In-vehicle target detection method and related device

Abstract

The application discloses a kind of in-vehicle target detection method and related equipment, the method includes obtaining the first point cloud data of in-vehicle space, wherein the point cloud data of vehicle body structure is not included in first point cloud data;First point cloud data is converted to in-vehicle coordinate system of in-vehicle space, and second point cloud data is obtained;In response to the quantity of second point cloud data included in the first preset region of in-vehicle space reaches preset quantity, the point cloud density of second point cloud data included in the second preset region of in-vehicle space is obtained, wherein the second preset region is arranged inside first preset region, and the volume of second preset region is less than the volume of first preset region;Based on point cloud density, determine whether high-density point cloud zone that meets preset determination rule exists in in-vehicle space;In response to the existence of high-density point cloud zone in in-vehicle space, it is determined that specific detection target exists in in-vehicle space.

Description

Technical Field

[0001] This application belongs to the field of vehicle safety technology, specifically relating to an in-vehicle target detection method, an in-vehicle target detection device, an electronic device, a computer-readable storage medium, a computer program product, and a vehicle. Background Technology

[0002] With the continuous increase in car ownership and the increasing complexity of travel scenarios, safety accidents caused by living beings left in vehicles occur frequently, becoming a pressing vehicle safety technology issue. Currently, in-vehicle cabin liveness detection mainly employs three technologies: vision, infrared, and pressure sensing. Visual detection uses cameras to capture images for identification, but its accuracy drops significantly under adverse lighting conditions such as nighttime or strong light. Infrared detection relies on capturing thermal radiation, but it is easily interfered with by heat sources such as insulated cups and heaters, leading to false positives, and its sensitivity also decreases significantly in low-temperature environments. Pressure sensing technology collects data through seat sensors, but it has blind spots for detecting lightweight infants, small pets, and living beings in curled-up positions. These technologies generally suffer from poor environmental adaptability, limited judgment criteria, and high false positive and false negative rates. Therefore, there is an urgent need to develop a solution that can overcome these technological shortcomings and achieve accurate, stable, and non-invasive in-vehicle target detection. Summary of the Invention

[0003] This application aims to address at least one of the technical problems existing in the prior art. To this end, this application proposes an in-vehicle target detection method that can solve problems such as in-vehicle detection being susceptible to interference from the environment and occlusion, difficulty in accurately distinguishing between living and non-living objects, frequent false positives and false negatives, and inability to guarantee privacy and security.

[0004] To achieve the above objectives, a first aspect of the present invention provides a method for detecting targets inside a vehicle, comprising: acquiring first point cloud data of the vehicle interior space, wherein the first point cloud data does not include point cloud data of the vehicle body structure; converting the first point cloud data to the vehicle interior coordinate system to obtain second point cloud data; in response to the number of second point cloud data included in a first preset area of ​​the vehicle interior space reaching a preset number, acquiring the point cloud density of the second point cloud data included in the second preset area of ​​the vehicle interior space, wherein the second preset area is set inside the first preset area and the volume of the second preset area is smaller than the volume of the first preset area; determining whether there is a high-density point cloud band satisfying a preset judgment rule in the vehicle interior space based on the point cloud density; and in response to the existence of a high-density point cloud band in the vehicle interior space, determining that a specific detection target exists in the vehicle interior space.

[0005] In some embodiments, acquiring first point cloud data of the vehicle interior space includes: acquiring raw point cloud data obtained by scanning the vehicle interior space with radar, wherein the raw point cloud data includes the three-dimensional spatial coordinates and point cloud reflection intensity of each point in the vehicle interior space; and removing point cloud data of the vehicle body structure and point cloud data whose Euclidean distance from other raw point cloud data exceeds a preset threshold from the raw point cloud data using a preset point cloud matching template to obtain the first point cloud data.

[0006] In some embodiments, converting the first point cloud data to the in-vehicle coordinate system of the in-vehicle space to obtain the second point cloud data includes: determining a rotation matrix for correcting the radar installation angle deviation and a translation vector for correcting the radar installation position deviation; and obtaining the second point cloud data based on the first point cloud data, the rotation matrix, and the translation vector.

[0007] In some embodiments, the method further includes: dividing each seat area into a first preset area based on the seating layout of the vehicle interior space; and dividing each first preset area into several second preset areas along the longitudinal direction of the vehicle interior space.

[0008] In some embodiments, determining whether a high-density point cloud band exists in the vehicle interior space based on point cloud density includes: traversing all second preset regions along the longitudinal direction of the vehicle interior space and calculating the point cloud density of the second point cloud data included in each second preset region; in response to the point cloud density of at least a preset number of consecutive adjacent second preset regions exceeding a preset density threshold, determining that a high-density point cloud band exists in the vehicle interior space.

[0009] In some embodiments, the method further includes: extracting the 1D signal time series corresponding to the point cloud of the high-density point cloud band, filtering the 1D signal time series to obtain the target signal sequence; extracting the respiratory micro-motion signal in the target signal sequence, and determining whether the respiratory frequency represented by the respiratory micro-motion signal is within a preset frequency range; and determining the specific detection target as a human target in response to the respiratory frequency being within the preset frequency range.

[0010] In some embodiments, the method further includes: in response to the number of second point cloud data included in a first preset area of ​​the vehicle interior space reaching a preset number, determining that a detection target exists in the vehicle interior space, wherein the detection target is a human target or a non-human target.

[0011] In some embodiments, the method further includes: determining that there is no detection target in the vehicle interior space in response to the fact that the number of second point cloud data included in a first preset area of ​​the vehicle interior space does not reach a preset number.

[0012] In some embodiments, the method further includes: in response to a respiratory rate not being within a preset frequency range, determining a specific detection target as a living target but not a human target.

