Idle parking space detection method and device, vehicle and medium

Abstract

The invention relates to an idle parking space detection method and device, a vehicle and a medium, and the method comprises the steps: determining an idle parking space based on the collected obstacle information and the driving information of the vehicle in the collection process, so as to improve the accuracy and reliability of the detection of the idle parking space.

Description

Technical Field

[0001] This disclosure relates to the field of vehicle detection technology, specifically to a method, equipment, vehicle, and medium for detecting vacant parking spaces. Background Technology

[0002] Smart parking technology has been widely applied in the field of intelligent driving, and accurately identifying vacant parking spaces is crucial for improving parking efficiency and user experience.

[0003] In existing technical solutions, vehicle sensors can be used to acquire the coordinates of obstacles, and then the distance between these coordinates can be used to determine available parking spaces. However, this method has significant limitations and is typically only applicable to the ideal situation where the vehicle is traveling at a constant speed in a straight line. In practical applications, as the vehicle moves, its sensors also move. When the sensors move at different speeds and in different directions, the acquired obstacle coordinates cannot accurately reflect the actual distance between obstacles, thus affecting the accuracy and reliability of available parking space detection. Summary of the Invention

[0004] The purpose of this disclosure is to provide a method, device, vehicle, and medium for detecting vacant parking spaces, so as to improve the accuracy and reliability of vacant parking space detection.

[0005] To achieve the above objectives, in a first aspect, this disclosure provides a method for detecting vacant parking spaces, the method comprising: Available parking spaces are determined based on the obstacle information collected and the vehicle driving information collected during the collection process.

[0006] Optionally, determining available parking spaces based on the collected obstacle information and vehicle driving information during the collection process includes: Based on the obstacle information, determine the data collection points for the obstacles; The available parking spaces are determined based on the data collected from the collection points and the driving information.

[0007] Optionally, determining the available parking space based on the data collected from the collection points and the driving information includes: Based on the driving information, the collected point data is divided into multiple point clusters; The available parking spaces are determined based on the distance between adjacent point clusters.

[0008] Optionally, the driving information includes vehicle speed; the step of dividing the collected point data into multiple point clusters based on the driving information includes: Determine the target distance threshold based on the vehicle speed; The collected data points are divided into multiple clusters based on the target distance threshold.

[0009] Optionally, the target distance threshold is positively correlated with the vehicle speed.

[0010] Optionally, dividing the collected point data into multiple point clusters according to the target distance threshold includes: Based on the distance between any two adjacent collection points in the collection point data, collection points whose distance is greater than the target distance threshold are divided into different point clusters, and collection points whose distance is less than the target distance threshold are divided into the same point cluster.

[0011] Optionally, the method further includes: Obtain the vehicle's pose information at each acquisition moment; Based on the vehicle's pose information at each of the acquisition times, establish a vehicle coordinate system corresponding to each acquisition time; and / or A map coordinate system is established based on the initial pose information of the vehicle at the initial acquisition time.

[0012] Optionally, the vehicle includes a distance detection device mounted on the side of the vehicle body, and the obstacle information includes the distance between the distance detection device and the obstacle at each of the acquisition times; The step of determining the obstacle collection point data based on the obstacle information includes: For any given acquisition time, based on the obstacle distance, the installation position of the distance detection device, and the vehicle width, the acquisition point data of the obstacle corresponding to the vehicle at that acquisition time in the vehicle's coordinate system at that acquisition time is determined.

[0013] Optionally, the method further includes: The obstacle acquisition point data is converted from the corresponding vehicle coordinate system to the map coordinate system.

[0014] Optionally, the step of converting the obstacle's data collection points from the corresponding vehicle coordinate system to the map coordinate system includes: Determine the vehicle's position coordinates and yaw angle in the map coordinate system at any given time of data collection. Based on the vehicle position coordinates and the yaw angle, the data of the obstacle in the vehicle coordinate system at the corresponding acquisition time is converted into the data of the acquisition point in the map coordinate system.

[0015] Optionally, the method further includes: If the obstacle distance collected at any of the aforementioned collection times is invalid data, the minimum obstacle distance corresponding to multiple collection times within a preset time range near that collection time shall be taken as the obstacle distance at that collection time.

[0016] Optionally, determining the available parking space based on the distance between adjacent point clusters includes: If the distance between any two adjacent point clusters is greater than a preset parking space width threshold, the spatial area between the adjacent point clusters is determined as the vacant parking space.

[0017] Optionally, the vehicle includes a distance detection device mounted on the side of the vehicle body, the emission medium of the distance detection device having a diffusion angle, and prior to determining an available parking space based on the distance between adjacent point clusters, the following steps are included: For any of the aforementioned point clusters, determine the target acquisition points that deviate from the true boundary contour of the obstacle due to the diffusion angle of the emission medium and the shape of the obstacle; Remove the target acquisition point from the point cluster.

[0018] Optionally, determining, for any of the point clusters, target acquisition points that deviate from the true boundary contour of the obstacle due to the diffusion angle of the emission medium and the shape of the obstacle includes: For any given cluster of points, determine the obstacle distance of each collection point in the cluster corresponding to the vehicle position at the current time, and the average obstacle distance corresponding to all collection points in the cluster. The collection points in the point cluster whose obstacle distance exceeds the average obstacle distance and reaches a preset error threshold are determined as the target collection points.

[0019] Secondly, this disclosure also provides an electronic device, comprising: A memory on which computer programs are stored; A processor is configured to execute the computer program in the memory to implement the vacant parking space detection method provided in the first aspect of this disclosure.

[0020] Thirdly, this disclosure also provides a vehicle that includes the electronic equipment provided in the second aspect of this disclosure.

[0021] Fourthly, this disclosure also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the vacant parking space detection method provided in the first aspect of this disclosure.

