A parking space recognition method, device and vehicle

Abstract

This application provides a parking space identification method, apparatus, and vehicle. The method includes: associating all identified parking spaces in the current frame with tracked parking spaces in a parking space list to obtain an association result; updating the location information of the tracked parking spaces based on the association result; updating the status of the tracked parking spaces based on the updated location information and vehicle position; and updating and outputting an updated parking space list based on the updated location information and status of the tracked parking spaces. This method effectively filters out false detections or temporary loss of real parking spaces caused by transient interference, resulting in more stable and reliable output results, improving the location accuracy of parking spaces, and providing more reliable input for parking path planning and control.

Description

Technical Field

[0001] This application relates to the field of intelligent parking technology, and in particular to a parking space recognition method, device and vehicle. Background Technology

[0002] With the rapid development of intelligent driving technology, automatic parking systems have become one of the core functions for improving parking efficiency and safety. This system relies on accurate perception of the surrounding environment, especially available parking spaces, and the stability and accuracy of parking space recognition technology directly determine the parking success rate and user experience. However, existing solutions mostly rely on single-frame image data for parking space detection and recognition, which is susceptible to instantaneous environmental interference, such as sudden changes in lighting, partial occlusion, and sensor noise. For example, falling leaves or pedestrians in adjacent frames may cause intermittent detection results for the same parking space, leading to low stability and accuracy of parking space recognition results, thus affecting the stability and reliability of the system. Summary of the Invention

[0003] This application provides a parking space identification method, device, and vehicle to solve the above-mentioned technical problems.

[0004] A parking space identification method includes: associating all identified parking spaces in the current frame with tracked parking spaces in a parking space list to obtain an association result; updating the location information of the tracked parking spaces based on the association result; updating the status of the tracked parking spaces based on the updated location information and vehicle location; and updating and outputting an updated parking space list based on the updated location information and status of the tracked parking spaces.

[0005] Further, the step of associating all identified current frame parking spaces with the tracked parking spaces in the parking space list to obtain an association result includes: calculating the intersection-union ratio (IUR) of the current frame parking space contour and the tracked parking space contour based on the parking space contour of the current frame parking space and the parking space contour of the tracked parking space; if the IUR is greater than a preset IUR threshold, then the current frame parking space and the tracked parking space are successfully associated; otherwise, the association is unsuccessful.

[0006] Further, updating the location information of the tracked parking space based on the association result includes: if the current frame parking space and the tracked parking space are successfully associated, then the location information of the associated tracked parking space is updated using the location information of the current frame parking space; if the tracked parking space is not successfully associated, then the location information of the unassociated tracked parking space is updated by dead reckoning; if the current frame parking space is not successfully associated, then the location information of the current frame parking space is updated and added to the parking space list.

[0007] Furthermore, updating the status of the tracking parking space based on the updated location information of the tracking parking space and the vehicle location includes: the location information includes the center position of the parking space, and the status of the tracking parking space includes a test status, a confirmed status, an extrapolation status, and a terminated status; when the center position of the tracking parking space in the test status is located behind the rear axle center of the vehicle, the status of the tracking parking space is adjusted to the confirmed status; when the nearest distance between the center position of the tracking parking space in the confirmed status or the extrapolation status and the vehicle exceeds a preset distance threshold, the status of the tracking parking space is adjusted to the terminated status; when the tracking parking space in the confirmed status is not associated or overlaps with the vehicle, the status of the tracking parking space is adjusted to the extrapolation status.

[0008] Furthermore, the method for identifying the parking space in the current frame is as follows: detecting and identifying obstacles and obtaining the detection information of the obstacles; identifying and removing abnormal obstacles to obtain normal obstacles; correcting the detection information of the normal obstacles to obtain corrected detection information of the normal obstacles; and constructing the parking space in the current frame based on two adjacent normal obstacles and their corrected detection information.

[0009] Further, the step of constructing the current frame parking space based on two adjacent normal obstacles and their correction detection information includes: selecting available obstacles from the normal obstacles for constructing the current frame parking space; determining the type of the current frame parking space based on the correction detection information of two adjacent available obstacles, and obtaining the initial corner point of the current frame parking space; and generating the parking space outline of the current frame parking space based on the initial corner point.

[0010] Further, generating the parking space outline of the current frame parking space based on the initial corner point includes: generating an initial outline of the current frame parking space based on the initial corner point; if the size of the initial outline differs from the size of a preset standard parking space by more than a preset size threshold, then adjusting the initial corner point to obtain an optimized corner point based on the size of the preset standard parking space; and generating the parking space outline of the current frame parking space based on the optimized corner point.

[0011] Further, generating the parking space contour of the current frame parking space based on the initial corner point includes: generating an initial contour of the current frame parking space based on the initial corner point; if the initial contour overlaps with the contour of the available obstacle, determining the coordinates of the deepest vertex of the available obstacle extending into the initial contour; adjusting the initial corner point based on the coordinates of the deepest vertex to obtain an optimized corner point; and generating the parking space contour of the current frame parking space based on the optimized corner point.

[0012] Based on the same technical concept, an in-vehicle device is designed, including a processor and a memory storing computer programs or instructions that can run on the processor, wherein the computer programs or instructions are executed by the processor to implement the above-mentioned parking space recognition method.

[0013] Based on the same technical concept, a vehicle is designed, including the aforementioned vehicle-mounted device.

[0014] Compared with the prior art, the beneficial effects of this application are as follows: In this scheme, the parking spaces identified in the current frame are dynamically associated and matched with the historically maintained list of tracked parking spaces, eliminating the need to consider the identification results of each frame in isolation. Even if a parking space is not identified or is incorrectly identified in a frame due to changes in lighting, temporary occlusion (such as pedestrians or other vehicles passing by), or sensor noise, it can still be associated with the correct identification results in subsequent frames based on historical tracking information. This avoids intermittent or abrupt changes in the perception of parking spaces, ensuring continuous perception of their existence. The scheme maintains the status of each tracked parking space and updates the status based on the association results and vehicle position, effectively filtering out false detections or temporary loss of true parking spaces caused by transient interference, resulting in more stable and reliable output results. Furthermore, the location information of the parking spaces is updated based on the association results between the current frame and the tracked parking spaces, improving the location accuracy of the parking spaces and providing more reliable input for parking path planning and control. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the vertical parking space described in this application.