[0013] According to an embodiment of the present invention, an in-vehicle target detection method includes: acquiring first point cloud data of the in-vehicle space, wherein the first point cloud data does not include point cloud data of the vehicle body structure; converting the first point cloud data to an in-vehicle coordinate system of the in-vehicle space to obtain second point cloud data; in response to the number of second point cloud data included in a first preset area of ​​the in-vehicle space reaching a preset number, acquiring the point cloud density of the second point cloud data included in the second preset area of ​​the in-vehicle space, wherein the second preset area is set inside the first preset area and the volume of the second preset area is smaller than the volume of the first preset area; determining whether there is a high-density point cloud band satisfying a preset judgment rule in the in-vehicle space based on the point cloud density; and in response to the existence of a high-density point cloud band in the in-vehicle space, determining that a specific detection target exists in the in-vehicle space. Therefore, this application can solve the problems of in-vehicle detection being easily affected by environmental and occlusion interference, difficulty in accurately distinguishing between living and non-living objects, frequent false positives and false negatives, and lack of privacy and security. It achieves accurate identification and positioning of in-vehicle targets in all weather and all areas, effectively filters out various interference factors, eliminates false positives and false negatives caused by changes in light, occlusion, heat source interference, and abnormal posture, and improves the accuracy and stability of detection.

[0014] To achieve the above objectives, a second aspect of the present invention provides an in-vehicle target detection device, comprising: an acquisition module configured to acquire first point cloud data of the in-vehicle space, wherein the first point cloud data does not include point cloud data of the vehicle body structure; a conversion module configured to convert the first point cloud data to an in-vehicle coordinate system of the in-vehicle space to obtain second point cloud data; a response module configured to, in response to a preset number of second point cloud data included in a first preset area of ​​the in-vehicle space reaching a preset number, acquire the point cloud density of the second point cloud data included in the second preset area of ​​the in-vehicle space, wherein the second preset area is located inside the first preset area and the volume of the second preset area is smaller than the volume of the first preset area; a determination module configured to determine whether a high-density point cloud band satisfying a preset determination rule exists in the in-vehicle space based on the point cloud density; and a determination module configured to, in response to the presence of a high-density point cloud band in the in-vehicle space, determine that a specific detection target exists in the in-vehicle space.

[0015] An in-vehicle target detection device according to an embodiment of the present invention includes: an acquisition module configured to acquire first point cloud data of an in-vehicle space, wherein the first point cloud data does not include point cloud data of the vehicle body structure; a conversion module configured to convert the first point cloud data to an in-vehicle coordinate system of the in-vehicle space to obtain second point cloud data; a response module configured to acquire the point cloud density of the second point cloud data included in a second preset area of ​​the in-vehicle space in response to the number of second point cloud data included in a first preset area of ​​the in-vehicle space reaching a preset number, wherein the second preset area is located inside the first preset area and the volume of the second preset area is smaller than the volume of the first preset area; a determination module configured to determine whether a high-density point cloud band satisfying a preset determination rule exists in the in-vehicle space based on the point cloud density; and a determination module configured to determine the existence of a specific detection target in the in-vehicle space in response to the existence of a high-density point cloud band in the in-vehicle space. Therefore, this application can solve the problems of in-vehicle detection being easily affected by environmental and occlusion interference, difficulty in accurately distinguishing between living and non-living objects, frequent false positives and false negatives, and lack of privacy and security. It achieves accurate identification and positioning of in-vehicle targets in all weather and all areas, effectively filters out various interference factors, eliminates false positives and false negatives caused by changes in light, occlusion, heat source interference, and abnormal posture, and improves the accuracy and stability of detection.

[0016] To achieve the above objectives, a third aspect of the present invention provides an electronic device, comprising: a processor and a memory, wherein the memory stores a program or instructions executable on the processor, and the program or instructions, when executed by the processor, implement the steps of the in-vehicle target detection method as described in any of the foregoing.

[0017] The electronic device according to the embodiments of the present invention, by executing the aforementioned in-vehicle target detection method, can solve the problems of in-vehicle detection being easily affected by environmental and occlusion interference, difficulty in accurately distinguishing between living and non-living objects, frequent false positives and false negatives, and inability to guarantee privacy and security. It achieves accurate identification and positioning of in-vehicle targets in all weather and all areas, effectively filters various interference factors, eliminates detection false positives and false negatives caused by changes in light, occlusion, heat source interference, and abnormal posture, and improves the accuracy and stability of detection.

[0018] To achieve the above objectives, a fourth aspect of the present invention provides a computer-readable storage medium storing a program or instructions that, when executed by a processor, implement the steps of the in-vehicle target detection method as described in any of the preceding claims.

[0019] According to the computer-readable storage medium of the present invention, by executing the aforementioned in-vehicle target detection method, the problems of in-vehicle detection being susceptible to environmental and occlusion interference, difficulty in accurately distinguishing between living and non-living objects, frequent false positives and false negatives, and lack of privacy and security can be solved. It achieves accurate identification and positioning of in-vehicle targets in all weather and all areas, effectively filters various interference factors, eliminates detection false positives and false negatives caused by changes in light, occlusion, heat source interference, and abnormal posture, and improves the accuracy and stability of detection.

[0020] To achieve the above objectives, a fifth aspect of the present invention provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the in-vehicle target detection method as described in any of the preceding claims.

[0021] The computer program product according to the embodiments of the present invention, by executing the aforementioned in-vehicle target detection method, can solve the problems of in-vehicle detection being easily affected by environmental and occlusion interference, difficulty in accurately distinguishing between living and non-living objects, frequent false positives and false negatives, and inability to guarantee privacy and security. It achieves accurate identification and positioning of in-vehicle targets in all weather and all areas, effectively filters various interference factors, eliminates detection false positives and false negatives caused by changes in light, occlusion, heat source interference, and abnormal posture, and improves the accuracy and stability of detection.