[0022] By combining the collected obstacle information with vehicle driving information, the dynamic changes of vehicles on the acquired obstacle information are fully considered, thereby enabling accurate judgment of the distance between obstacles and improving the accuracy and reliability of vacant parking space detection.

[0023] Other features and advantages of this disclosure will be described in detail in the following detailed description section. Attached Figure Description

[0024] The accompanying drawings are provided to further illustrate the present disclosure and form part of the specification. They are used together with the following detailed description to explain the present disclosure, but do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart illustrating an vacant parking space detection method according to an exemplary embodiment.

[0025] Figure 2 This is a schematic diagram illustrating obstacle detection in a map coordinate system according to an exemplary embodiment.

[0026] Figure 3 This is a schematic diagram illustrating obstacle detection in a vehicle coordinate system according to an exemplary embodiment.

[0027] Figure 4 This is a schematic diagram of a point cluster before removing outliers, according to an exemplary embodiment.

[0028] Figure 5 This is a schematic diagram of a point cluster after outlier removal, according to an exemplary embodiment. Figure 6 This is a block diagram illustrating an vacant parking space detection device according to an exemplary embodiment.

[0029] Figure 7 This is a block diagram of a vehicle proposed in an exemplary embodiment of the present disclosure. Detailed Implementation

[0030] The specific embodiments of this disclosure will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit this disclosure.

[0031] Figure 1 This is a flowchart illustrating an vacant parking space detection method according to an exemplary embodiment, such as... Figure 1 As shown, the vacant parking space detection method includes the following steps.

[0032] Step S1: Determine available parking spaces based on the collected obstacle information and the vehicle driving information collected during the process.

[0033] In one embodiment, obstacle information includes at least one of obstacle acquisition point data and obstacle distance. Obstacle information can be acquired using sensors or cameras installed on the vehicle. The vehicle's sensors include at least one of radar sensors, distance sensors, ultrasonic sensors, and infrared sensors. Driving information includes vehicle speed, driving direction (yaw angle), and yaw rate. Vehicle speed can be obtained using a vehicle speed sensor or other vehicle speed monitoring equipment, and the vehicle's driving direction can be obtained using the vehicle's steering wheel angle or a vehicle positioning device.

[0034] When obstacle information is presented as data points, each data point represents the obstacle's position in space. Vehicle speed affects the density of data point distribution, and the driving direction changes the sensor's relative detection direction, thus affecting the shape of the data point distribution. Yaw rate causes non-inertial distortion in the data point data. When obstacle information is presented as obstacle distances, each obstacle distance represents the distance between the sensor and the obstacle in its directly opposite direction.

[0035] In one embodiment, whether two adjacent data collection points belong to the same obstacle is typically determined by a set distance threshold. When the vehicle speed is high, the area scanned by the sensor per unit time is larger, resulting in relatively sparse coordinate points. Conversely, when the vehicle speed is low, the data collection points are relatively dense. Therefore, it is necessary to establish a correlation between vehicle speed and the distance threshold. If the frequency of data collection for obstacle collection points is fixed, then at higher speeds, the data collection points are sparser, meaning the distance between two adjacent data collection points is greater. To reasonably classify these points as representing different obstacles, a relatively large distance threshold needs to be set. Conversely, at lower speeds, the data collection points are denser, requiring a relatively small distance threshold to accommodate this variation in coordinate point spacing. For example, a correspondence table or function expression between vehicle speed and distance threshold can be pre-determined through experimental testing, mathematical modeling, etc., and then the corresponding distance threshold can be found or calculated from the correspondence based on the acquired real-time vehicle speed.

[0036] It is also worth noting that because the vehicle collects obstacle data points very frequently, the vehicle speed corresponding to two or more data points is the same when they are collected. For example, if the vehicle's collection frequency is 50Hz, then the vehicle will collect 50 data points in 1 second. To simplify the calculation, the vehicle speed corresponding to these 50 data points can be regarded as the same speed.

[0037] In another embodiment, assuming a vehicle is driving towards a parking space area in a certain direction, when the vehicle's direction of travel forms a 30-degree angle with the parking space layout, the obstacle information acquired by the sensor may be obstructed and deviated due to the viewing angle. For example, if a pillar next to a parking space is partially obscured by other vehicles in the current direction, from the sensor's perspective, it may incorrectly determine that the parking space is occupied. However, by using the vehicle's direction of travel and combining the collected obstacle information for 3D spatial modeling and analysis, and by calculating the projection relationship and obstruction of obstacles in different directions, the true position and spacing of obstacles can be restored, accurately identifying that the parking space is actually vacant. When the vehicle turns and changes its direction of travel, the direction of travel can be updated in real time, continuously adjusting the method of analyzing obstacle information to always ensure that vacant parking spaces can be accurately determined based on the vehicle's direction of travel.

[0038] By combining the collected obstacle information with vehicle driving information, the dynamic changes of vehicles on the acquired obstacle information are fully considered, thereby enabling accurate judgment of the distance between obstacles and improving the accuracy and reliability of vacant parking space detection.

[0039] In one implementation, step S1 can be performed by the following steps: S11. Based on the obstacle information, determine the data collection points for the obstacles.

[0040] It is worth noting that the obstacle acquisition data obtained by the vehicle can be multiple coordinate points in a fixed coordinate system or a dynamic coordinate system. For example, the fixed coordinate system can be a pre-calibrated map coordinate system, and the dynamic coordinate system can be a vehicle coordinate system that changes dynamically as the vehicle moves.

[0041] In one embodiment, the obstacle acquisition point data includes the coordinates of multiple obstacle acquisition points. The acquisition point data can be obtained directly by sensors, such as by radar sensors, or indirectly by sensors, such as by ultrasonic sensors, to obtain the distance between the obstacle and the vehicle, and then the coordinates of the obstacle are calculated based on the distance, which is the obstacle acquisition point data.

