have the same polar angle. This solution eliminates abnormally positioned collection points, which can improve the accuracy of similarity calculation.

[0009] Furthermore, when constructing a suspicious object, the steps include: obtaining a new feature vertex and adding it to the end of the first feature sequence or the second feature sequence where the feature vertex marked as the same suspicious human leg as the new feature vertex is located; if there is no feature vertex marked as the same suspicious human leg as the new feature vertex, setting the new feature vertex as the waiting feature point; within a waiting time period, when the waiting feature point marked as a different suspicious human leg appears, constructing a new first feature sequence with the waiting feature point marked as the same suspicious human leg, constructing a new second feature sequence with the waiting feature point marked as another same suspicious human leg, and constructing a new suspicious object with the new first feature sequence and the new second feature sequence.

[0010] Furthermore, the suspicious step features include a first step length sequence, a second step length sequence and a step frequency sequence, and the human standard step features include a human standard step length and a human standard step frequency. When the first step length errors between N of the M first step lengths in the first step length sequence and the human standard step length are less than a step length error threshold, and the second step length errors between N of the M second step lengths in the second step length sequence and the human standard step are less than a step length error threshold; and the step frequency errors between N of the M step frequencies in the step frequency sequence and the human standard step frequency are less than a step frequency error threshold, then the detected object is human.

[0011] Furthermore, the method includes the steps of obtaining the temperature of the detected object and determining whether the detected object is human based on a comparison of the detected object's temperature with human body temperature characteristics. Simultaneously, the method integrates the detection results of human leg characteristics, human gait characteristics, and human body temperature characteristics to determine whether the detected object is human, further improving the alarm accuracy of the security system.

[0012] Based on the same inventive concept, the present invention also provides a human intrusion security system based on laser radar, comprising: a data acquisition and preprocessing unit, a human leg detection unit and a human step detection unit, wherein the data acquisition and preprocessing unit includes a laser radar, which is used to collect the distance ρ and orientation α of each collection point on the detection object relative to the laser emission point to obtain a collection point set; the human leg detection unit includes a contour comparison module, which is used to calculate the similarity between the contour formed by the collection points in the collection point set and the human leg contour model, and when the similarity is greater than the similarity threshold E, the collection point set is determined to be a suspicious feature point set; the human step detection unit includes a suspicious feature point tracking module, a suspicious object construction module and a human step feature comparison module, the suspicious feature point tracking module is used to obtain a feature vertex from the suspicious feature point set of a scanning cycle, and obtain the absolute distance dis between the feature vertex and each collection point in the suspicious feature point set of the previous scanning cycle. If there is no collection point X in the suspicious feature point set of the previous scanning cycle whose absolute distance dis is less than the suspicious threshold β, i , then mark the feature vertex as a new suspicious human leg; if the suspicious feature point set of the previous scanning cycle has a collection point X whose absolute distance dis is less than the suspicious threshold β i , the feature vertex is marked as the suspicious human leg that is the same as the feature vertex in the previous scanning cycle; wherein the feature vertex is the collection point with the smallest distance ρ in the collection point set; the suspicious object construction module is used to construct suspicious objects, each suspicious object corresponds to a first feature sequence and a second feature sequence, and the first feature sequence and the second feature sequence each contain multiple feature vertices marked as the same suspicious human leg; the human gait feature comparison module is used to calculate the suspicious gait feature based on the first feature sequence and the second feature sequence of the same suspicious object, and determine whether the detected object is human based on the comparison result of the suspicious gait feature and the human gait feature.

[0013] Furthermore, the data acquisition and preprocessing unit also includes: a data clustering module, which is used to screen out acquisition points with abnormal positions in the acquisition point set through a sliding window clustering method; the human leg detection unit also includes a contour adjustment module, which is used to obtain the contour formed by the acquisition point set and move the contour in a polar coordinate system so that the feature vertices of the acquisition point set are aligned with the model vertices of the human leg contour model and the polar angles are the same.

[0014] Furthermore, when the suspicious object construction module constructs a suspicious object, it includes the steps of: obtaining a new feature vertex and adding it to the end of the first feature sequence or the second feature sequence where the feature vertex marked as the same suspicious human leg as the new feature vertex is located; if there is no feature vertex marked as the same suspicious human leg as the new feature vertex, setting the new feature vertex as the waiting feature point; within a waiting time period, when the waiting feature point marked as a different suspicious human leg appears, constructing a new first feature sequence with the waiting feature point marked as the same suspicious human leg, constructing a new second feature sequence with the waiting feature point marked as another same suspicious human leg, and constructing a new suspicious object with the new first feature sequence and the new second feature sequence.