[0022] To achieve the above objectives, a sixth aspect of the present invention provides a vehicle including the aforementioned electronic equipment.

[0023] The vehicle according to embodiments of the present invention, including the aforementioned electronic equipment, can solve the problems of in-vehicle detection being susceptible to environmental and occlusion interference, difficulty in accurately distinguishing between living and non-living objects, frequent false positives and false negatives, and inability to guarantee privacy and security. It achieves accurate identification and positioning of in-vehicle targets in all weather and all areas, effectively filters various interference factors, eliminates detection false positives and false negatives caused by changes in light, occlusion, heat source interference, and abnormal posture, and improves the accuracy and stability of detection.

[0024] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0025] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a flowchart illustrating an in-vehicle target detection method according to an embodiment of this application; Figure 2 This is a schematic diagram of the interference point cloud removal process of an in-vehicle target detection method in an embodiment of this application; Figure 3 This is a schematic diagram of the coordinate transformation process of an in-vehicle target detection method according to an embodiment of this application; Figure 4 This is a schematic diagram illustrating the process of dividing a first preset region and a second preset region in an in-vehicle target detection method according to an embodiment of this application. Figure 5 This is a simplified schematic diagram of the first and second preset regions of an in-vehicle target detection method according to an embodiment of this application. Figure 6 This is a schematic diagram illustrating the high-density point cloud band determination process of an in-vehicle target detection method in an embodiment of this application; Figure 7 This is a schematic diagram of the structure of an in-vehicle target detection device according to an embodiment of this application; Figure 8 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application.

[0026] Reference numerals: Acquisition module 701, Conversion module 702, Response module 703, Determination module 704, Judgment module 705, Processor 810, Memory 820, Input / Output interface 830, Communication interface 840, Bus 850. Detailed Implementation

[0027] Embodiments of this application will now be described in more detail with reference to the accompanying drawings. While some embodiments of this application are shown in the drawings, it should be understood that this application can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this application. It should be understood that the drawings and embodiments of this application are for illustrative purposes only and are not intended to limit the scope of protection of this application.

[0028] It should be understood that the steps described in the method embodiments of this application may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of this application is not limited in this respect.

[0029] As described in the background section, with the continuous increase in car ownership and the increasing complexity of travel scenarios, safety accidents caused by leaving living beings inside vehicles occur frequently, becoming a pressing social issue. Young children often enter unlocked vehicles out of curiosity, or are inadvertently left in the back seats by parents or school bus staff. The enclosed car interior heats up extremely quickly under sunlight; even at an outdoor temperature of only 20°C, the interior temperature can rise to 46°C within an hour, and can exceed 60°C in summer. Children have weaker thermoregulation abilities, making them highly vulnerable to life-threatening situations within a short period. Currently, the mainstream technologies for in-vehicle liveness detection include visual detection, infrared detection, and pressure sensing. Among these technologies, visual detection relies on cameras, but its accuracy drops significantly under adverse lighting conditions such as nighttime, strong light, and heavy rain. It is also susceptible to obstruction by seats and clothing, and poses a risk of privacy data leakage. Infrared detection captures thermal radiation signals, but is easily interfered with by heat sources such as thermos cups and heaters, leading to false alarms. Its sensitivity also decreases significantly in low temperatures. Pressure sensing technology, based on seat pressure sensors, is insufficient for detecting lighter infants, non-standard postures, or curled-up individuals, and is prone to false alarms triggered by items such as backpacks and pillows. All of these technologies suffer from limitations such as poor adaptability, high false alarm rates, and privacy risks, making it difficult to meet the requirements for all-weather, all-round, and highly reliable in-cabin liveness detection.

[0030] To address the shortcomings of the aforementioned detection methods, the in-vehicle target detection method of this invention can solve problems such as the susceptibility of in-vehicle detection to environmental and occlusion interference, difficulty in accurately distinguishing between living and non-living objects, frequent false positives and false negatives, and the inability to guarantee privacy and security. It achieves accurate identification and positioning of in-vehicle targets in all weather conditions and all areas, effectively filters out various interference factors, eliminates detection false positives and false negatives caused by changes in light, occlusion, heat source interference, and abnormal posture, and improves the accuracy and stability of detection.

[0031] The following is for reference. Figures 1-6 This application describes the in-vehicle target detection method provided in its embodiments.

[0032] like Figure 1 The diagram shown is a flowchart of an in-vehicle target detection method according to an embodiment of this application.

[0033] Step S101: Obtain the first point cloud data of the vehicle interior space, wherein the first point cloud data does not include the point cloud data of the vehicle body structure.

[0034] Specifically, the entire interior space is scanned and collected using millimeter-wave radar or UWB radar to obtain raw point cloud data containing the three-dimensional coordinates and reflection intensity information of each measuring point. This data is then combined with point cloud feature templates of pre-collected fixed vehicle structures such as seat backs and headrests to complete matching and elimination. The average distance between each measuring point and its neighboring points is calculated, and isolated noise points exceeding the distance threshold are filtered out to obtain the first point cloud data of the interior space.

[0035] It should be noted that millimeter-wave radar operates in the 30–300 GHz frequency band, has all-weather adaptability and a certain penetration capability, and can accurately capture target distance and speed information, making it suitable for complex in-vehicle environments. UWB radar has ultra-high temporal resolution and centimeter-level ranging accuracy, with extremely strong micro-motion sensing capabilities, and can accurately capture weak life signals from living beings. Neither of them requires image acquisition, fundamentally avoiding the risk of privacy leaks. The choice can be made based on the in-vehicle detection accuracy and cost requirements.