[0042] S12. Determine the available parking spaces based on the obstacle collection point data and vehicle driving information.

[0043] Optionally, step S12 above includes the following sub-steps: S1201, Based on the driving information, the collected point data is divided into multiple point clusters.

[0044] S1202, determine the available parking space based on the distance between adjacent point clusters.

[0045] Optionally, the driving information includes vehicle speed. The above sub-step S1201 can be performed in the following manner: The first step is to determine the target distance threshold based on the vehicle speed.

[0046] It's worth noting that the area scanned by the sensor per unit time varies with vehicle speed, resulting in different density of data collection points. Therefore, it's necessary to establish a correlation between vehicle speed and a distance threshold. If the frequency of data collection at obstacle points is fixed, then at higher speeds, the data collection points are sparser, meaning the distance between two adjacent points is greater. To reasonably divide these points into clusters representing different obstacles, a relatively large distance threshold needs to be set. Conversely, at slower speeds, the data collection points are denser, requiring a relatively small distance threshold.

[0047] In one implementation, a correspondence table or function expression between vehicle speed and distance threshold can be pre-determined through experimental testing, mathematical modeling, etc., and then the corresponding target distance threshold can be found or calculated from the correspondence based on the acquired real-time vehicle speed.

[0048] By dynamically adjusting the target distance threshold based on vehicle speed, it can adapt to changes in the distribution of collection points at different vehicle speeds, providing a suitable standard for accurately dividing point clusters, improving the accuracy of point cluster division, and thus enhancing the precision of vacant parking space detection.

[0049] The second step is to divide the collected point data into multiple point clusters based on the target distance threshold.

[0050] Based on the target distance threshold determined in the first step above, the acquired obstacle collection point data is clustered. The basic idea of ​​clustering is to group collection points that are close to each other into the same cluster.

[0051] In one embodiment, an algorithm such as DBSCAN (Density-Based Spatial Clustering of Applications with Noise) can be used. This algorithm uses a specific data collection point as the core and examines other data collection points whose distance is within a target distance threshold. If the number of such points reaches the threshold set in the clustering algorithm, these data collection points are grouped into a single cluster. This process is then repeated for any unclassified data collection points until all data collection points are assigned to their respective clusters.

[0052] For example, assuming the target distance threshold is set to 1m, the DBSCAN algorithm is used for point clustering. A point P1 is selected from the collected data. Other points within 1m of P1 are checked. If points P2, P3, and P4 meet the clustering criteria, and the algorithm's threshold of at least 3 points is met, then P1, P2, P3, and P4 are grouped into a single point cluster C1. Next, another point P5 is selected from the ungrouped points, and the above process is repeated. If not enough points within 1m of P5 meet the clustering criteria, P5 may be marked as a noise point, or other neighboring points may be searched until all collected points have been processed, ultimately resulting in multiple point clusters C1, C2, C3, etc.

[0053] In another embodiment, point clusters can be divided according to the distance between any two adjacent data points and the corresponding target distance threshold. That is, when the distance between two adjacent data points is greater than or equal to the corresponding target distance threshold, the two adjacent data points belong to the same point cluster. Conversely, when the distance between two adjacent data points is less than the corresponding target distance threshold, the two adjacent data points belong to different point clusters. Here, adjacent data points refer to two data points collected in two adjacent acquisition frames or adjacent acquisition times.

[0054] By dividing the data points into clusters based on dynamic target distance thresholds, the influence of vehicle speed on the distribution of data points can be fully considered. This allows data points representing different obstacles to be accurately divided into different clusters, avoiding errors in cluster division due to changes in vehicle speed. This provides a reliable data foundation for accurately determining available parking spaces in the future.

[0055] Optionally, the target distance threshold is positively correlated with the vehicle speed.

[0056] For example, suppose the vehicle's ultrasonic sensor has a frame rate of 50Hz, meaning it collects obstacle data every 0.02 seconds. When the vehicle speed is slow, such as when it's traveling at 1 m / s, the vehicle moves 0.02 m within the 0.02-second interval between two consecutive data collections. In this case, the obstacle data collection points are relatively dense, and we can set a target distance threshold of 0.1 m. This allows us to accurately group data collection points belonging to the same obstacle into one cluster. When the vehicle speed is faster, such as when it increases to 5 m / s, the vehicle moves 0.1 m within the 0.02-second interval between two consecutive data collections. The collected obstacle data points become sparse. If we still use a target distance threshold of 0.1 m, data collection points that originally belonged to the same obstacle might be grouped into different clusters. Therefore, we need to increase the target distance threshold, for example, to 0.3 m, to more reasonably divide the data into clusters.

[0057] Alternatively, the collected data can be divided into multiple clusters of points in the following manner: First, determine the distance between any two adjacent collection points in the collection point data.

[0058] It is worth noting that adjacent sampling points are sampling points collected at two adjacent sampling times.

[0059] In one embodiment, the distance between two adjacent sampling points can be calculated using Euclidean distance.

[0060] Then, two adjacent sampling points whose distance is greater than the target distance threshold are divided into different point clusters, and two adjacent sampling points whose distance is less than the target distance threshold are divided into the same point cluster.

[0061] For example, suppose in a parking lot environment, a vehicle acquires data from a series of obstacle collection points via sensors. These collection points are distributed at different locations within the parking lot. The target distance threshold is set to 1 meter. The data sequence is: P1(1,1), P2(1.5,1), P3(3,1), P4(3.5,1). The distance between P1 and P2 is 0.5 meters, less than 1 meter, so P1 and P2 are assigned to the same point cluster, designated as point cluster C1. The distance between P2 and P3 is 1.5 meters, greater than 1 meter, so P3 is assigned to a new point cluster, designated as point cluster C2. The distance between P3 and P4 is 0.5 meters, less than 1 meter, so P4 is assigned to point cluster C2.