[0015] Furthermore, the suspicious step features include a first step length sequence, a second step length sequence and a step frequency sequence, and the human standard step features include a human standard step length and a human standard step frequency. When the first step length errors between N of the M first step lengths in the first step length sequence and the human standard step length are less than a step length error threshold, and the second step length errors between N of the M second step lengths in the second step length sequence and the human standard step are less than a step length error threshold; and the step frequency errors between N of the M step frequencies in the step frequency sequence and the human standard step frequency are less than a step frequency error threshold, then the detected object is human.

[0016] Furthermore, it also includes a human body temperature detection unit, which is used to obtain the temperature of the detection object and determine whether the detection object is a human based on the comparison result of the temperature of the detection object and the human body temperature characteristics. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic structural diagram of a human intrusion detection security system according to an embodiment of the present invention;

[0018] Figure 2 Schematic diagram of the process of the sliding window clustering method in the embodiment;

[0019] Figure 3 A schematic diagram of the adjustment process of the profile adjustment module in the embodiment;

[0020] Figure 4 This is an example diagram of a suspicious feature point tracking table in the embodiment;

[0021] Figure 5 This is an example diagram of a suspicious object data table in the embodiment;

[0022] Figure 6 This is a storage structure diagram of the waiting space in the embodiment;

[0023] Figure 7A schematic structural diagram of a human intrusion protection system according to a preferred embodiment;

[0024] Figure 8 Schematic diagram of the process of human intrusion detection method in the embodiment;

[0025] Figure 9 Schematic diagram of the process of constructing a suspicious object in an embodiment.

[0026] The technical solution of the present invention is described in detail below with reference to the accompanying drawings. DETAILED DESCRIPTION

[0027] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0028] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. As used in this application and the appended claims, the singular forms "a," "an," "the," and "the" are 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 encompasses any and all possible combinations of one or more of the associated listed items.

[0029] 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 each other. 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 "at the time of" or "when" or "in response to determining".

[0030] The present invention uses a laser radar to collect the outline and position information of objects within its detection range. Based on this information, it detects human legs and steps, as well as body temperature, using these three dimensions to comprehensively determine whether the intruder is human. This is illustrated in the following examples.

[0031] See also Figure 1, which is a schematic diagram of the structure of the human intrusion detection security system of the present invention. The security system includes a data acquisition and preprocessing unit 10, a human leg detection unit 20, a human gait detection unit 30, and an early warning unit 40. The data acquisition and preprocessing unit 10 is used to collect valid position information of the detection object within the detection range; the human leg detection unit 20 is used to detect whether the detection object has human leg characteristics based on the position information collected by the data acquisition and preprocessing unit 10; the human gait detection unit 30 detects whether the detection object has human gait characteristics based on the result data of the human leg detection unit 20; and the early warning unit 40 sends different types of alarm information to the user terminal based on the detection results of the human leg detection unit 20 and the human gait detection unit 30.

[0032] Specifically, the data acquisition and preprocessing unit 10 includes a laser radar 11 and a data clustering module 12. The laser radar 11 collects the position information of the detection object through multiple laser beam collection points falling on the detection object, obtains a collection point set, and transmits it to the data clustering module 12. The position information is the distance ρ and orientation α of each collection point relative to the laser emission point. The laser radar 11 rotates and scans around its own axis, with a scanning cycle of 360°. Each scanning cycle corresponds to a collection point set. Of course, in other embodiments, the rotational scanning angle of the laser radar 11 can be set to any angle according to actual needs.

[0033] The data clustering module 12 is used to filter out the collection points with abnormal positions in the collection point set by using a sliding window clustering method. Figure 2 , which is a flow chart of the sliding window clustering method, specifically comprising the following steps: 1. Calculating the window size X of the clustering window according to the distance ρ of the first acquisition point in a scanning cycle, wherein the window size X of the clustering window decreases as the distance ρ of the first acquisition point in the scanning cycle increases; 2. When the number of acquisition points reaches a set number, selecting the first X acquisition points of the current scanning cycle through the clustering window; 3. Selecting the first acquisition point in the clustering window as the reference point prev; 4. Selecting the next acquisition point in the clustering window as the current point cur, and calculating the absolute distance г between the current point cur and the reference point prev. The calculation formula for the absolute distance г is:

[0034]

[0035] Fifth, when the absolute distance г≤clustering threshold θ, the acquisition point corresponding to the current point cur is added to the clustering point set of the scanning cycle, and the reference point prev is updated to the acquisition point; when the absolute distance г>clustering threshold θ, the acquisition point corresponding to the current point cur is skipped; Sixth, if the acquisition point corresponding to the current point cur is not the last acquisition point in the clustering window, return to step four; if the acquisition point corresponding to the current point cur is the last acquisition point in the clustering window, slide the clustering window by a certain step length, and return to step three; if the acquisition point corresponding to the current point cur is the last acquisition point in the scanning cycle, output the clustering point set of the scanning cycle to the human leg detection unit 20.

[0036] The human leg detection unit 20 includes a contour adjustment module 21 and a contour comparison module 22. The contour adjustment module 21 is used to move the contour formed by the collection points in the cluster point set in the polar coordinate system until the feature vertex K is aligned with the model vertex K0 of the human leg contour model and has the same polar angle. The feature vertex K is the collection point with the smallest distance ρ in the cluster point set. The human leg contour model is a function model representing the standard human leg cross-sectional contour, and the model vertex K0 is the arc vertex of the human leg contour model. Figure 3 , which is a schematic diagram of the adjustment process of the contour adjustment module 21, wherein a1~a4 are the contours formed by the collection points in the cluster point set, and b1~b4 are the contours gradually formed by the collection points in the cluster point set in the polar coordinate system. The specific adjustment method is: move the feature vertex K by a distance Δρ to the origin of the polar coordinate system, and rotate the angle Δα to an orientation α of 90°; change the distance ρ of the remaining collection points in the cluster point set by Δρ, and rotate the orientation α by an angle Δα. Of course, in another embodiment, the polar coordinates of the human leg wheel touch model can also be adjusted so that the feature vertex K in the cluster point set of each scanning cycle is aligned with the vertex K0 of the human leg contour model in the polar coordinate system and has the same polar angle.

[0037] The contour comparison module 22 is used to determine whether the contour of the detection object corresponding to the cluster point set is a suspicious human leg based on the similarity between the contour formed by the cluster point set adjusted by the contour adjustment module 21 and the human leg contour model. Specifically, the contour formed by the adjusted cluster point set is fitted with the human leg contour model, and the determination coefficient (goodness of fit) R is calculated. 2 ; The coefficient of determination R 2 Compare with the similarity threshold E: If the determination coefficient R 2 >similarity threshold E, it means that the detection object outline corresponding to the cluster point set is a suspicious human leg, and the cluster point set is marked as a suspicious feature point set and transmitted to the human step detection unit 30; if the determination coefficient R 2≤ similarity threshold E, it means that the detection object contour corresponding to the cluster point set is not a suspicious human leg, and the cluster point set is marked as a common feature point set.

[0038] The human step detection unit 30 includes a suspicious feature point tracking module 31, a suspicious object construction module 32, and a human step feature comparison module 33. The suspicious feature point tracking module 31 is used to determine whether the suspicious feature point sets of adjacent scanning cycles point to the same suspicious human leg, and mark the feature vertices K of the suspicious feature point sets that point to the same suspicious human leg as the same suspicious human leg and transmit the feature vertices K with the suspicious human leg mark to the suspicious object construction module 32. Specifically, the feature vertices K of the suspicious feature point set of a scanning cycle are calculated with respect to each collection point X of the suspicious feature point set of the previous scanning cycle. i The absolute distance dis is compared with the suspicious threshold β: If there is no collection point X whose absolute distance dis is less than the suspicious threshold β i , then the suspicious feature point set of this scanning cycle and the suspicious feature point set of the previous scanning cycle point to different suspicious human legs. At this time, the feature vertex K is marked as a new suspicious human leg. If there is a collection point X whose absolute distance dis is less than the suspicious threshold β i , then the suspicious feature point set of this scanning cycle and the suspicious feature point set of the previous scanning cycle point to the same suspicious human leg, and the feature vertex K is marked as the same suspicious human leg. The calculation formula of the absolute distance dis is:

[0039] in, is the distance ρ of the i-th collection point of the suspicious feature point set in the previous scanning cycle, is the orientation α, ρ of the i-th acquisition point in the suspicious feature point set of the previous scanning cycle K is the distance ρ, α of the feature vertex K of the suspicious feature point set in the current scanning cycle K is the orientation α of the feature vertex K of the suspicious feature point set in the current scanning cycle.