[0036] Step S102: Convert the first point cloud data to the in-vehicle coordinate system of the vehicle interior space to obtain the second point cloud data.

[0037] Specifically, the second point cloud data is obtained by transforming the radar's own coordinate system to a unified in-vehicle coordinate system.

[0038] Step S103: In response to the number of second point cloud data included in the first preset area of ​​the vehicle interior space reaching a preset number, the point cloud density of the second point cloud data included in the second preset area of ​​the vehicle interior space is obtained, wherein the second preset area is set inside the first preset area and the volume of the second preset area is smaller than the volume of the first preset area.

[0039] Specifically, the first preset area is the area set within the actual space of each seat, covering the space from the front and back of the seat, the side, and the seat surface to the human shoulder. The number of second point cloud data contained in each first preset area is counted. When the number of point cloud data in a certain first preset area reaches a preset threshold, it is determined that there is a suspected target in the area. The point cloud density of multiple second preset areas set inside the first preset area is obtained. The second preset areas are located inside the first preset area and have a smaller volume than the first preset area. They are used to perform detailed point cloud distribution feature analysis on the suspected target area.

[0040] Step S104: Determine whether there is a high-density point cloud band in the vehicle interior space that meets the preset judgment rules based on the point cloud density.

[0041] Specifically, all second preset areas are traversed along the longitudinal direction of the vehicle interior space, and the presence of high-density point cloud bands that meet the conditions is determined according to preset judgment rules.

[0042] Step S105: In response to the presence of a high-density point cloud band in the vehicle interior space, it is determined that a specific detection target exists in the vehicle interior space.

[0043] Specifically, the high-density point cloud band is composed of a continuous preset number of adjacent second preset regions. The continuity and aggregation of its point cloud distribution match the spatial distribution characteristics of the human body contour, which can reflect the actual occupancy status of the human body in the seat area of ​​the vehicle. The presence of a specific detection target in the vehicle can be determined based on the high-density point cloud band.

[0044] like Figure 2 The diagram illustrates the interference point cloud removal process of an in-vehicle target detection method according to an embodiment of this application. The process involves acquiring first point cloud data of the in-vehicle space, including: Step S201: Obtain raw point cloud data obtained by scanning the interior space of the vehicle with radar. The raw point cloud data includes the three-dimensional spatial coordinates of each point in the interior space and the point cloud reflection intensity.

[0045] Specifically, the radar scans and collects data across the entire interior space of the vehicle, outputting a single frame of raw point cloud data. This raw point cloud data represents the information of each measuring point within the vehicle space in three-dimensional coordinates. A single frame of point cloud data can be represented as a set:

[0046] A radar coordinate system is established with the radar installation location as the origin. The positive x-axis points directly in front of the radar and extends along the illumination direction of the radar's main beam. The positive y-axis points to the right side of the radar, and the positive z-axis points vertically upward, perpendicular to the plane formed by the x-axis and y-axis. For the first point cloud in a single frame One point, These are the three-dimensional spatial coordinates of the point in the radar coordinate system. The point cloud reflection intensity corresponding to that point. This represents the total number of points contained in a single frame of point cloud data.

[0047] Step S202: Remove the point cloud data of the vehicle body structure and the point cloud data whose Euclidean distance from other original point cloud data exceeds a preset threshold from the original point cloud data using a preset point cloud matching template to obtain the first point cloud data.

[0048] Specifically, the pre-collected point cloud feature templates of fixed structures in the vehicle body, such as seat backs and headrests, are matched and compared with the original point cloud data. Point cloud data corresponding to static fixed structures are identified and removed to eliminate inherent static interference terms within the vehicle. Neighborhood distance statistical analysis is then performed on the remaining point cloud data, calculating the Euclidean distance between each point and its neighbors. Isolated points whose distance exceeds a threshold are removed. The threshold formula is:

[0049] in, For point With the Euclidean distance between neighboring points The mean of the neighborhood distance. The standard deviation of the neighborhood distance. Set the value to 15. When the Euclidean distance between a point and its neighboring points exceeds a preset threshold... When an isolated point is identified as having no practical significance, it is removed.

[0050] like Figure 3 The diagram illustrates the coordinate transformation process of an in-vehicle target detection method according to an embodiment of this application. The first point cloud data is transformed into the in-vehicle coordinate system to obtain the second point cloud data, including: Step S301: Determine the rotation matrix used to correct the radar installation angle deviation and the translation vector used to correct the radar installation position deviation.

[0051] Specifically, during radar installation, there may be angular deviations and positional offsets. Directly using point cloud data in the radar coordinate system will not match the actual seating layout inside the vehicle, affecting the accuracy of subsequent area division and target detection. Coordinate correction is required using pre-calibrated parameters, and the transformation relationship is as follows:

[0052] in, for A rotation matrix is ​​used to correct for radar installation angle deviations. for Translation vector, used to correct for radar installation position deviations. The coordinates of the first point cloud data in the radar coordinate system. These are the coordinates in the in-vehicle coordinate system after the transformation.

[0053] Step S302: Obtain the second point cloud data based on the first point cloud data, the rotation matrix, and the translation vector.

[0054] Specifically, in order to reduce the impact of radar installation angle deviation and position offset on data spatial positioning, the radar installation angle error is corrected by a rotation matrix, and the radar installation position deviation is compensated by a translation vector. The point cloud data in the radar's own coordinate system is mapped to a unified in-vehicle coordinate system to complete the coordinate correction and transformation of the point cloud data and obtain the second point cloud data.

[0055] like Figure 4 The diagram shown illustrates the process of dividing a first preset region and a second preset region in an in-vehicle target detection method according to an embodiment of this application. The method further includes: Step S401: Based on the seating layout of the vehicle interior space, divide each seating area into a first preset area.