[0016] Figure 2 This is a schematic diagram of the horizontal parking space described in this application.

[0017] Figure 3 This is a schematic diagram of the inclined train position described in this application.

[0018] Figure 4 This is a flowchart illustrating the parking space recognition method described in this application.

[0019] Figure 5 This is a schematic diagram of the current frame parking space detection and recognition process described in this application.

[0020] Figure 6 This is a schematic diagram of the vehicle body coordinate system described in this application.

[0021] Figure 7 This is a schematic diagram of the process described in this application for constructing the parking space of the current frame based on two adjacent normal obstacles and their corrected detection information.

[0022] Figure 8 This is a schematic diagram showing the corner points of the obstacles described in this application arranged according to a uniform pattern.

[0023] Figure 9 This is a schematic diagram illustrating the dimensional adjustment of the initial outline of the parking space as described in this application.

[0024] Figure 10 This is a schematic diagram showing the overlap between the outline of the parking space described in this application and the obstacles that constitute it.

[0025] Figure 11 This is a schematic diagram illustrating the contour correction of the initial outline of the parking space as described in this application.

[0026] Figure 12 This is a schematic diagram of the vehicle-mounted device described in this application.

[0027] Figure 13 This is a schematic diagram illustrating an example of the arrangement of an ultrasonic radar array on a vehicle as described in this application.

[0028] Figure 14 This is a schematic diagram of the vehicle functional framework described in this application. Detailed Implementation

[0029] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0030] To better understand the solutions in the embodiments of this application, some terms involved in the embodiments of this application will be explained below.

[0031] An ultrasonic radar sensor array is a collaborative sensor network composed of multiple ultrasonic radar sensors (typically 6-16) installed at specific locations around the vehicle, arranged in a systematic layout. Ultrasonic radar is a sensor used to detect surrounding obstacles. It measures information such as distance, angle, and shape between the vehicle and surrounding objects by emitting ultrasonic waves and receiving echo signals. The sensor positions are carefully designed to ensure that their detection ranges overlap and cover each other, eliminating blind spots for individual sensors and forming a complete protective network and sensing field around the vehicle. All sensors are centrally controlled and can operate alternately according to a specific time sequence to avoid interference between their acoustic signals.

[0032] Perpendicular parking spaces are those where the angle between the obstacle and the vehicle being parked is close to 90 degrees, such as... Figure 1 Examples are shown in the text.

[0033] A level parking space is a parking space where the angle between the obstacle and the vehicle is close to 0 degrees, such as... Figure 2Examples are shown in the text.

[0034] Angled parking spaces, where the angle between the obstacle and the vehicle is 30, 45, or 60 degrees, such as... Figure 3 Examples are shown in the text.

[0035] IOU, short for Intersection over Union, is a core indicator for measuring the degree of overlap between two regions. It is the ratio of the area of ​​the overlapping region to the total area of ​​the merged region. The value ranges from [0, 1], where 0 indicates that the two regions do not overlap at all, 1 indicates that the two regions completely overlap, and (0, 1) indicates that they partially overlap. The larger the value, the higher the degree of overlap.

[0036] With the rapid development of intelligent driving technology, automatic parking systems have become one of the core functions for improving parking efficiency and safety. This system relies on accurate perception of the surrounding environment, especially available parking spaces, and the stability and accuracy of parking space recognition technology directly determine the parking success rate and user experience. Currently, mainstream parking space recognition technologies include visual recognition and radar recognition. Visual recognition technology relies on clear parking lines, making it difficult to work effectively in poor lighting conditions or when parking lines are blurred, and it cannot recognize unmarked parking spaces. Radar recognition technology is mostly based on single-frame image data for parking space detection and recognition, making it susceptible to instantaneous environmental interference, such as sudden changes in lighting, partial occlusion, and sensor noise. For example, falling leaves or pedestrians in adjacent frames may cause intermittent detection results for the same parking space, leading to low stability and accuracy in parking space recognition. Furthermore, it performs poorly in recognizing angled parking spaces, affecting the system's stability and reliability. Therefore, there is an urgent need for a new parking space recognition solution that can accurately identify various types of parking spaces, especially angled parking spaces, and has anti-interference capabilities. Example 1

[0037] like Figure 4 As shown, in order to solve the above-mentioned technical problems, this application proposes a parking space recognition method, which specifically includes the following steps S10 to S40.

[0038] S10: Associate all identified parking spaces in the current frame with the tracked parking spaces in the parking space list to obtain the association result.

[0039] S20, based on the association result, update the location information of the tracked parking space.

[0040] S30, based on the updated location information of the tracked parking space and the vehicle location, update the status of the tracked parking space.

[0041] S40 updates and outputs the updated parking space list based on the updated location information and status of the tracked parking spaces.

[0042] This method dynamically correlates and matches the parking spaces identified in the current frame with the historically maintained list of tracked parking spaces, rather than considering the identification results of each frame in isolation. Even if a parking space is not identified or is incorrectly identified in a frame due to changes in lighting, temporary occlusion (such as pedestrians or other vehicles passing by), or sensor noise, it can still be correlated with the correct identification results of subsequent frames based on historical tracking information. This avoids intermittent or abrupt changes in parking space perception, ensuring continuous perception of parking space presence. The method maintains the state of each tracked parking space and dynamically updates the state based on the correlation results and vehicle positions, effectively filtering out false detections or temporary loss of true parking spaces caused by transient interference, resulting in more stable and reliable output results. Furthermore, the location information of the parking spaces is updated based on the correlation results between the current frame and the tracked parking spaces, improving the location accuracy of the parking spaces and providing more reliable input for parking path planning and control.