[0040] In a more specific implementation, a suspicious feature point tracking table can be used to store the suspicious feature point set and related information of the previous scanning cycle, see Figure 4, which is an example diagram of the suspicious feature point tracking table. The suspicious feature point tracking table includes a suspicious feature point set, a suspicious leg number, and suspicious feature information. The suspicious leg number is used to mark different suspicious legs. The suspicious feature information includes the distance ρ, orientation α, and scanning cycle number of the feature vertex K of the suspicious feature point set in the previous scanning cycle. Whenever the suspicious feature point tracking module 31 completes the determination of whether the suspicious feature point set in a scanning cycle and the suspicious feature point set in the scanning cycle before the scanning cycle point to the same suspicious leg, the suspicious feature point tracking table is updated with the suspicious feature point set, suspicious leg number, and suspicious feature information of the current scanning cycle.

[0041] The suspicious object construction module 32 is used to construct suspicious objects based on the feature vertices K marked as different suspicious human legs transmitted by the suspicious feature point tracking module 31. Each suspicious object corresponds to a first feature sequence and a second feature sequence. The first feature sequence and the second feature sequence each contain multiple feature vertices K marked as the same suspicious human leg sorted by scanning period. Constructing a suspicious object specifically includes the following steps: 1. Adding a new feature vertex K to the end of a first feature sequence or a second feature sequence that includes a feature vertex marked as the same suspicious human leg as the new feature vertex K; 2. If there is no first feature sequence or second feature sequence that includes a feature vertex marked as the same suspicious human leg as the new feature vertex K, that is, there is no feature vertex marked as the same suspicious human leg as the new feature vertex K, then setting the new feature vertex K as a waiting feature point; 3. Within a waiting time period T1, when waiting feature points marked as two different suspicious human legs appear among the obtained waiting feature points, constructing a new first feature sequence with the waiting feature points marked as one identical suspicious human leg, constructing a new second feature sequence with the waiting feature points marked as another identical suspicious human leg, and constructing a new suspicious object with the new first feature sequence and the new second feature sequence.

[0042] This embodiment uses the suspicious object data table to store the constructed suspicious objects. Figure 5 , which is an example diagram of the suspicious object data table, wherein the suspicious object data table stores multiple suspicious objects, each suspicious object includes a suspicious object number, a suspicious object validity period, two different suspicious human leg numbers, a first feature sequence and a second feature sequence. Whenever a feature vertex K is added to the first feature sequence or the second feature sequence, the suspicious object validity period of the suspicious object corresponding to the first feature sequence or the second feature sequence is updated with the added time. When the suspicious object validity period corresponding to each suspicious object information expires, it will be cleared from the suspicious object data table to improve the table lookup speed and save storage space.

[0043] This embodiment selects a waiting space to store the waiting feature points. Figure 6, which is a storage structure diagram of the waiting space, the waiting space stores a suspicious person leg number, at least one waiting feature point corresponding to the suspicious person leg number and a waiting timestamp, the waiting timestamp is the time when the first waiting feature point corresponding to the suspicious person leg number is stored, if no new waiting feature point corresponding to a different suspicious person leg number appears within the waiting time period T1 starting from the waiting timestamp, the information stored in the waiting space is cleared; if a new waiting feature point corresponding to a different suspicious person leg number appears within the waiting time period T1 starting from the waiting timestamp, the first feature sequence and the second feature sequence are respectively constructed with the waiting feature points of the two suspicious person leg numbers, and a corresponding suspicious object is constructed.