[0062] This division method helps to structure the data collected at the points, making it easier to conduct separate analyses of different point clusters. For example, in the scenario of determining vacant parking spaces, it is possible to distinguish point sets belonging to different obstacles or different areas.

[0063] Optionally, the vacant parking space detection method further includes: The vehicle's pose information at each acquisition moment is obtained.

[0064] Based on the vehicle's pose information at each acquisition time, establish a vehicle coordinate system corresponding to each acquisition time; and / or, based on the vehicle's initial pose information at the initial acquisition time, establish a map coordinate system.

[0065] The vehicle's pose information includes its position and heading. The vehicle's position can be either its center of mass or its geometric center.

[0066] In one embodiment, for each acquisition time, the vehicle's centroid position at that acquisition time is taken as the center of the vehicle body coordinate system at that acquisition time, and the vehicle's front direction at that acquisition time is taken as the y-axis to establish the vehicle body coordinate system at that time.

[0067] In another embodiment, the vehicle has a parking space detection program. After the vehicle enters the parking lot, the vehicle can automatically start the program or the user can manually start the program. In response to the start of the vehicle's parking space detection program, the initial centroid position of the vehicle is used as the initial centroid position of the vehicle, and the initial head direction of the vehicle is used as the initial head direction of the vehicle to establish a map coordinate system. That is, the initial centroid position is used as the center of the map coordinate system, and the initial head direction of the vehicle is used as the y-axis to establish the map coordinate system.

[0068] Optionally, the vehicle includes a distance detection device mounted on the side of the vehicle body, and the obstacle information includes the obstacle distance between the distance detection device and the obstacle at each acquisition time.

[0069] Optionally, step S11 above includes the following implementation: For any given acquisition time, based on the obstacle distance, the installation position of the distance detection device, and the vehicle width, the acquisition point data of the obstacle corresponding to the vehicle at that acquisition time in the vehicle's coordinate system at that acquisition time is determined.

[0070] For example, at each acquisition moment, the distance detection device emits a signal (such as ultrasound, laser, etc.). When the signal encounters an obstacle, it reflects back. The distance detection device calculates the distance between itself and the obstacle based on information such as the time difference between the signal's emission and reception. For instance, for an ultrasonic sensor, the obstacle distance d = vt / 2 can be calculated based on the speed of sound v in the air and the round-trip time t.

[0071] Then see Figure 2 As shown, d is the probe's detection distance, i.e., the distance to the obstacle, and X... d Let W be the longitudinal distance from the installation position of the distance detection device to the center of the rear axle, and let W be the width of the vehicle body. The coordinate data of the obstacle corresponding to the vehicle in the vehicle coordinate system at the time of data acquisition are calculated using the following formula 1: Formula 1:

[0072] In one optional implementation, the distance detection device can be a sensor probe such as an ultrasonic probe, laser probe, or electromagnetic probe that can acquire the distance of the vehicle to the obstacle. When acquiring obstacle information, the distance detection device may be subject to external interference (such as noise or multipath effect), which may lead to missed data or measurement errors. Therefore, after acquiring obstacle information through the distance detection device, the obstacle information can be corrected and optimized to improve detection efficiency. The obstacle information includes obstacle distances acquired at multiple acquisition times.

[0073] It's worth noting that the data collected in the vehicle coordinate system were acquired at different times, and the coordinates of each data point at each time point are represented based on the vehicle's coordinate system at that moment. The vehicle coordinate system is established with the vehicle itself as a reference, and its origin and coordinate axis directions change continuously as the vehicle moves. For example, the vehicle may turn, move forward, or reverse during travel, causing changes in the position and orientation of the vehicle coordinate system. Therefore, the coordinates of data points at different times are obtained under different vehicle coordinate systems and cannot be directly processed and analyzed uniformly.

[0074] A map coordinate system is a relatively fixed global coordinate system that serves as a unified reference framework for a specific area, such as a parking lot. Transforming the coordinate data of collected points from the vehicle coordinate system to the map coordinate system unifies the data from all collected points under a fixed reference frame. This facilitates subsequent overall processing and analysis of the collected points, such as performing point clustering and determining available parking spaces based on cluster spacing. Without coordinate transformation, differences in the vehicle coordinate system at different times may lead to confusion regarding the positional relationships of the collected points, making it impossible to accurately identify obstacles and available parking spaces.

[0075] In order to effectively process and analyze the collected obstacle data, it is necessary to unify all the data points into the same coordinate system.

[0076] In one implementation, the coordinate system can be a vehicle body coordinate system established based on the vehicle itself. The vehicle body coordinate system is a coordinate system that dynamically changes according to the movement of the vehicle. That is, during the movement of the vehicle, each acquisition moment or acquisition frame corresponds to a vehicle body coordinate system. Each acquisition point obtained by acquisition corresponds to the vehicle body coordinate system at its own acquisition moment. Therefore, when calculating the distance between two acquisition points in the vehicle body coordinate system, it is first necessary to convert the data of the two coordinate points to the same vehicle body coordinate system, and then calculate the distance between the two coordinate points using Euclidean distance.

[0077] In another implementation, the coordinate system can also be the parking lot's map coordinate system, which is a fixed coordinate system. In the map coordinate system, each data collection point has a unique coordinate representation. Using the coordinate representations of two data collection points, the distance between any two data collection points can be calculated using a distance formula (such as Euclidean distance).

[0078] By calculating the distance between collection points in the same coordinate system and combining it with the target distance threshold to divide the points into clusters, it is possible to accurately distinguish the collection points representing different obstacles, forming a clear point cluster structure, which provides a reliable data foundation for the subsequent determination of vacant parking spaces.

[0079] Optionally, after determining the obstacle's data points in the vehicle coordinate system at each data collection moment, the obstacle's data points can be transformed from the corresponding vehicle coordinate system to the map coordinate system through the following steps: Step 1: Determine the vehicle's position coordinates and yaw angle in the map coordinate system at any of the aforementioned collection times.