[0043] It should be noted that the parking space recognition in this embodiment is based on ultrasonic radar installed on the vehicle. Of course, in other possible embodiments, other types of radar, such as lidar, millimeter-wave radar, or combinations thereof, may also be used.

[0044] In step S10, both the parking spaces in the current frame and the tracked parking spaces in the parking space list are detected and identified based on ultrasonic radar recognition technology. The parking spaces in the current frame are constructed based on the ultrasonic radar data at the current moment (the latest time). Tracked parking spaces are those constructed in the previous frame, which were identified and tracked in the previous sampling period.

[0045] Therefore, before step S10, it is also necessary to detect and identify the parking space in the current frame, such as... Figure 5 As shown, the specific method includes the following steps.

[0046] S01, detect and identify obstacles and obtain obstacle detection information.

[0047] S02, identify and remove abnormal obstacles to obtain normal obstacles.

[0048] S03, correct the detection information of normal obstacles to obtain corrected detection information of normal obstacles.

[0049] S04, construct the parking space of the current frame based on two adjacent normal obstacles and their corrected detection information.

[0050] In step S01, data is first acquired and preprocessed using an ultrasonic radar sensor array, and then obstacle detection is performed using a deep learning model, as detailed below: An ultrasonic radar sensor array emits ultrasonic pulses at a fixed frequency (typically 10-20Hz) and receives the echoes. The distance to an obstacle is determined by calculating the round-trip time of the sound waves, and the polar coordinate data is converted to a vehicle coordinate system with the rear axle center as the origin (e.g., ...). Figure 6 The discrete point cloud data (shown in the image) is converted into a continuous bird's-eye view image using Cartesian coordinates. A pre-trained deep learning model (such as a convolutional neural network) is used to detect obstacles in the bird's-eye view. The model can identify various obstacles in the image, such as vehicles, pillars, walls, and curbs, and outputs the rotated bounding box information for each obstacle, including corner coordinates, center coordinates, length, width, and orientation angle, along with category labels and detection confidence. The raw detection results output by the model are converted into standardized structured data to form obstacle detection information. The detection information for each obstacle includes: geometric features, such as center position, size, orientation angle, and coordinates of the four corners; attribute features, such as category, height attribute, and static / dynamic label; confidence assessments, such as detection confidence, classification confidence, and positional uncertainty; and timestamps and unique identifiers.

[0051] It should be noted that, as Figure 6 As shown, in the vehicle body coordinate system with the rear axle center as the origin, the X-axis points to the right side of the vehicle, and the Y-axis points to the direction of vehicle movement.

[0052] In step S02, the detected obstacles are screened for quality, and unreasonable or unreliable detection results are removed to ensure that only high-quality obstacle information enters the subsequent parking space recognition process, thereby improving the accuracy of parking space detection and recognition. The specific steps are as follows: Abnormal obstacle filtering is performed using confidence levels. First, a basic detection confidence threshold is applied to filter out obviously unreliable detections. Second, differentiated detection confidence thresholds are set for different obstacle categories; for example, vehicle obstacles require higher classification confidence. Finally, the detection confidence threshold is adjusted based on the distance between the obstacle and the vehicle. Obstacles with confidence levels below the detection confidence threshold are marked as abnormal obstacles. It should be noted that the detection confidence thresholds for different obstacle categories are set according to actual needs, and specific requirements are not specified here.

[0053] Simultaneously, it is necessary to verify whether the obstacle's location conforms to physical and logical constraints. First, check whether the obstacle spatially overlaps with the vehicle; overlapping obstacles are likely sensor noise or false detections. Second, verify whether the obstacle is within the sensor's effective detection range, eliminating detections that exceed physical limits. Next, assess the reasonableness of the obstacle's height, excluding unrealistic detections located below ground level or excessively high (e.g., exceeding 3 meters).

[0054] In step S03, corrections are made using the detection information of normal obstacles to optimize the accuracy and stability of the obstacle detection information and improve data quality. The correction process is as follows: For geometric features in the detection information, such as center position, size, orientation angle, and coordinates of the four corner points, α-β filtering can be performed using the prediction information from the current frame detection information and the previous frame detection information (i.e., Where X represents the estimated information for the current frame. For the detection information of the current frame, The method updates the information based on the prediction information of the previous frame (α and β are weights). For updating high and low attributes, the attribute that appears most frequently in the obstacle detection history is taken as the final high or low attribute. For example, if a detected object is judged as high in 8 out of 10 consecutive frames and as low in 2 frames, it is finally confirmed as a high attribute. This solves the jitter problem of single-frame detection and improves the stability of attribute judgment.

[0055] like Figure 7 As shown, in step S04, the parking space of the current frame is constructed based on two adjacent normal obstacles and their corrected detection information, specifically including the following steps S041 to S043.

[0056] S041, Filter from normal obstacles to find available obstacles to build the parking space in the current frame.

[0057] S042, based on the corrected detection information of two adjacent available obstacles, determine the type of parking space in the current frame and obtain the initial corner point of the parking space in the current frame.

[0058] S043, Based on the initial corner points, generate the parking space outline of the parking space in the current frame.

[0059] In step S041, obstacle pairs that could potentially form parking spaces are searched from all normal obstacles. Not every pair of obstacles needs to be searched for and constructed. First, the distance between obstacles is calculated, and candidate pairs with distances within a reasonable parking space width are selected. If there are other obstacles between two obstacles or the distance between two obstacles is too large, then constructing a parking space is not allowed. Secondly, obstacles with low attributes are also not allowed to participate in constructing parking spaces.

[0060] In step S042, the parking space type is determined based on the corrected detection information, i.e., geometric features, of the paired available obstacles. The average orientation angle of the two obstacles is calculated, and the parking space is classified according to the angle value, including perpendicular parking spaces, horizontal parking spaces, and angled parking spaces. It should be noted that when the angles of the two obstacles are inconsistent, the angle of the obstacle with a longer detection lifespan can be used as the parking space angle because the outline of the obstacle with a longer detection lifespan is clearer and more complete.