[0044] The human step feature comparison module 33 calculates the suspicious step feature based on the suspicious object information in the suspicious object construction module 32, and compares the suspicious step feature with the human step feature to determine whether the corresponding detection object is human. Specifically, the relative displacement of the same suspicious person's legs between adjacent scanning cycles is calculated, that is, the first feature sequence {K L1 , K L2 , K L3 ...}, the relative displacement of adjacent feature vertices K is expressed as the first displacement sequence {S L1 , S L2 , S L3 ...} and the second characteristic sequence {K R1 , K R2 , K R3 ...}, the relative displacement of adjacent feature vertices K is expressed as the second displacement sequence {S R1 , S R2 , S R3 ...}; for the first shift sequence {S L1 , S L2 , S L3 ...} and the second shift sequence {S R1 , S R2 , S R3 ...} are separated into groups with zero relative displacement, and the relative displacement and displacement time of each group of sequence are summed to obtain the first step length sequence {Z L1 , Z L2 , Z L3 ...}、First step time series {t L1 , t L2 , t L3 ...}, the second step sequence {Z R1 , Z R2 , Z R3 ...} and the second step time series {t R1 , t R2 , tR3 ...}; According to the first step long sequence {Z L1 , Z L2 , Z L3 ...}、First step time series {t L1 , t L2 , t L3 ...}, the second step sequence {Z R1 , Z R2 , Z R3 ...} and the second step time series {t R1 , t R2 , t R3 ...} calculate the step frequency sequence {f1, f2, f3...}, the calculation formula is Calculate the first step length sequence {Z L1 , Z L2 , Z L3 ...}, the first step length is Z L The error W between the first step length and the human standard step length Z0 ZL , the second step sequence {Z R1 , Z R2 , Z R3 ...}M second step length Z R The second step length error W compared to the human standard step length Z0 ZR The step frequency error W between the M step frequencies f in the step frequency sequence {f1, f2, f3...} and the human standard step frequency f0 f , if there are more than N first step length errors W ZL Less than the step error threshold W Z0 , and the second step error W is more than N times ZR Less than the step error threshold W Z0 , and the step frequency error W is more than N times f Less than the step frequency error W f0 , the detection object is human.

[0045] The types of alarm notifications sent by the early warning unit 40 include "pre-" type and "tight" type. When only the human leg detection unit 20 detects that the detection object is a suspicious human leg, the "pre-" type alarm notification is sent to the user terminal; when the human leg detection unit 20 detects that the detection object is a suspicious human leg, and the human step detection unit 30 detects that the detection object is a human, the "tight" type alarm notification is sent to the user terminal. In a specific implementation, the alarm notification can be a text, voice or picture message such as a short message or phone message. The user terminal can be any smart terminal with Internet access function, for example, it can be a computer, a mobile phone, a tablet computer, a PDA (Personal Digital Assistant), an e-book reader, a multimedia player, etc.

[0046] Furthermore, in order to further enhance the recognition accuracy of the security system, in a preferred embodiment, the security system also includes a human body temperature detection unit 50, which is used to detect whether the detection object has human body temperature characteristics. When the human pace detection unit 30 detects that the detection object within the detection range is a human, the human body temperature detection unit 50 measures the temperature of the detection object. If the measured temperature is within the human body temperature range, the early warning unit 40 sends a "tight" type alarm notification to the user terminal; if the measured temperature is outside the human body temperature range, no "tight" type alarm notification is sent to the user terminal. In a specific implementation, the human body temperature detection unit 50 can be selected as a non-contact infrared temperature sensor.

[0047] Furthermore, in order to expand the security detection range, in a preferred embodiment, the laser radar 11 can move according to a set route and stop moving within a set time period T2 when receiving an intrusion movement signal. In this preferred embodiment, please refer to Figure 7 , which is a structural diagram of a human intrusion security system according to a preferred embodiment, the human step detection unit 30 further includes an intrusion motion signal module 34, the intrusion motion signal module 34 is used to detect the human step detection unit 30 according to the characteristic vertex K of the suspicious feature point set of the current scanning cycle and the collection points X of the suspicious feature point set of the previous scanning cycle. i The absolute distance dis is used to determine whether the same suspicious person's leg moves between two adjacent scanning cycles: if the absolute distance dis is less than the suspicious threshold β and greater than the movement threshold γ, the suspicious person's leg has moved, and an intrusion movement signal is generated.

[0048] In this preferred embodiment, the data acquisition and preprocessing unit 10 further includes a motion sensing module 13 that receives intrusion motion signals and a data correction module 14. The motion sensing module 13 is configured to acquire a radar motion signal from the laser radar 11, indicating whether the laser radar 11 is in motion or stopped, and to acquire the total offset angle and total displacement of the laser radar 11's current position relative to its initial position. The data correction module 14 is configured to correct for positional deviations of acquisition points caused by the movement of the laser radar 11. Specifically, upon receiving the intrusion motion signal, the data correction module 14 calculates the periodic offset angle and periodic displacement of the laser radar 11's position change during the current scanning cycle based on the total offset angle and total displacement of the laser radar 11's current position relative to its initial position, as acquired by the motion sensing module 13. The data correction module 14 then corrects the distance ρ and azimuth α of the acquisition points for each scanning cycle within a set time period T2, starting from the next scanning cycle, based on the periodic offset angle and periodic displacement, to ensure that the acquisition point's position changes in the same manner as the laser radar 11 during the current scanning cycle.