[0080] The vehicle position coordinates refer to the specific location of the vehicle's center of mass in the map coordinate system, usually represented by two-dimensional coordinates (x, y). The yaw angle refers to the rotation angle of the vehicle's heading relative to the positive x-axis of the map coordinate system; the yaw angle is used to characterize the vehicle's orientation.

[0081] For example, see Figure 3 As shown, the vehicle's position coordinates (X0, Y0) and yaw angle α in the map coordinate system can be obtained through the vehicle's integrated positioning system. The integrated positioning system includes at least one of the following: GPS (Global Positioning System), BDS (Beidou Navigation Satellite System), IMU (Inertial Measurement Unit), and wireless communication positioning system.

[0082] Step 2: Based on the vehicle's position coordinates and yaw angle, convert the obstacle's data points in the vehicle's coordinate system at the corresponding data collection time into data points in the map coordinate system.

[0083] For example, for the coordinates in the vehicle body coordinate system at any given time of data collection, the coordinate data of the data collection point in the vehicle body coordinate system can be transformed to the map coordinate system using the following formula: Formula 2: .

[0084] Where (X, Y) are the coordinates of the data collection point in the map coordinate system, (X... r Y r (X0, Y0) represents the coordinates of the data collection point in its corresponding vehicle coordinate system, and (X0, Y0) represents the coordinates of the vehicle's center of mass in the map coordinate system.

[0085] Optionally, the vacant parking space detection method further includes: If the obstacle distance collected at any of the aforementioned collection times is invalid data, the minimum obstacle distance corresponding to multiple collection times within a preset time range near that collection time shall be taken as the obstacle distance at that collection time.

[0086] In one embodiment, obstacle information can be optimized through the following steps: Step 1: Perform minimum value filtering on the acquired obstacle information.

[0087] For example, a data queue of length N can be created to store obstacle distances from the most recent N frames. When the obstacle distance for the current frame is obtained, it is compared with the N data points in the queue. The minimum distance value is then identified and output as the valid data for the current frame. Simultaneously, the data queue is updated by removing the oldest frame and adding the current frame. This minimum-value filtering process prevents erroneous data caused by accidental interference from affecting subsequent processing, thus improving data reliability.

[0088] For example, assuming N=5, the vehicle has collected obstacle distances for the last 5 frames (at the time of collection), which are 2 meters, 2.2 meters, 1.8 meters, 2.1 meters, and 1.9 meters respectively. The obstacle distance collected in the current frame is 5 meters (an abnormally large value due to interference). The current frame data is compared with the data from the last 5 frames. The smallest distance value is 1.8 meters, so 1.8 meters is taken as the valid data for the current frame. Simultaneously, the data queue is updated, removing the oldest frame data (2 meters) and adding the current frame data (5 meters). The new data queue contains 2.2 meters, 1.8 meters, 2.1 meters, 1.9 meters, and 5 meters.

[0089] It is worth noting that during the data acquisition and filtering process, there may be situations where the obstacle distance data at certain acquisition moments is invalid, such as sensor failure or signal loss. In order to ensure the continuity and reliability of the data, the obstacle information can be further processed in step two.

[0090] Step 2: If the obstacle distance collected at any time is invalid data, take the minimum obstacle distance corresponding to multiple collection times within a preset time range near that collection time as the obstacle distance at that collection time.

[0091] Invalid data includes missed data where a valid distance could not be obtained, and abnormal data where the distance value is too large or too small. When an obstacle distance at a certain acquisition time is detected as invalid data, the system will search for obstacle distances corresponding to all acquisition times within a preset time range (e.g., M acquisition times before and after) around that acquisition time. Then, the minimum obstacle distance is selected from these data and used as the obstacle distance at the acquisition time of the invalid data.

[0092] For example, suppose that an invalid obstacle distance is detected at the 10th acquisition time (e.g., displayed as -1 to indicate invalid). The preset time range is the three acquisition times before and after, that is, searching for obstacle distances at acquisition times 7 to 13. Assuming that the distances at these acquisition times are 2 meters, 2.1 meters, 1.9 meters, invalid, 2.2 meters, 2 meters, and 2.3 meters respectively, the minimum obstacle distance of 1.9 meters is selected and used as the obstacle distance at the 10th acquisition time.

[0093] By processing invalid data, the continuity of obstacle information over time is ensured, enabling vehicles to more smoothly track changes in the position and state of obstacles, thus providing better data support for vehicle vacant parking space detection.

[0094] Optionally, step S1202 above includes the following implementation: If the distance between any two adjacent point clusters is greater than the preset parking space width threshold, the space between the adjacent point clusters will be identified as an vacant parking space.

[0095] The preset parking space width threshold can be determined based on factors such as the common size of parking spaces in actual parking lots, the average size of vehicles, and the space required for parking operations. For example, in a typical small car parking lot, the preset parking space width threshold is set to 3 meters.

[0096] In addition, the preset parking space width threshold can be adjusted according to the vehicle's maximum lateral dimension at the current moment. For example, if the vehicle's maximum lateral dimension is 2.5 meters when the left and right side mirrors are open, the preset parking space width threshold can be set to 2.7 meters. If the vehicle's maximum lateral dimension is 2.7 meters when the left and right side mirrors are closed, the preset parking space width threshold can be set to 2.9 meters. In this way, the vehicle's vacant parking space detection can be applied to more types of parking lots and usage scenarios.

[0097] In one embodiment, different point clusters represent different obstacles. Since the obstacle collection points are collected sequentially according to the collection frequency, the adjacency relationship between point clusters is determined based on the time relationship between the collection times of the collection points. For example, collection points P1, P2, P3, P4, P5, and P6 are collected sequentially. Among them, P1 and P2 are divided into point cluster C1, P3 and P4 are divided into point cluster C2, and P5 and P6 are divided into point cluster C3. Thus, point clusters C1 and C2 are adjacent point clusters, and C2 and C3 are adjacent point clusters.