[0056] Specifically, based on the actual layout of the driver's seat, front passenger seat, and rear seats inside the vehicle, a corresponding first preset area is defined for each seating area. Each first preset area... The dimensions are adapted to the actual space of the seat, and the space range can be expressed as:

[0057] With the vehicle body coordinate system as the reference, the x-axis runs along the longitudinal direction of the vehicle body, corresponding to the front and rear range of the seat. This is the minimum coordinate of the seating area along the longitudinal direction of the vehicle body (near the boundary of the front of the vehicle). The y-axis represents the maximum longitudinal coordinate of the seating area within the vehicle body (near the rear boundary), which can be fine-tuned depending on the vehicle model; the y-axis runs laterally along the vehicle body, corresponding to the left and right boundaries of the seat. Let this be the minimum left coordinate of the seating area on the lateral side of the vehicle body. This represents the maximum right-hand coordinate of the seating area across the vehicle body; it can be set in segments for rear-seat configurations with multiple seats. The z-axis is perpendicular to the vehicle body and extends upwards, covering the height range from the seat surface to the passenger's shoulder. This is the lowest coordinate of the seating area (usually the height of the seat cushion surface). The highest coordinate of the seating area (usually the spatial position corresponding to the height of a person's shoulder) is used to exclude interference from irrelevant areas such as the ground and the roof.

[0058] Step S402: Divide each first preset area into several second preset areas along the longitudinal direction of the vehicle interior space.

[0059] Specifically, such as Figure 5 The diagram shown is a simplified illustration of a first preset region and a second preset region in an embodiment of this application for a target detection method inside a vehicle. The second preset region... The spatial extent can be represented as:

[0060] in, For the first The starting coordinates of the second preset region in the vertical direction. The equally spaced step sizes are used to divide the longitudinal direction. For the first The starting coordinates of the second preset region in the horizontal direction. Equal intervals are used to divide the space in the horizontal direction. For the first The starting coordinates of the second preset region in the vertical direction. The vertically divided equally spaced step size. The segmentation range is in the vertical direction. This refers to the segmentation range in the horizontal direction. The segmentation range is in the vertical direction (height direction). This represents the total number of regions divided horizontally. This represents the total number of regions divided in the vertical direction. This represents the total number of regions divided vertically. The volume of the second preset region is smaller than that of the first preset region, which can capture the point cloud aggregation characteristics of local contours such as the human head and torso.

[0061] The number of second point cloud data points within each first preset region can be represented as:

[0062] in, This refers to the spatial range corresponding to the first preset region in the current statistics. For the first A second cloud data point that falls within the first preset area As an indicator function, set a threshold for the number of point clouds. If the value exceeds the threshold, it is determined that there is a target in the seating area, and the second preset area is entered for further analysis; otherwise, it is determined that there is no target, and the subsequent steps are skipped.

[0063] like Figure 6 The diagram illustrates the process of determining a high-density point cloud band in a vehicle interior target detection method according to an embodiment of this application. The determination of whether a high-density point cloud band satisfying a preset determination rule exists in the vehicle interior space based on the point cloud density includes: Step S601: Traverse all second preset regions along the longitudinal direction of the vehicle interior space and calculate the point cloud density of the second point cloud data included in each second preset region.

[0064] Specifically, the effective second point cloud data within the second preset area is filtered according to the longitudinal direction of the vehicle interior space. Invalid interference point clouds such as vehicle body noise and discrete noise are removed. The number of effective point clouds within each area is counted by using the fixed spatial volume of each area, which can be expressed as:

[0065] in, This refers to the spatial range corresponding to the second preset region in the current statistics. For the first The second cloud data point falls within the second preset area. As an indicator function, set a threshold for the number of point clouds. When the value exceeds the threshold, it is determined that a target exists in the seating area, and the point cloud density of the second point cloud data included in each second preset area is calculated.

[0066] Step S602: In response to the point cloud density of at least a preset number of adjacent second preset regions exceeding a preset density threshold, it is determined that a high-density point cloud band exists in the vehicle interior space.

[0067] Specifically, the point cloud density of each second preset region is compared with a preset density threshold. If the point cloud density of at least a preset number of consecutive adjacent second preset regions exceeds the preset density threshold, it is determined that these continuously distributed regions together constitute a high-density point cloud band in the vehicle interior space. This feature is consistent with the distribution characteristics of continuous human body contours and strong point cloud aggregation. If, after traversing all second preset regions, no situation is found where the point cloud density of consecutively adjacent regions exceeds the threshold, it is determined that there is no high-density point cloud band that meets the conditions in the vehicle interior space.

[0068] As an optional embodiment, the method further includes: extracting the 1D signal time series corresponding to the point cloud of the high-density point cloud band, filtering the 1D signal time series to obtain the target signal sequence; extracting the respiratory micro-motion signal in the target signal sequence, and determining whether the respiratory frequency represented by the respiratory micro-motion signal is within a preset frequency range; and determining the specific detection target as a human target in response to the respiratory frequency being within the preset frequency range.