[0049] The system embodiments described above are merely illustrative. The units and modules described as separate components may or may not be physically separate, and the components shown as units and modules may or may not be physical units and modules, that is, they may be located in one place or distributed across multiple network elements. Some or all of the modules may be selected based on actual needs to achieve the objectives of the disclosed solution. Persons of ordinary skill in the art can understand and implement the disclosed solution without inventive effort.

[0050] Based on the same inventive concept, the present invention also provides a method for detecting human intrusion. Figure 8 , which is a flow chart of the human intrusion detection method, specifically comprising the steps of:

[0051] S1: Obtain the distance ρ and azimuth α of each acquisition point where the laser beam of the laser radar falls on the detection object relative to the laser emission point, and obtain the acquisition point set;

[0052] S2: Calculating the similarity between the contour formed by the collection points in the collection point set and the human leg contour model, and when the similarity is greater than a similarity threshold E, determining that the collection point set is a suspicious feature point set;

[0053] S3: Obtain a feature vertex from the suspicious feature point set of a scanning cycle, and obtain the absolute distance dis between the feature vertex and each collection point in the suspicious feature point set of the previous scanning cycle. If the suspicious feature point set of the previous scanning cycle does not have a collection point X whose absolute distance dis is less than the suspicious threshold β, i, then mark the feature vertex as a new suspicious human leg; if the suspicious feature point set of the previous scanning cycle has a collection point X whose absolute distance dis is less than the suspicious threshold β i , then the feature vertex is marked as the suspicious human leg that is the same as the feature vertex in the previous scanning cycle; wherein the feature vertex is the collection point with the smallest distance ρ in the collection point set;

[0054] S4: Construct suspicious objects. Each suspicious object corresponds to a first feature sequence and a second feature sequence. The first feature sequence and the second feature sequence each contain multiple feature vertices K marked as the same suspicious person's legs.

[0055] S5: Calculate a suspicious gait feature based on the first feature sequence and the second feature sequence of the same suspicious object, and determine whether the detected object is a human based on a comparison result of the suspicious gait feature and a human gait feature.

[0056] In order to improve the accuracy of calculating the similarity in step S2, in a preferred embodiment, before calculating the similarity between the contour formed by the collection points in the collection point set and the human leg contour model, the collection points in the collection point set are subjected to the following processing: collection points with abnormal positions in the collection point set are screened out by a sliding window clustering method; and the contour formed by the collection point set is obtained and moved in a polar coordinate system so that the feature vertex K of the collection point set is aligned with the model vertex of the human leg contour model and the polar angle is the same.

[0057] See also Figure 9 , which is a flow chart of constructing a suspicious object in step S4. When constructing a suspicious object, the steps specifically include:

[0058] S41: Acquire a new feature vertex K and add it to the end of the first feature sequence or the second feature sequence where the feature vertex K marked as the same suspicious human leg as the new feature vertex K is located;

[0059] S42: If there is no feature vertex marked as the same suspicious human leg as the new feature vertex K, then the new feature vertex K is set as the waiting feature point;

[0060] S43: When the waiting feature points marked as different suspicious human legs appear within a waiting time period, a new first feature sequence is constructed using the waiting feature points marked as the same suspicious human leg, a new second feature sequence is constructed using the waiting feature points marked as another same suspicious human leg, and a new suspicious object is constructed using the new first feature sequence and the new second feature sequence.

[0061] In order to further improve the accuracy of the detection results of human intrusion, in a preferred embodiment, step S6 is further included: obtaining the temperature of the detection object, and determining whether the detection object is a human based on the comparison result of human body temperature characteristics.

[0062] The remaining specific implementation details of the human intrusion detection method are the same as those of the aforementioned human intrusion detection security system and will not be repeated here.

[0063] The present invention first uses the human leg detection unit 20 to compare the contour data of the detection object collected by the laser radar 11, and determines whether the detection object meets the human characteristics based on the dimension of human leg characteristics. After determining that the detection object has human leg characteristics, the human gait detection unit 30 calculates the detection object's leg stride length and cadence characteristics based on the detection object's position information collected by the laser radar 11, and compares them with the standard human stride length and standard human cadence characteristics to determine whether the detection object meets the human characteristics based on the dimension of human gait characteristics. In addition, the human body temperature detection unit 50 continues to determine whether the detection object meets the human characteristics based on the dimension of human body temperature characteristics. The comprehensive detection results of human leg characteristics, human gait characteristics, and human body temperature characteristics determine whether the detection object is human. Compared with the existing technology, the detection of human intrusion is more targeted, and can detect human intrusion within the security range, avoiding false alarms caused by non-human moving objects entering the security detection range, thereby improving the alarm accuracy of the security system. In addition, the present invention uses laser radar to collect information about the detection object, which is not affected by environmental factors such as ambient light, has strong robustness, and is more private and secure than video surveillance.