[0098] Furthermore, the distance between any two adjacent point clusters can be determined by the distance between the two closest sampling points in those two point clusters.

[0099] Using the example from the above embodiment for further explanation, the two closest collection points in adjacent point clusters C1 and C2 are P2 and P3, and the distance between P2 and P3 is 2 meters. Therefore, the distance between point clusters C1 and C2 is 2 meters. If the preset parking space width threshold is 3.2 meters, and the distance between C1 and C2 is less than the preset parking space width threshold, then the area between C1 and C2 is not an empty parking space. If the distance between C2 and C3 is 4 meters, which is greater than the preset parking space width threshold, then the area between C2 and C3 is an empty parking space.

[0100] By comparing the distance between adjacent point clusters and the preset parking space width threshold, the system can accurately determine the vacant parking space area in the parking lot based on the distribution of obstacles, providing drivers with reliable parking guidance information. As a result, drivers can quickly find a suitable parking space based on the vacant parking space information determined by the system, reducing the time spent searching for a parking space in the parking lot and thus improving the overall utilization efficiency of the parking lot.

[0101] Optionally, the vehicle includes a distance detection device mounted on the side of the vehicle body. The emission medium of the distance detection device has a diffusion angle. Before determining an available parking space based on the distance between adjacent clusters of points, the available parking space detection method further includes the following steps: Step 1: For any of the aforementioned point clusters, determine the target acquisition points that deviate from the true boundary contour of the obstacle due to the diffusion angle of the emission medium and the shape of the obstacle.

[0102] Step 2: Remove the target acquisition point from the point cluster.

[0103] It is worth noting that, see Figure 4 and Figure 5 As shown, the emitting medium (such as ultrasound) of the distance detection device has a diffusion angle during propagation, indicating that the emitted signal does not propagate in a completely straight line, but rather diffuses out at a certain angle. When encountering an obstacle, due to the combined effect of the diffusion angle and the shape of the obstacle, the signal may be reflected back from the side, edge, or other positions not directly facing the sensor, causing the collected points to deviate from the true boundary contour of the obstacle. Therefore, by considering the positional relationship between the collected points in the point cluster, their relative position to the vehicle, and the diffusion characteristics of the emitting medium and the possible shape of the obstacle, collected points that deviate from the true boundary contour of the obstacle (abnormal points) are identified and discarded as target collected points.

[0104] By eliminating target acquisition points that deviate from the true boundary due to the diffusion angle of the emission medium through the above scheme, the point clusters can more accurately depict the actual boundary contour of the obstacle, providing a more reliable basis for subsequent judgment on whether the space between obstacles is an empty parking space. Accurate obstacle boundary information helps to more accurately calculate the distance between adjacent point clusters, thereby more accurately judging whether these spaces meet the width requirements of the empty parking space, reducing the possibility of misjudgment, and improving the accuracy of empty parking space detection.

[0105] Alternatively, step 1 above can determine the target collection point in the following way: For any of the aforementioned point clusters: First, determine the obstacle distance of each collection point in the point cluster corresponding to the vehicle position at the current time, and the average obstacle distance corresponding to all collection points in the point cluster.

[0106] Then, the collection points in the point cluster whose obstacle distance exceeds the average obstacle distance and reaches a preset error threshold are determined as the target collection points.

[0107] In one embodiment, the acquired data points are located in the vehicle coordinate system. Since the vehicle coordinate system changes at each acquisition moment as the vehicle moves, it is first necessary to convert all data points in the vehicle coordinate system to the map coordinate system. The map coordinate system is a fixed coordinate system; converting the data points to it allows for processing within a unified framework, facilitating subsequent operations. Then, this data in the map coordinate system is converted back to the current vehicle coordinate system. Finally, based on the acquisition point positions in the current vehicle coordinate system, the obstacle distance of each acquisition point relative to the vehicle is determined.

[0108] In another embodiment, the acquired data points are located in a map coordinate system, and this data is directly converted to the vehicle's coordinate system at the current moment. After the conversion, the obstacle distance at each data point is determined in the vehicle's coordinate system at the current moment.

[0109] For any given collection point, after determining the coordinate data (X2, Y2) of the collection point in the vehicle coordinate system at the current moment, the difference between the absolute value of Y2 and half the width of the vehicle body can be determined as the obstacle distance of the collection point relative to the vehicle.

[0110] For example, suppose in a parking lot, an ultrasonic sensor on the side of a vehicle detects a cluster of points C, which contains 6 sampling points P1, P2, P3, P4, P5, and P6, with a preset error threshold of 1.5 meters.

[0111] First, determine the vehicle's coordinate system at the current moment by using the vehicle's current position.

[0112] Then, the coordinate data of each collection point in the vehicle body coordinate system at the current time is calculated. Based on the coordinate data of each collection point and the width of the vehicle body, the obstacle distance d of each collection point relative to the current vehicle is determined, where d1=3 meters, d2=3.2 meters, d3=2.8 meters, d4=6 meters, d5=3.1 meters, and d6=2.9 meters.

[0113] Secondly, calculate the average obstacle distance d. a =3.5 meters.

[0114] Finally, the differences between the obstacle distance and the average obstacle distance at each sampling point P1, P2, P3, P4, P5, and P6 are calculated, which are 0.5 meters, 0.3 meters, 0.7 meters, 2.5 meters, 0.4 meters, and 0.6 meters, respectively. Since P4 and d... a If the difference between the two values ​​is greater than the preset error threshold of 1.5 meters, then P4 is determined as the target acquisition point.