[0061] After determining the parking space type, for perpendicular and angled parking spaces, the minimum distance between the two obstacles in the width direction is calculated. If the minimum distance is less than a preset parking space width threshold, such as 2.2m, the parking space is not constructed. For horizontal parking spaces, the minimum distance between the two obstacles in the length direction is calculated. If the minimum distance is less than a preset parking space length threshold, such as 4.8m, the parking space is not constructed. It should be noted that the preset parking space width and length thresholds are determined based on actual conditions and are not specifically limited here.

[0062] When determining the initial corner points of the parking space in the current frame, all obstacle corner points involved in the parking space construction are first sorted according to a uniform rule to reduce the logical complexity of subsequent modules. Then, the initial corner points of the parking space are calculated based on the principle of minimizing the area of ​​the shape formed between two obstacle corner points. Specifically: like Figure 8 As shown, the four corner points of each obstacle are sorted according to a uniform rule, specifically: 0 point: The corner point with the smallest x and y values ​​(closest to the origin); Point 1: The corner point closest to point 0; Point 2: The corner point farthest from point 0; 3 points: the remaining corner point.

[0063] For perpendicular or angled parking spaces, calculate the area between the long sides of the two obstacles, and take the four vertices that form the minimum area as the initial corner points of the constructed parking space (as shown in the attached diagram). Figure 8 In the diagram, (OD1_0, OD1_3, OD2_2, OD2_1) are the initial corner points. For a horizontal parking space, the area between the short sides of the obstacle is calculated, and the four vertices that make up the smallest area are taken as the initial corner points of the constructed parking space.

[0064] In step S043, the parking space outline of the current frame is generated based on the initial corner point, which is to transform the initial corner point into a regular, practical, and standard parking space outline.

[0065] First, based on the initial corner points, the initial outline of the parking space in the current frame is generated. A computational geometry algorithm is used to fit the four initial corner points into a rotated rectangle. For example, the `minAreaRect` function from the OpenCV library is used to convert an irregular quadrilateral into a regular rectangular parking space bounding box that can be used for subsequent judgment. The OpenCV library function is then used to fit the initial corner points into an initial rectangle with the smallest area, i.e., the initial outline of the parking space, and can output the center coordinates, length, width, and orientation angle of the rectangle.

[0066] Secondly, when the initial outline of the parking space differs significantly from the standard parking space dimensions, size adjustments are necessary, as follows: If the difference between the initial outline size and the preset standard parking space size exceeds a preset size threshold, the initial corner points are adjusted based on the preset standard parking space size to obtain optimized corner points. Based on the optimized corner points, the parking space outline for the current frame is generated. The preset standard parking space size and preset size threshold are determined according to actual conditions. Generally, the preset standard parking space size is 5.3m long and 2.4m wide, and the preset size threshold is 0.3m. The parking space size correction steps are as follows: like Figure 9 As shown, if the original parking space outline points are (A, B, C, D), the parking space length that needs to be adjusted is length (i.e., the length of the preset standard parking space), and the adjusted parking space outline points are (A, B', C', D). Of course, in other possible embodiments, the width of the preset standard parking space can also be used for correction.

[0067] Assume the coordinates of parking space A are The coordinates of point B are The coordinates of point C are The coordinates of point D are The adjusted coordinates of point B' are: The coordinates of point C' are The coordinates of points B' and C' are determined using the vector method. Then, vector for:

[0068] The length of AB is:

[0069] The unit vector in the AB direction is:

[0070] Starting from point A, travel along the u direction for a length length to reach point B'. The coordinates of point B' are:

[0071]

[0072] Similarly, the coordinates of point C' can be obtained as follows:

[0073]

[0074] The length between points C and D is:

[0075] In addition, such as Figure 10As shown, when the angles of the two obstacles involved in the parking space construction differ significantly, the initial parking space outline generated using the minAreaRect function may be over-expanded, causing the parking space outline to overlap with the obstacles that constitute it. Therefore, overlap correction is required, as follows: Determine the coordinates of the deepest vertex into which an available obstacle can extend within the initial contour; based on the coordinates of the deepest vertex, adjust the initial corner points to obtain optimized corner points; based on the optimized corner points, generate the parking space contour for the current frame. The specific steps for correcting obstacle and parking space overlap are as follows: Detect and calculate the IOU between the obstacle and the parking space. When the IOU is greater than 0, obtain the coordinates O of the deepest overlapping vertex. .

[0076] The straight-line distance from point O to line AB can be calculated using the vector method. The straight-line distance from point O to line AB is: If the straight-line distance from point O to line AB is less than the width of the preset standard parking space, for example, 2.4m, then the parking space is cleared. Otherwise, geometric correction is performed.

[0077] For example, such as Figure 11 As shown, if the original parking space outline points are (A, B, C, D), the parking space orientation angle (i.e., the angle between the long side of the parking space and the positive y-axis) is: There exists a vehicle and a parking space that overlap, with the deepest overlap vertex being O. The adjusted parking space outline points are (A, B, C', D'). Geometric principles are then used to determine the coordinates of C' and D'.

[0078] The slope of the equation of line C'D' is... The intercept is ; The coordinates of point C' are:

[0079]

[0080] The coordinates of point D' are:

[0081]

[0082] in, , , , Let AD and BC be the slopes and intercepts, respectively. The coordinates are determined using geometric principles and existing mathematical techniques; the specific calculation process will not be detailed here.

[0083] It should be noted that the above method can obtain the parking space's location information, including the center position, dimensions, orientation angle, and coordinates of the four corner points. In step S10, all identified parking spaces in the current frame are associated with the tracked parking spaces in the parking space list to obtain an association result. This association confirmation maintains the stability of the parking space ID and the continuity of the trajectory. Specifically, this includes: calculating the intersection-union ratio (IUR) of the parking space contours of the current frame parking space and the tracked parking space based on their contours; if the IUR is greater than a preset IUR threshold, the current frame parking space and the tracked parking space are successfully associated; otherwise, the association is unsuccessful.