[0069] Specifically, after the radar received signal is sampled by AD, it first undergoes 1DFFT (range-dimensional FFT) to obtain a 1D range-dimensional signal containing phase information, then undergoes 2DFFT (velocity-dimensional FFT) to obtain velocity information, and then undergoes angle FFT to obtain angle information. These range, velocity, and angle information are combined to generate point cloud data. The point cloud itself does not contain the original phase information, but subtle movements during human breathing cause minute changes in the distance between the target and the radar. This change directly manifests as a periodic phase change in the corresponding 1D range-dimensional signal. The 1D signal time series corresponding to each point within the high-density point cloud band is extracted to construct an echo change sequence reflecting the subtle movement characteristics of the target position. This sequence is then bandpass filtered. The target signal sequence is obtained by filtering out noise and interference within the normal human respiratory frequency range (e.g., 0.1Hz~1Hz). The MUSIC isospectral estimation algorithm is then used to analyze the frequency of the target signal sequence. This algorithm can identify periodic frequency components in the signal at high resolution, extracting respiratory micro-motion signals. Finally, it is determined whether the respiratory frequency represented by the respiratory micro-motion signal falls within a preset frequency range. The preset frequency range can be set according to human physiological characteristics (e.g., 0.1Hz~0.8Hz corresponds to 6 to 48 breaths per minute, covering the normal respiratory frequency range for adults and children). If the frequency of the respiratory micro-motion signal falls within the preset frequency range, the specific detection target is identified as a human target. To reduce the false positive rate caused by transient noise or abnormal reflections in single-frame data, a sliding window with a window size of 10 frames is set. The number of consecutive single frames within the window that are identified as human targets is counted. When 5 or more frames meet the requirement, the existence of a human target within the first preset area is confirmed, and its location is marked.

[0070] As an optional embodiment, the method further includes: in response to the number of second point cloud data included in a first preset area of ​​the vehicle interior space reaching a preset number, determining that a detection target exists in the vehicle interior space, wherein the detection target is a human target or a non-human target.

[0071] Specifically, the number of second point cloud data falling within the first preset area is counted. When the number of point cloud data in a certain first preset area reaches the preset threshold, it indicates that there is an object in the area that can generate effective radar reflection. It is determined that there is a detection target in the area inside the vehicle. The target is divided into two categories: human target or non-human target. The target type can be distinguished by high-density point cloud band determination, respiratory micro-motion signal detection, etc.

[0072] As an optional embodiment, the method further includes: in response to the fact that the number of second point cloud data included in the first preset area of ​​the vehicle interior space does not reach a preset number, determining that there is no detection target in the vehicle interior space.

[0073] Specifically, when the number of point clouds in a certain first preset area does not reach the preset number threshold, it indicates that there is no object in the seating area that can generate sufficient radar echoes, no human body occupies the seat, and no other non-human body objects, thus determining that there is no detection target in the area inside the vehicle.

[0074] As an optional embodiment, the method further includes: in response to the respiratory rate not being within a preset frequency range, determining that a specific detection target is a living target but not a human target.

[0075] Specifically, when a high-density point cloud band is detected and a periodic respiratory micro-motion signal is extracted, but the frequency of the signal is not within the preset normal respiratory frequency range for humans, the specific detection target can be determined to be a living target but not a human target. The normal respiratory frequency of pets such as cats and dogs is generally higher than that of humans. For example, the respiratory frequency of dogs is about 0.2Hz to 1.5Hz and that of cats is about 0.3Hz to 2Hz. These living targets will also produce periodic micro-motions due to breathing, but the frequency characteristics are significantly different from those of humans.

[0076] In summary, the in-vehicle target detection method provided in this application includes: acquiring first point cloud data of the in-vehicle space, wherein the first point cloud data does not include point cloud data of the vehicle body structure; converting the first point cloud data to the in-vehicle coordinate system of the in-vehicle space to obtain second point cloud data; in response to the number of second point cloud data included in a first preset area of ​​the in-vehicle space reaching a preset number, acquiring the point cloud density of the second point cloud data included in the second preset area of ​​the in-vehicle space, wherein the second preset area is set inside the first preset area and the volume of the second preset area is smaller than the volume of the first preset area; determining whether there is a high-density point cloud band satisfying a preset judgment rule in the in-vehicle space based on the point cloud density; and in response to the existence of a high-density point cloud band in the in-vehicle space, determining that a specific detection target exists in the in-vehicle space. Therefore, this application can solve the problems of in-vehicle detection being easily affected by environmental and occlusion interference, difficulty in accurately distinguishing between living and non-living objects, frequent false positives and false negatives, and lack of privacy and security. It achieves accurate identification and positioning of in-vehicle targets in all weather and all areas, effectively filters out various interference factors, eliminates false positives and false negatives caused by changes in light, occlusion, heat source interference, and abnormal posture, and improves the accuracy and stability of detection.

[0077] It should be noted that the method of this embodiment can be executed by a single device, such as a computer or server. The method of this embodiment can also be applied to a distributed scenario, where multiple devices cooperate to complete the task. In such a distributed scenario, one of these devices may execute only one or more steps of the method of this embodiment, and the multiple devices will interact with each other to complete the above method.

[0078] It should be noted that the above description describes some embodiments of the present invention. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims may be performed in a different order than that shown in the above embodiments and still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0079] refer to Figure 7 This is a schematic diagram of the structure of an in-vehicle target detection device in an embodiment of this application.

[0080] The in-vehicle target detection device includes: an acquisition module 701, a conversion module 702, a response module 703, a determination module 704, and a judgment module 705.

[0081] The acquisition module 701 is configured to acquire first point cloud data of the vehicle interior space, wherein the first point cloud data does not include point cloud data of the vehicle body structure; the conversion module 702 is configured to convert the first point cloud data to the vehicle interior coordinate system to obtain second point cloud data; the response module 703 is configured to acquire the point cloud density of the second point cloud data included in the second preset area of ​​the vehicle interior space in response to the number of second point cloud data included in the first preset area of ​​the vehicle interior space reaching a preset number, wherein the second preset area is set inside the first preset area and the volume of the second preset area is smaller than the volume of the first preset area; the determination module 704 is configured to determine whether there is a high-density point cloud band satisfying a preset judgment rule in the vehicle interior space based on the point cloud density; the judgment module 705 is configured to determine the existence of a specific detection target in the vehicle interior space in response to the existence of a high-density point cloud band in the vehicle interior space.