[0064] Other embodiments of the present disclosure will readily occur to those skilled in the art after considering the specification and practicing the invention disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, with the true scope and spirit of the present disclosure being indicated by the appended claims.

[0065] It should be understood that the present disclosure 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 the scope of the present disclosure. The scope of the present disclosure is limited only by the appended claims. The above description is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present disclosure shall be included in the scope of protection of the present disclosure.

Claims

1. A human intrusion detection method based on laser radar, characterized in that: Including steps: Obtain the distance ρ and azimuth α of each acquisition point where the laser beam of the laser radar falls on the detection object relative to the laser emission point to obtain a collection point set; Acquire a contour formed by the collection point set, and move the contour in a polar coordinate system so that feature vertices of the collection point set are aligned with model vertices of the human leg contour model and have the same polar angle; Calculating the similarity between the contour formed by the collection points in the collection point set and the human leg contour model, and when the similarity is greater than a similarity threshold E, determining that the collection point set is a suspicious feature point set; Obtain a feature vertex from the suspicious feature point set of a scanning cycle, and obtain the absolute distance dis between the feature vertex and each collection point in the suspicious feature point set of the previous scanning cycle. If the suspicious feature point set of the previous scanning cycle does not have an absolute distance The collection points that are less than the suspicious threshold β , then mark the feature vertex as a new suspicious human leg; if the suspicious feature point set of the previous scanning cycle has an absolute distance The collection points that are less than the suspicious threshold β , then the feature vertex is marked as the suspicious human leg that is the same as the feature vertex in the previous scanning cycle; wherein the feature vertex is the collection point with the smallest distance ρ in the collection point set; The absolute distance The calculation formula is ,in, is the distance ρ of the i-th collection point of the suspicious feature point set in the previous scanning cycle, is the orientation α of the i-th collection point in the suspicious feature point set of the previous scanning cycle, is the distance ρ of the feature vertex K of the suspicious feature point set in the current scanning cycle, is the orientation α of the feature vertex K of the suspicious feature point set in the current scanning cycle; Constructing suspicious objects, each of the suspicious objects corresponding to a first feature sequence and a second feature sequence, wherein the first feature sequence and the second feature sequence each contain a plurality of feature vertices marked as the same suspicious person's legs; A suspicious gait feature is calculated based on the first feature sequence and the second feature sequence of the same suspicious object, and whether the detected object is a human is determined based on a comparison result of the suspicious gait feature and a human gait feature.

2. The human intrusion detection method according to claim 1, characterized in that: Before calculating the similarity between the contour formed by the collection points in the collection point set and the human leg contour model, the method further includes the following steps: The collection points with abnormal positions in the collection point set are screened out by a sliding window clustering method.

3. The human intrusion detection method according to claim 1, characterized in that: When constructing a suspicious object, the following steps are included: Acquire a new feature vertex and add it to the end of the first feature sequence or the second feature sequence where the feature vertex marked as the same suspicious human leg as the new feature vertex is located; If there is no feature vertex marked as the same suspicious human leg as the new feature vertex, then the new feature vertex is set as the waiting feature point; During a waiting time period, when the waiting feature points marked as different suspicious human legs appear, a new first feature sequence is constructed using the waiting feature points marked as the same suspicious human leg, a new second feature sequence is constructed using the waiting feature points marked as another same suspicious human leg, and a new suspicious object is constructed using the new first feature sequence and the new second feature sequence.

4. The human intrusion detection method according to claim 1, characterized in that: The suspicious step features include a first step length sequence, a second step length sequence, and a step frequency sequence; the human step features include a standard human step length and a standard human step frequency. When the first step length errors between N of the M first step lengths in the first step length sequence and the standard human step length are less than a step length error threshold, and the second step length errors between N of the M second step lengths in the second step length sequence and the standard human step are less than a step length error threshold; and the step frequency errors between N of the M step frequencies in the step frequency sequence and the standard human step frequency are less than a step frequency error threshold, then the detected object is human.