[0084] It should be noted that the parking space list is a dynamically maintained data structure used to store and manage real-time and historical information of all parking spaces being tracked in the system. It includes: A unique identifier (ID) uniquely identifies a parking space throughout the entire identification and detection lifecycle. It is the foundation for parking space tracking and ensures that the same physical parking space has the same ID in consecutive frames.

[0085] The parking space location information includes the center location of the parking space, the size of the parking space, the direction angle, and the coordinates of the four corner points.

[0086] Parking space status directly determines the reliability and availability of parking space information, including: STATUS_NULL (Invalid status): Indicates that the parking space information is invalid and cannot be used for parking decisions. STATUS_TEST (Test Status): Indicates that the parking space is a potential parking space, but has not yet been confirmed as a real parking space; STATUS_UPDATED (Confirmed Status): Indicates that the parking space has been confirmed as a real parking space and can be used for parking decisions. STATUS_COASTED (Extrapolation Status): Indicates that current sensor information is temporarily missing. The system extrapolates parking space information based on historical information, which can be used for parking decision-making. STATUS_TERMINATED (Termination Status): Indicates that the parking space has been determined to be invalid or deactivated and is no longer used for parking decisions.

[0087] When confirming the association, the IOU (Intersection over Union) of all parking space outlines detected in the current frame is calculated and matched with the IOU of all parking space rows in the parking space list that are in the STATUS_UPDATED and STATUS_COASTED states. If the IOU value exceeds the preset IOU threshold, such as 0.6, the data association is considered successful and the two are the same parking space; otherwise, it is considered a new parking space.

[0088] In step S20, the location information of the tracked parking space is updated based on the association result, including: If the parking space in the current frame and the tracked parking space are successfully associated, the location information of the associated tracked parking space will be updated using the location information of the parking space in the current frame.

[0089] If the tracking of a parking space fails to be successfully associated, the location information of the unassociated tracking parking space will be updated using dead reckoning.

[0090] If the parking space in the current frame is not successfully associated, the location information of the parking space in the current frame will be updated and added to the parking space list.

[0091] Specifically, for successfully matched tracking parking spaces, the space dimensions (length and width) and orientation angle can be updated using the dimensions (length and width) and orientation angle of the matched parking space in the current frame, such as through mean filtering or Kalman filtering. The length, width, and orientation angle of the parking space are key parameters describing its geometry. Single-frame detection may be affected by noise, causing jitter. Filtering updates smooth these parameters using historical data from multiple frames, suppressing random noise. The center and corner positions of the parking space can be fused and updated using the center and corner positions obtained through Dead Reckoning (DR) with those detected in the current frame, such as through α-β filtering or Kalman filtering, to obtain optimal estimates of the parking space's corner and center positions. DR DR extrapolation predicts the position of the parking space in the vehicle coordinate system in the current frame based on the parking space position in the previous frame and the vehicle's own motion information (such as vehicle speed and yaw rate) using a motion model (such as a constant velocity model). DR extrapolation is essentially a coordinate transformation. Since parking spaces are fixed objects on the ground, their coordinates in the vehicle coordinate system change when the vehicle moves. Assuming the vehicle's movement is known, the parking space coordinates from the previous frame can be transformed into the vehicle coordinate system of the current frame to obtain the predicted position.

[0092] For unmatched tracking parking spaces, meaning a tracking parking space in the parking space list does not match any detected parking spaces in the current frame (possibly due to occlusion, sensor miss, or the parking space disappearing), the system lacks parking space data for the current frame and can only rely on historical information for state maintenance. In this case, DR extrapolation is used to update the parking space information. That is, based on the vehicle's own motion information, the parking space is transformed from the vehicle coordinate system of the previous frame to the vehicle coordinate system of the current frame, thereby updating the coordinates of the center position and corner points of the parking space. The size of the parking space can remain unchanged.

[0093] In step S30, based on the updated location information of the tracked parking spaces and the vehicle's location, the state of the tracked parking spaces is updated to address the decision-making problem regarding the credibility of parking spaces—that is, when to believe a parking space is real and when to discard it. A state is introduced for each parking space, and its validity and availability are dynamically managed based on its location and interaction with the vehicle. Specifically, this includes the following: When the center of the tracked parking space in the test state is located behind the center of the vehicle's rear axle, the status of the tracked parking space is adjusted to the confirmed state.

[0094] When the nearest distance between the center of a tracked parking space and a vehicle exceeds a preset distance threshold while the tracking space is in the confirmed or extrapolated state, the tracking status of the parking space is changed to the terminated state, and it is removed from the parking space list. The preset distance threshold is set according to the actual situation, such as 15m.

[0095] For example, when a parking space is in front of your car, the shortest distance between the center of the parking space and the car is the distance from the center of the parking space to the front bumper of the car; when a parking space is behind your car, the shortest distance between the center of the parking space and the car is the distance from the center of the parking space to the rear bumper of the car; when a parking space is directly to the side of your car, the shortest distance between the center of the parking space and the car is the distance from the center of the parking space to the side door of the car.

[0096] When a tracked parking space in the confirmed state is not associated or overlaps with a vehicle, the status of the tracked parking space will be adjusted to the extrapolation state.

[0097] In step S40, based on the updated location information and status of the tracked parking spaces, an updated parking space list is updated and output. The updated parking space list contains all available parking spaces. By selecting the optimal parking space, automatic parking path planning can be performed to execute the automatic parking function.

[0098] Example 2 like Figure 12 As shown, this embodiment provides a vehicle-mounted device 10, including a processor 110, an ultrasonic radar sensor array 120, a display screen 130, an audio module 140, a memory 150, a mobile communication module 160, a wireless communication module 170, an antenna 1, and an antenna 2.

[0099] It is understood that the structures illustrated in the embodiments of the present invention do not constitute a specific limitation on the vehicle-mounted device. In other embodiments of this application, the vehicle-mounted device may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0100] Processor 110 may include one or more processing units, such as: application processor (AP), modem processor, graphics processing unit (GPU), image signal processor (ISP), controller, video codec, digital signal processor (DSP), baseband processor, and / or neural network processing unit (NPU), etc. Different processing units may be independent devices or integrated into one or more processors.