[0082] The acquisition module 701 is also configured as follows: Acquire the first point of cloud data of the vehicle interior space, including: The raw point cloud data obtained by scanning the interior space of the vehicle with radar is acquired. The raw point cloud data includes the three-dimensional spatial coordinates of each point in the interior space and the point cloud reflection intensity. The first point cloud data is obtained by removing the point cloud data of the vehicle body structure and the point cloud data whose Euclidean distance from other original point cloud data exceeds a preset threshold from the original point cloud data using a preset point cloud matching template.

[0083] The conversion module 702 is also configured as follows: The first point cloud data is transformed into the in-vehicle coordinate system to obtain the second point cloud data, including: Determine the rotation matrix used to correct the radar installation angle deviation and the translation vector used to correct the radar installation position deviation; The second point cloud data is obtained based on the first point cloud data, the rotation matrix, and the translation vector.

[0084] Response module 703 is also configured as follows: Based on the seating layout inside the vehicle, a first preset area is divided for each seating area; Within each first preset area, several second preset areas are divided along the longitudinal direction of the vehicle interior space.

[0085] Module 704 is also configured as follows: Determining whether a high-density point cloud band exists in the vehicle interior space based on point cloud density, including: Traverse all second preset regions along the longitudinal direction of the vehicle interior space and calculate the point cloud density of the second point cloud data included in each second preset region; In response to the point cloud density of at least a preset number of consecutive adjacent second preset regions exceeding a preset density threshold, it is determined that a high-density point cloud band exists in the vehicle interior space.

[0086] The determination module 705 is also configured as follows: Extract the time series of point cloud reflection intensity from the high-density point cloud band, filter the point cloud reflection intensity time series, and obtain the target signal sequence; Extract respiratory micro-motion signals from the target signal sequence and determine whether the respiratory frequency represented by the respiratory micro-motion signals is within the preset frequency range; In response to a respiratory rate within a preset frequency range, a specific detection target is identified as a human target.

[0087] Optionally, the determination module 705 is also configured as follows: In response to the number of second point cloud data included in the first preset area of ​​the vehicle interior space reaching a preset number, it is determined that there is a detection target in the vehicle interior space, wherein the detection target is a human target or a non-human target.

[0088] Optionally, the determination module 705 is also configured as follows: If the number of second point cloud data points included in the first preset area of ​​the vehicle interior space does not reach the preset number, it is determined that there is no detection target in the vehicle interior space.

[0089] Optionally, the determination module 705 is also configured as follows: In response to a respiratory rate that is not within the preset frequency range, a specific detection target is identified as a living target but not a human target.

[0090] The in-vehicle target detection device according to embodiments of the present invention can solve the problems of in-vehicle detection being easily affected by environmental and occlusion interference, difficulty in accurately distinguishing between living and non-living objects, frequent false positives and false negatives, and lack of privacy and security. It can achieve accurate identification and positioning of in-vehicle targets in all weather and all areas, effectively filter various interference factors, eliminate detection false positives and false negatives caused by changes in light, occlusion, heat source interference and abnormal posture, and improve the accuracy and stability of detection.

[0091] refer to Figure 8 The diagram below is a block diagram of an electronic device according to some embodiments of the present invention. It illustrates a more specific hardware structure of an electronic device provided in this application embodiment. The device may include: a processor 810, a memory 820, an input / output interface 830, a communication interface 840, and a bus 850. The processor 810, memory 820, input / output interface 830, and communication interface 840 are interconnected internally via the bus 850.

[0092] The processor 810 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0093] The memory 820 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 820 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program code is stored in the memory 820 and is called and executed by the processor 810.

[0094] The input / output interface 830 is used to connect input / output modules to enable information input and output. Input / output modules can be configured as components within the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touchscreens, microphones, various sensors, etc., while output devices may include displays, speakers, vibrators, indicator lights, etc.

[0095] The communication interface 840 is used to connect the communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0096] Bus 850 includes a pathway for transmitting information between various components of the device, such as processor 810, memory 820, input / output interface 830, and communication interface 840.

[0097] It should be noted that although the above-described device only shows the processor 810, memory 820, input / output interface 830, communication interface 840, and bus 850, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.

[0098] The electronic devices described above are used to implement the corresponding in-vehicle target detection methods in any of the foregoing embodiments, and have the beneficial effects of the corresponding in-vehicle target detection method embodiments, which will not be repeated here.

[0099] Based on the same concept, corresponding to the in-vehicle target detection method provided in any of the above embodiments, this application also provides a computer-readable storage medium storing a program or instructions, which, when executed by a processor, implements the in-vehicle target detection method as described in the first aspect.

[0100] The aforementioned computer-readable storage medium can be any available medium or data storage device that a computer can access, including but not limited to magnetic storage (e.g., floppy disks, hard disks, magnetic tapes, magneto-optical disks (MOs), etc.), optical storage (e.g., CDs, DVDs, BDs, HVDs, etc.), and semiconductor storage (e.g., ROMs, EPROMs, EEPROMs, non-volatile memory (NAND flash), solid-state drives (SSDs)).

[0101] The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to execute the corresponding in-vehicle target detection method in any of the foregoing embodiments, and have the beneficial effects of the corresponding in-vehicle target detection method embodiments, which will not be repeated here.

[0102] Based on the same inventive concept, corresponding to the in-vehicle target detection method described in any of the above embodiments, this application also provides a computer program product, which includes computer program instructions. In some embodiments, the computer program instructions can be executed by one or more processors of a computer to cause the computer and / or the processor to perform the in-vehicle target detection method. Corresponding to the execution entity for each step in each embodiment of the in-vehicle target detection method, the processor executing the corresponding step can belong to the corresponding execution entity.