5. The human intrusion detection method according to any one of claims 1 to 4, characterized in that: The method further includes the steps of obtaining the temperature of the detection object, and determining whether the detection object is a human being based on a comparison result between the temperature of the detection object and a human body temperature characteristic.

6. A human intrusion security system based on laser radar, characterized in that: include: Data acquisition and preprocessing unit, human leg detection unit and human step detection unit, among which, The data acquisition and preprocessing unit includes a laser radar, which is used to collect the distance ρ and the orientation α of each collection point on the detection object relative to the laser emission point to obtain a collection point set; Acquire a contour formed by the collection point set, and move the contour in a polar coordinate system so that feature vertices of the collection point set are aligned with model vertices of the human leg contour model and have the same polar angle; The human leg detection unit includes a contour comparison module, which is used to calculate the similarity between the contour formed by the collection points in the collection point set and the human leg contour model, and when the similarity is greater than a similarity threshold E, determine that the collection point set is a suspicious feature point set; The human step detection unit includes a suspicious feature point tracking module, a suspicious object construction module and a human step feature comparison module. The suspicious feature point tracking module is used to obtain a feature vertex from the suspicious feature point set of a scanning cycle, and obtain the absolute distance dis between the feature vertex and each collection point in the suspicious feature point set of the previous scanning cycle. If the suspicious feature point set of the previous scanning cycle does not have an absolute distance The collection points that are less than the suspicious threshold β , then mark the feature vertex as a new suspicious human leg; if the suspicious feature point set of the previous scanning cycle has an absolute distance The collection points that are less than the suspicious threshold β , then the feature vertex is marked as the suspicious human leg that is the same as the feature vertex in the previous scanning cycle; wherein the feature vertex is the collection point with the smallest distance ρ in the collection point set; The absolute distance The calculation formula is ,in, is the distance ρ of the i-th collection point of the suspicious feature point set in the previous scanning cycle, is the orientation α of the i-th collection point in the suspicious feature point set of the previous scanning cycle, is the distance ρ of the feature vertex K of the suspicious feature point set in the current scanning cycle, is the orientation α of the feature vertex K of the suspicious feature point set in the current scanning cycle; The suspicious object construction module is used to construct suspicious objects, each of which corresponds to a first feature sequence and a second feature sequence, wherein the first feature sequence and the second feature sequence each contain a plurality of feature vertices marked as the same suspicious person's legs; The human step feature comparison module is used to calculate the suspicious step feature based on the first feature sequence and the second feature sequence of the same suspicious object, and determine whether the detected object is a human based on the comparison result of the suspicious step feature and the human step feature.

7. The human intrusion protection system according to claim 6, characterized in that: The data collection and preprocessing unit further includes: a data clustering module, which is used to screen out collection points with abnormal positions in the collection point set by using a sliding window clustering method.

8. The human intrusion protection system according to claim 6, characterized in that: When the suspicious object construction module constructs a suspicious object, the steps include: Acquire a new feature vertex and add it to the end of the first feature sequence or the second feature sequence where the feature vertex marked as the same suspicious human leg as the new feature vertex is located; If there is no feature vertex marked as the same suspicious human leg as the new feature vertex, then the new feature vertex is set as the waiting feature point; During a waiting time period, when the waiting feature points marked as different suspicious human legs appear, a new first feature sequence is constructed using the waiting feature points marked as the same suspicious human leg, a new second feature sequence is constructed using the waiting feature points marked as another same suspicious human leg, and a new suspicious object is constructed using the new first feature sequence and the new second feature sequence.

9. The human intrusion protection system according to claim 6, characterized in that: The suspicious step features include a first step length sequence, a second step length sequence, and a step frequency sequence; the human step features include a standard human step length and a standard human step frequency. When the first step length errors between N of the M first step lengths in the first step length sequence and the standard human step length are less than a step length error threshold, and the second step length errors between N of the M second step lengths in the second step length sequence and the standard human step are less than a step length error threshold; and the step frequency errors between N of the M step frequencies in the step frequency sequence and the standard human step frequency are less than a step frequency error threshold, then the detected object is human.

10. The human intrusion protection system according to any one of claims 6 to 9, characterized in that: Also includes: A human body temperature detection unit is used to obtain the temperature of the detection object and determine whether the detection object is a human based on the comparison result between the temperature of the detection object and the human body temperature characteristics.

Images