[0115] In another example, after determining the coordinates of each acquisition point in the point cluster relative to the vehicle's coordinate system at the current moment, the y-axis coordinate of each point in the cluster is obtained. Since the y-axis coordinate can be positive or negative, to uniformly measure lateral distance, the absolute value of the y-axis coordinate of each acquisition point is taken. Then, the smallest value among these absolute values ​​is denoted as y. min y min The corresponding points can be viewed as points in the point cluster that are relatively closer to the center of the vehicle or the core of the obstacle in the lateral direction. For each point in the point cluster, subtract the absolute value of its y-axis coordinate from the value of its y-axis coordinate. min The result is that the point is related to y. min The difference in distance between corresponding points in the horizontal direction. This difference reflects the distance of that point relative to the y-axis. min The degree of deviation of a point in the lateral direction. A threshold is preset, determined based on actual application scenarios and experience. If the lateral distance difference of a sampling point exceeds this threshold, it indicates that the sampling point deviates significantly from the main body of the point cluster in the lateral direction, and is likely an anomaly. These sampling points exceeding the threshold are removed from the point cluster, so that the remaining sampling points can more accurately represent the lateral range of the obstacle.

[0116] By removing outliers with large lateral deviations, the point clusters are made to more closely surround the main body of the obstacle, thus improving the accuracy of the point clusters in describing the lateral extent of the obstacle.

[0117] By identifying and eliminating target acquisition points using this method, the impact of inaccurate acquisition points caused by the diffusion angle of the emission medium on the judgment of obstacle boundaries can be reduced, making the point cluster more accurately reflect the true boundary contour of the obstacle.

[0118] See Figure 6As shown, this disclosure also provides an vacant parking space detection device 200, which includes a determination module 201.

[0119] The determination module 201 is configured to determine available parking spaces based on the collected obstacle information and the vehicle driving information during the collection process.

[0120] Optionally, the determining module 201 is configured to: Based on the obstacle information, determine the data collection points for the obstacles; The available parking spaces are determined based on the data collected from the collection points and the driving information.

[0121] Optionally, the determining module 201 is configured to: Based on the driving information, the collected point data is divided into multiple point clusters; The available parking spaces are determined based on the distance between adjacent point clusters.

[0122] Optionally, the driving information includes vehicle speed.

[0123] The determining module 201 is configured as follows: Determine the target distance threshold based on the vehicle speed; The collected data points are divided into multiple clusters based on the target distance threshold.

[0124] Optionally, the target distance threshold is positively correlated with the vehicle speed.

[0125] Optionally, the determining module 201 is configured to: Determine the distance between any two adjacent collection points in the collection point data; Two adjacent sampling points whose distance is greater than the target distance threshold are divided into different point clusters, and two adjacent sampling points whose distance is less than the target distance threshold are divided into the same point cluster.

[0126] Optionally, the determining module 201 is configured to: Obtain the vehicle's pose information at each acquisition moment; Based on the vehicle's pose information at each of the acquisition times, establish a vehicle coordinate system corresponding to each acquisition time; and / or A map coordinate system is established based on the initial pose information of the vehicle at the initial acquisition time.

[0127] Optionally, the vehicle includes a distance detection device mounted on the side of the vehicle body, and the obstacle information includes the obstacle distance between the distance detection device and the obstacle at each acquisition time.

[0128] The determining module 201 is configured as follows: For any given acquisition time, based on the obstacle distance, the installation position of the distance detection device, and the vehicle width, the acquisition point data of the obstacle corresponding to the vehicle at that acquisition time in the vehicle's coordinate system at that acquisition time is determined.

[0129] Optionally, the determining module 201 is configured to: The obstacle acquisition point data is converted from the corresponding vehicle coordinate system to the map coordinate system.

[0130] Optionally, the determining module 201 is configured to: Determine the vehicle's position coordinates and yaw angle in the map coordinate system at any given time of data collection. Based on the vehicle position coordinates and the yaw angle, the data of the obstacle in the vehicle coordinate system at the corresponding acquisition time is converted into the data of the acquisition point in the map coordinate system.

[0131] Optionally, the determining module 201 is configured to: If the obstacle distance collected at any of the aforementioned collection times is invalid data, the minimum obstacle distance corresponding to multiple collection times within a preset time range near that collection time shall be taken as the obstacle distance at that collection time.

[0132] Optionally, the determining module 201 is configured to: If the distance between any two adjacent point clusters is greater than a preset parking space width threshold, the spatial area between the adjacent point clusters is determined as the vacant parking space.

[0133] Optionally, the vehicle includes a distance detection device mounted on the side of the vehicle body, the emission medium of the distance detection device having a diffusion angle, and the determining module 201 is configured to: For any of the aforementioned point clusters, determine the target acquisition points that deviate from the true boundary contour of the obstacle due to the diffusion angle of the emission medium and the shape of the obstacle; Remove the target acquisition point from the point cluster.

[0134] Optionally, the determining module 201 is configured to: For any given cluster of points, determine the obstacle distance of each collection point in the cluster corresponding to the vehicle position at the current time, and the average obstacle distance corresponding to all collection points in the cluster. The collection points in the point cluster whose obstacle distance exceeds the average obstacle distance and reaches a preset error threshold are determined as the target collection points.

[0135] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0136] This disclosure also provides an electronic device, including: A memory that stores computer programs.

[0137] A processor is configured to execute the computer program in the memory to implement the vacant parking space detection method provided in this disclosure.

[0138] This disclosure also provides a vehicle including the electronic equipment described herein.

[0139] This disclosure also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the vacant parking space detection method provided in this disclosure.

[0140] Figure 7 This is a block diagram illustrating a vehicle 600 according to an exemplary embodiment. For example, vehicle 600 can be a hybrid vehicle, a non-hybrid vehicle, an electric vehicle, a fuel cell vehicle, or other types of vehicle. Vehicle 600 can be an autonomous vehicle, a semi-autonomous vehicle, or a non-autonomous vehicle.