[0101] In some embodiments, the processor 110 may include one or more interfaces. These interfaces may include an Inter-integrated Circuit (I2C) interface, an Inter-integrated Circuit Sound (I2S) interface, a Pulse Code Modulation (PCM) interface, a Universal Asynchronous Receiver / Transmitter (UART) interface, a Mobile Industry Processor Interface (MIPI) interface, a General-Purpose Input / Output (GPIO) interface, etc.

[0102] The I2C interface is a bidirectional synchronous serial bus, including a serial data line (SDL) and a serial clock line (SCL). In some embodiments, the processor 110 may include multiple I2C buses. The processor 110 can couple to the ultrasonic radar sensor array 120, the display screen 130, etc., through different I2C bus interfaces. For example, the processor 110 can couple to the ultrasonic radar sensor array 120 through the I2C interface, enabling the processor 110 and the ultrasonic radar sensor array 120 to communicate through the I2C bus interface, thereby realizing the radar detection function of the vehicle-mounted device.

[0103] The I2S interface can be used for audio communication. In some embodiments, the processor 110 may include multiple I2S buses. The processor 110 can be coupled to the audio module 140 via the I2S bus to realize communication between the processor 110 and the audio module 140. In some embodiments, the audio module 140 can transmit audio signals to the wireless communication module 170 via the I2S interface to realize the function of answering telephone calls.

[0104] The MIPI interface can be used to connect the processor 110 to peripheral devices such as the display screen 130 and the camera 193. The MIPI interface includes a Display Serial Interface (DSI). In some embodiments, the processor 110 and the display screen 130 communicate via the DSI interface to realize the display function of the vehicle-mounted device.

[0105] The GPIO interface is configurable via software. It can be configured as a control signal or a data signal. In some embodiments, the GPIO interface can be used to connect the processor 110 to the ultrasonic radar sensor array 120, the display screen 130, the audio module 140, the wireless communication module 170, etc. The GPIO interface can also be configured as an I2C interface, an I2S interface, a UATT interface, a MIPI interface, etc.

[0106] It is understood that the interface connection relationships between the modules illustrated in the embodiments of the present invention are merely illustrative and do not constitute a structural limitation on the vehicle-mounted device 10. In other embodiments of this application, the vehicle-mounted device 10 may also employ different interface connection methods or combinations of multiple interface connection methods as described in the above embodiments.

[0107] The wireless communication function of the vehicle-mounted device 10 can be implemented through antenna 1, antenna 2, mobile communication module 160, wireless communication module 170, modem processor, and baseband processor.

[0108] Antennas 1 and 2 are used to transmit and receive electromagnetic wave signals. Each antenna in the vehicle-mounted device 10 can be used to cover one or more communication frequency bands. Different antennas can also be reused to improve antenna utilization. For example, antenna 1 can be reused as a diversity antenna for a wireless local area network. In some other embodiments, the antennas can be used in conjunction with a tuning switch.

[0109] The mobile communication module 160 can provide wireless communication solutions, including 2G / 3G / 4G / 5G, for use in the vehicle-mounted device 10. The mobile communication module 160 may include at least one filter, switch, power amplifier, low-noise amplifier (LNA), etc. The mobile communication module 160 can receive electromagnetic waves via the antenna 1, and perform filtering, amplification, and other processing on the received electromagnetic waves before transmitting them to the modem processor for demodulation. The mobile communication module 160 can also amplify the signal modulated by the modem processor and convert it into electromagnetic waves for radiation via the antenna 1. In some embodiments, at least some functional modules of the mobile communication module 160 may be housed in the processor 110. In some embodiments, at least some functional modules of the mobile communication module 160 and at least some modules of the processor 110 may be housed in the same device.

[0110] The wireless communication module 170 can provide solutions for wireless communication applications on the vehicle-mounted device 10, including Wireless Local Area Networks (WLAN) (such as Wireless Fidelity (WiFi) networks), Bluetooth (BT), Global Navigation Satellite System (GNSS), Frequency Modulation (FM), Near Field Communication (NFC), and Infrared (IF) technologies. The wireless communication module 170 can be one or more devices integrating at least one communication processing module. The wireless communication module 170 receives electromagnetic waves via antenna 2, modulates and filters the electromagnetic wave signals, and sends the processed signal to processor 110. The wireless communication module 170 can also receive signals to be transmitted from processor 110, modulate and amplify them, and then convert them into electromagnetic waves for radiation via antenna 2.

[0111] In some embodiments, antenna 1 of vehicle-mounted device 10 is coupled to mobile communication module 160, and antenna 2 is coupled to wireless communication module 170, enabling vehicle-mounted device 10 or vehicle-mounted device 20 to communicate with networks and other devices via wireless communication technology. The wireless communication technology may include Global System for Mobile Communications (GSM), General Packet Radio Service (GPRS), Code Division Multiple Access (CDMA), Wideband Code Division Multiple Access (WCDMA), Time Division Code Division Multiple Access (TD-SCDMA), Long Term Evolution (LTE), BT, GNSS, WLAN, NFC, FM, and / or IT technologies, etc. The GNSS may include the Global Positioning System (GPS), the Global Navigation Satellite System (GLONASS), the BeiDou Navigation Satellite System (BDS), the Quasi-Zenith Satellite System (QZSS), and / or Satellite Based Augmentation Systems (SBAS).

[0112] The display screen 130 is used to display images, videos, etc. The display screen 130 includes a display panel. The display panel can be a Liquid Crystal Display (LCD), an Organic Light-Emitting Diode (OLED), an Active-Matrix Organic Light-Emitting Diode (AMOLED), a Flexible Light-Emitting Diode (FLED), Mini-LED, Micro-LED, Micro-OLED, Quantum Dot Light-Emitting Diodes (QLED), etc. In some embodiments, the vehicle-mounted device 10 may include one or more display screens 130, where N is a positive integer greater than 1. For example, the display screen 130 can display an automatic parking path.