[0103] The computer program product of the above embodiments is used to cause the computer and / or the processor to execute the in-vehicle target detection method as described in any of the above embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0104] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0105] From the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of computer software products plus necessary general-purpose hardware platforms, and of course, they can also be implemented by hardware. The computer software product is stored in a storage medium (such as ROM, RAM, magnetic disk, optical disk, etc.) and includes several instructions to cause the terminal or network-side device to execute the methods described in the various embodiments of this application.

[0106] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other implementations under the guidance of this application without departing from the spirit and scope of the claims. All of these implementations are within the protection scope of this application.

Claims

1. A method of in-vehicle object detection, characterized by, include: Acquire first point cloud data of the vehicle interior space, wherein the first point cloud data does not include point cloud data of the vehicle body structure; The first point cloud data is converted into the in-vehicle coordinate system of the in-vehicle space to obtain the second point cloud data; In response to the number of second point cloud data included in the first preset area of ​​the vehicle interior space reaching a preset number, the point cloud density of the second point cloud data included in the second preset area of ​​the vehicle interior space is obtained, wherein the second preset area is set inside the first preset area and the volume of the second preset area is smaller than the volume of the first preset area; Based on the point cloud density, determine whether there is a high-density point cloud band in the vehicle interior space that meets the preset judgment rules; In response to the presence of a high-density point cloud band in the vehicle interior space, it is determined that a specific detection target exists in the vehicle interior space.

2. The in-vehicle target detection method according to claim 1, characterized by, The acquisition of the first point cloud data of the vehicle interior space includes: Acquire raw point cloud data obtained by scanning the interior space of the vehicle with radar, wherein the raw point cloud data includes the three-dimensional spatial coordinates of each point in the interior space and the point cloud reflection intensity; The first point cloud data is obtained by removing the point cloud data of the vehicle body structure and the point cloud data whose Euclidean distance from other original point cloud data exceeds a preset threshold from the original point cloud data using a preset point cloud matching template.

3. The in-vehicle target detection method according to claim 2, characterized by, The step of converting the first point cloud data to the in-vehicle coordinate system of the vehicle interior space to obtain the second point cloud data includes: Determine the rotation matrix used to correct the radar installation angle deviation and the translation vector used to correct the radar installation position deviation; The second point cloud data is obtained based on the first point cloud data, the rotation matrix, and the translation vector.

4. The in-vehicle target detection method according to claim 1, characterized in that, The method further includes: Based on the seating layout of the vehicle interior space, the first preset area is divided for each seating area; Each of the first preset areas is divided into several second preset areas along the longitudinal direction of the vehicle interior space.

5. The in-vehicle target detection method according to claim 4, characterized in that, The step of determining whether a high-density point cloud band satisfying a preset judgment rule exists in the vehicle interior space based on the point cloud density includes: Traverse all the second preset regions along the longitudinal direction of the vehicle interior space, and calculate the point cloud density of the second point cloud data included in each second preset region; In response to the point cloud density of at least a preset number of consecutive adjacent second preset regions exceeding a preset density threshold, it is determined that a high-density point cloud band exists in the vehicle interior space.

6. The in-vehicle target detection method according to claim 1, characterized in that, The method further includes: Extract the 1D signal time series corresponding to the point cloud of the high-density point cloud band, and filter the 1D signal time series to obtain the target signal sequence; Extract the respiratory micro-motion signal from the target signal sequence, and determine whether the respiratory frequency represented by the respiratory micro-motion signal is within a preset frequency range; In response to the respiratory rate being within the preset frequency range, the specific detection target is determined to be a human target.

7. The in-vehicle target detection method according to claim 1, characterized in that, The method further includes: In response to the number of second point cloud data included in the first preset area of ​​the vehicle interior space reaching a preset number, it is determined that a detection target exists in the vehicle interior space, wherein the detection target is a human target or a non-human target.

8. The in-vehicle target detection method according to claim 7, characterized in that, The method further includes: If the number of second point cloud data included in the first preset area of ​​the vehicle interior space does not reach the preset number, it is determined that the detection target does not exist in the vehicle interior space.

9. The in-vehicle target detection method according to claim 6, characterized in that, The method further includes: In response to the respiratory rate not being within the preset frequency range, the specific detection target is determined to be a living target but not a human target.

10. An in-vehicle target detection device, characterized in that, include: The acquisition module is configured to acquire first point cloud data of the in-vehicle space, wherein the first point cloud data does not include point cloud data of the vehicle body structure; The conversion module is configured to convert the first point cloud data to the in-vehicle coordinate system of the in-vehicle space to obtain the second point cloud data; The response module is configured to, in response to the number of the second point cloud data included in the first preset area of ​​the vehicle interior space reaching a preset number, obtain the point cloud density of the second point cloud data included in the second preset area of ​​the vehicle interior space, wherein the second preset area is set inside the first preset area and the volume of the second preset area is smaller than the volume of the first preset area. The determination module is configured to determine whether there is a high-density point cloud band that meets a preset judgment rule in the vehicle interior space based on the point cloud density. The determination module is configured to determine the presence of a specific detection target in the vehicle interior space in response to the presence of a high-density point cloud band in the vehicle interior space.

11. An electronic device, characterized in that, include: A processor and a memory, the memory storing a program or instructions executable on the processor, the program or instructions, when executed by the processor, implementing the steps of the in-vehicle target detection method as described in any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the in-vehicle target detection method as described in any one of claims 1 to 9.

13. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the in-vehicle target detection method as described in any one of claims 1 to 9.

14. A vehicle, characterized in that, The vehicle includes the electronic equipment as described in claim 11.

Images