[0113] The audio module 140 is used to convert digital audio information into analog audio signals for output, and also to convert analog audio input into digital audio signals. The audio module 140 can also be used for encoding and decoding audio signals. In some embodiments, the audio module 140 may be located in the processor 110, or some functional modules of the audio module 140 may be located in the processor 110.

[0114] The memory 150 can be used to store computer executable program code, including instructions. The memory 150 may include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback, image playback, etc.), etc. The data storage area may store data created during the use of the vehicle-mounted device 10 (such as radar data, etc.). Furthermore, the memory 150 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, Universal Flash Storage (UFS), etc. The processor 110 executes various functional applications and data processing of the vehicle-mounted device 10 by running instructions stored in the memory 150 and / or instructions stored in memory disposed in the processor, to implement the parking space recognition method in Embodiment 1.

[0115] Example 3 like Figure 13As shown, this embodiment provides a vehicle, including the vehicle-mounted device of Embodiment 2. An ultrasonic radar array is disposed around the perimeter of the vehicle. One example of the ultrasonic radar array arrangement is that radars H1, H2, H3, H4, H5, and H6 are symmetrically distributed on the front bumper, and radars T1, T2, T3, T4, T5, and T6 are symmetrically distributed on the rear bumper. Radars H1, H6, T1, and T6 are long-range detection radars, while the remaining radars are short-range detection radars. Radars H1 and H6 on both sides of the front bumper are typically used to scan parking spaces and obstacles on both sides of the vehicle. Radars T1 and T6 on both sides of the rear bumper are typically used in conjunction with radars H1 and H6 on both sides of the front bumper to detect obstacles within the parking space. Therefore, the horizontal field of view (HOV) of radars H1, H6, T1, and T6 on both sides of the front and rear bumpers is typically small. Furthermore, radars that perform long-range detection operate at different frequencies than radars that perform short-range detection. Therefore, radars that perform long-range detection cannot perform triangulation calculations with radars that perform short-range detection to locate obstacles. They usually assume that the detected obstacle is directly in front of the radar.

[0116] like Figure 14 The diagram illustrates a possible functional framework of the vehicle in this embodiment. The vehicle's functional framework may include various subsystems, such as the onboard device 10, sensor system 20, control system 30, one or more peripheral devices 40 (one is shown as an example), power supply 50, and computer system 60. Optionally, the vehicle may also include other functional systems, such as an engine system that provides power to the vehicle, etc., which are not limited herein. The sensor system 20 may include several detection devices that can sense the measured information and convert the sensed information into electrical signals or other required forms of information output according to a certain rule. As shown in the figure, these detection devices may include a global positioning system, a vehicle speed sensor, an inertial measurement unit, a camera unit, a wheel speed sensor, a steering sensor, a gear position sensor, or other components for automatic detection, etc., and this application is not limited thereto.

[0117] The control system 30 may include several components, such as the steering unit, braking unit, lighting system, automatic driving system, navigation system, and network time synchronization system shown in the figure. Optionally, the control system 30 may also include components such as a throttle controller and an engine controller for controlling the vehicle's speed; this application is not limiting.

[0118] Peripheral device 40 may include several components, such as the communication system shown in the figure. The communication system is used to enable network communication between the vehicle and other devices besides the vehicle. In practical applications, the communication system employs wireless communication technology or wired communication technology to achieve network communication between the vehicle and other devices. The wired communication technology can refer to communication between the vehicle and other devices via network cables or fiber optic cables. The wireless communication technology includes, but is not limited to, Global System for Mobile Communications (GSMO), General Packet Radio Service (GSP), Code Division Multiple Access (CDMA), Wideband CDMA, Time Division Multiple Access (TDMA), LTE, Wireless Local Area Network (WLAN), Bluetooth, Global Navigation Satellite System (GNSS), Frequency Modulation (FM), Short Range Wireless Communication (SMR), and infrared technology, etc. Optionally, the peripheral device may also include a touchscreen, microphone, and speaker, etc.

[0119] Power source 50 represents a system that provides electricity or energy to the vehicle, which may include, but is not limited to, rechargeable lithium batteries or lead-acid batteries. In practical applications, one or more battery components in the power source are used to provide electrical energy or power for vehicle startup. The type and materials of power source 50 are not limited in this application. Optionally, power source 50 may also be an energy source used to provide energy to the vehicle, such as gasoline, diesel, ethanol, solar cells or solar panels, etc., which are not limited in this application.

[0120] Several functions of the vehicle are controlled and implemented by the computer system 60. The computer system 60 may include one or more processors (the figure shows one processor as an example) and memory (also referred to as a storage device). In practical applications, the memory may be located inside the computer system or outside the computer system, such as as a cache in the vehicle; this application does not limit this. The processor 6 may include one or more general-purpose processors, such as a graphics processor. The processor can be used to run relevant programs or instructions corresponding to programs stored in the memory to implement the corresponding functions of the vehicle 1.

[0121] The memory may include volatile memory, such as RAM; it may also include non-volatile memory, such as ROM, flash memory, HDD, or SSD; or it may include combinations of the above types of memory. The memory can be used to store a set of program code or instructions corresponding to the program code, so that the processor can call the program code or instructions stored in the memory to implement the corresponding functions of the vehicle. These functions include, but are not limited to, some or all of the functions shown in the schematic diagram of the vehicle's functional framework. In this application, the memory may store a set of program code for vehicle control, which the processor can call to control the safe operation of the vehicle.

[0122] Among them, this application Figure 14The systems shown are merely examples and do not constitute a limitation. In practical applications, vehicles can combine several components according to different functions to obtain subsystems with corresponding functions. For example, a vehicle may also include an Electronic Stability Program (ESP) and an Electric Power Steering (EPS) system, etc., not shown in the figure. The ESP system may consist of some sensors from the sensor system and some components from the control system. Specifically, the ESP system may include wheel speed sensors, steering sensors, lateral acceleration sensors, and control units involved in the control system, etc. The EPS system may consist of some sensors from the sensor system, some components from the control system, and power supply components, etc. Specifically, the EPS system may include steering sensors, generators and reducers involved in the control system, battery power supply, etc. It should be noted that the above... Figure 14 This is merely a schematic diagram of one possible functional framework for a vehicle. In practical applications, a vehicle may include more or fewer systems or components, and this application does not impose any limitations.

[0123] The aforementioned vehicles may include cars, trucks, motorcycles, buses, ships, airplanes, helicopters, lawnmowers, recreational vehicles, amusement park vehicles, construction equipment, trams, golf carts, trains, and handcarts, etc., and the embodiments of this application do not impose any special limitations.

[0124] Although exemplary embodiments have been described herein with reference to the accompanying drawings, it should be understood that the above exemplary embodiments are merely illustrative and are not intended to limit the scope of this application. Various changes and modifications can be made therein by those skilled in the art without departing from the scope and spirit of this application. All such changes and modifications are intended to be included within the scope of this application as claimed in the appended claims.

[0125] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0126] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed.

[0127] The various component embodiments of this application can be implemented in hardware, or as software modules running on one or more processors, or a combination thereof. Those skilled in the art will understand that microprocessors or digital signal processors (DSPs) can be used in practice to implement some or all of the functions of some modules according to the embodiments of this application. This application can also be implemented as an apparatus program (e.g., a computer program and computer program product) for performing part or all of the methods described herein. Such an implementation of this application can be stored on a computer-readable medium, or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.

[0128] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, 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 the element.

[0129] Although the description of this application has been made in conjunction with the specific embodiments described above, it will be apparent to those skilled in the art that many substitutions, modifications, and variations can be made based on the foregoing. Therefore, all such substitutions, modifications, and variations are included within the spirit and scope of the appended claims.

Claims

1. A parking space recognition method, characterized in that, The method includes: The identified parking spaces in the current frame are associated with the tracked parking spaces in the parking space list to obtain the association results; Based on the association results, the location information of the tracked parking space is updated; The status of the tracked parking space is updated based on the updated location information of the tracked parking space and the vehicle location; Based on the updated location information and status of the tracked parking spaces, update and output the updated list of parking spaces.

2. The parking space recognition method according to claim 1, characterized in that, The process of associating all identified parking spaces in the current frame with the tracked parking spaces in the parking space list to obtain the association result includes: Based on the parking space outline of the current frame parking space and the parking space outline of the tracked parking space, calculate the intersection-union ratio of the parking space outline of the current frame parking space and the parking space outline of the tracked parking space. If the cross-connection-union ratio is greater than the preset cross-connection-union ratio threshold, then the current frame parking space and the tracked parking space are successfully associated; otherwise, the association is unsuccessful.

3. The parking space recognition method according to claim 2, characterized in that, The step of updating the location information of the tracked parking space based on the association result includes: If the current frame parking space and the tracked parking space are successfully associated, the location information of the associated tracked parking space is updated using the location information of the current frame parking space. If the tracking parking space fails to be associated successfully, the location information of the unassociated tracking parking space will be updated by dead reckoning. If the parking space in the current frame is not successfully associated, the location information of the parking space in the current frame will be updated and added to the parking space list.

4. The parking space recognition method according to claim 3, characterized in that, The step of updating the status of the tracked parking space based on the updated location information of the tracked parking space and the vehicle location includes: The location information includes the center location of the parking space, and the status of the tracked parking space includes test status, confirmation status, extrapolation status, and termination status. When the center of the tracked parking space in the test state is located behind the center of the rear axle of the vehicle, the status of the tracked parking space is adjusted to the confirmed state. When the nearest distance between the center of the tracked parking space and the vehicle exceeds a preset distance threshold while the tracked parking space is in the confirmation or extrapolation state, the state of the tracked parking space will be adjusted to the termination state. When a tracked parking space in the confirmed state is not associated with a vehicle or overlaps with a vehicle, the state of the tracked parking space is adjusted to the extrapolation state.

5. The parking space recognition method according to claim 2, characterized in that, The method for identifying the parking space in the current frame is as follows: Detect and identify obstacles and obtain detection information of the obstacles; Identify and remove abnormal obstacles to obtain normal obstacles; The detection information of the normal obstacle is corrected to obtain the corrected detection information of the normal obstacle; The parking space in the current frame is constructed based on the two adjacent normal obstacles and their corrected detection information.

6. The parking space recognition method according to claim 5, characterized in that, The process of constructing the parking space in the current frame based on two adjacent normal obstacles and their corrected detection information includes: Filter from the normal obstacles to find available obstacles for constructing the parking space in the current frame; Based on the corrected detection information of two adjacent available obstacles, the type of the parking space in the current frame is determined, and the initial corner point of the parking space in the current frame is obtained; Based on the initial corner points, the parking space outline of the current frame parking space is generated.

7. The parking space recognition method according to claim 6, characterized in that, The step of generating the parking space outline of the current frame parking space based on the initial corner point includes: Based on the initial corner points, the initial outline of the parking space in the current frame is generated; If the difference between the size of the initial outline and the size of the preset standard parking space exceeds a preset size threshold, then the initial corner point is adjusted based on the size of the preset standard parking space to obtain the optimized corner point; Based on the optimized corner points, the parking space outline of the current frame parking space is generated.

8. The parking space recognition method according to claim 6, characterized in that, The step of generating the parking space outline of the current frame parking space based on the initial corner point includes: Based on the initial corner points, the initial outline of the parking space in the current frame is generated; If the initial contour overlaps with the contour of the available obstacle, then determine the coordinates of the deepest vertex of the available obstacle that extends into the initial contour; Based on the coordinates of the deepest vertex, the initial corner point is adjusted to obtain the optimized corner point; Based on the optimized corner points, the parking space outline of the current frame parking space is generated.

9. A vehicle-mounted device, comprising a processor and a memory storing a computer program or instructions capable of running on the processor, characterized in that, When the computer program or instructions are executed by the processor, they implement the parking space recognition method as described in any one of claims 1 to 8.

10. A vehicle, characterized in that, Includes the vehicle-mounted device as described in claim 9.

Images