[0141] Reference Figure 7 The vehicle 600 may include various subsystems, such as an infotainment system 610, a perception system 620, a decision control system 630, a drive system 640, and a computing platform 650. The vehicle 600 may also include more or fewer subsystems, and each subsystem may include multiple components. Furthermore, each subsystem and each component of the vehicle 600 can be interconnected via wired or wireless means.

[0142] In some embodiments, the infotainment system 610 may include a communication system, an entertainment system, and a navigation system, etc.

[0143] The perception system 620 may include several sensors for sensing information about the environment surrounding the vehicle 600. For example, the perception system 620 may include a global positioning system (which may be GPS, BeiDou, or other positioning systems), an inertial measurement unit (IMU), lidar, millimeter-wave radar, ultrasonic radar, and a camera device.

[0144] The decision control system 630 may include a computing system, a vehicle controller, a steering system, a throttle, and a braking system.

[0145] The drive system 640 may include components that provide powered motion to the vehicle 600. In one embodiment, the drive system 640 may include an engine, an energy source, a transmission system, and wheels. The engine may be one or a combination of internal combustion engines, electric motors, and compressed air engines. The engine is capable of converting energy provided by the energy source into mechanical energy.

[0146] Some or all of the functions of vehicle 600 are controlled by computing platform 650. Computing platform 650 may include at least one processor 651 and memory 652, processor 651 can execute instructions 653 stored in memory 652.

[0147] Processor 651 can be any conventional processor, such as a commercially available CPU. Processors may also include graphics processing units (GPUs), field-programmable gate arrays (FPGAs), systems-on-chips (SoCs), application-specific integrated circuits (ASICs), or combinations thereof.

[0148] The memory 652 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0149] In addition to instruction 653, memory 652 can also store data, such as road maps, route information, vehicle position, direction, speed, and other data. The data stored in memory 652 can be used by computing platform 650.

[0150] The preferred embodiments of this disclosure have been described in detail above with reference to the accompanying drawings. However, this disclosure is not limited to the specific details of the above embodiments. Within the scope of the technical concept of this disclosure, various simple modifications can be made to the technical solutions of this disclosure, and these simple modifications all fall within the protection scope of this disclosure.

[0151] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, this disclosure will not describe the various possible combinations separately.

[0152] Furthermore, various different embodiments of this disclosure can be combined in any way, as long as they do not violate the spirit of this disclosure, they should also be regarded as the content disclosed in this disclosure.

Claims

1. A method for detecting vacant parking spaces, characterized in that, The method includes: Available parking spaces are determined based on the obstacle information collected and the vehicle driving information collected during the collection process.

2. The method according to claim 1, characterized in that, The process of determining available parking spaces based on collected obstacle information and vehicle driving information during the collection process includes: Based on the obstacle information, determine the data collection points for the obstacles; The available parking spaces are determined based on the data collected from the collection points and the driving information.

3. The method according to claim 2, characterized in that, The step of determining the available parking space based on the data collected from the collection points and the driving information includes: Based on the driving information, the collected point data is divided into multiple point clusters; The available parking spaces are determined based on the distance between adjacent point clusters.

4. The method according to claim 3, characterized in that, The driving information includes vehicle speed; the step of dividing the collected point data into multiple point clusters based on the driving information includes: Determine the target distance threshold based on the vehicle speed; The collected data points are divided into multiple clusters based on the target distance threshold.

5. The method according to claim 4, characterized in that, The target distance threshold is positively correlated with the vehicle speed.

6. The method according to claim 4, characterized in that, The step of dividing the collected point data into multiple point clusters according to the target distance threshold includes: Determine the distance between any two adjacent collection points in the collection point data; Two adjacent sampling points whose distance is greater than the target distance threshold are divided into different point clusters, and two adjacent sampling points whose distance is less than the target distance threshold are divided into the same point cluster.

7. The method according to claim 2, characterized in that, The method further includes: Obtain the vehicle's pose information at each acquisition moment; Based on the vehicle's pose information at each of the acquisition times, establish a vehicle coordinate system corresponding to each acquisition time; and / or A map coordinate system is established based on the initial pose information of the vehicle at the initial acquisition time.

8. The method according to claim 7, characterized in that, The vehicle includes a distance detection device installed on the side of the vehicle body, and the obstacle information includes the distance between the distance detection device and the obstacle at each of the acquisition times; The step of determining the obstacle collection point data based on the obstacle information includes: For any given acquisition time, based on the obstacle distance, the installation position of the distance detection device, and the vehicle width, the acquisition point data of the obstacle corresponding to the vehicle at that acquisition time in the vehicle's coordinate system at that acquisition time is determined.

9. The method according to claim 8, characterized in that, The method further includes: The obstacle acquisition point data is converted from the corresponding vehicle coordinate system to the map coordinate system.

10. The method according to claim 9, characterized in that, The step of converting the obstacle acquisition point data from the corresponding vehicle coordinate system to the map coordinate system includes: Determine the vehicle's position coordinates and yaw angle in the map coordinate system at any given time of data collection. Based on the vehicle position coordinates and the yaw angle, the data of the obstacle in the vehicle coordinate system at the corresponding acquisition time is converted into the data of the acquisition point in the map coordinate system.

11. The method according to claim 8, characterized in that, The method further includes: If the obstacle distance collected at any of the aforementioned collection times is invalid data, the minimum obstacle distance corresponding to multiple collection times within a preset time range near that collection time shall be taken as the obstacle distance at that collection time.

12. The method according to claim 3, characterized in that, Determining the available parking space based on the distance between adjacent point clusters includes: If the distance between any two adjacent point clusters is greater than a preset parking space width threshold, the spatial area between the adjacent point clusters is determined as the vacant parking space.

13. The method according to claim 3, characterized in that, The vehicle includes a distance detection device mounted on the side of the vehicle body. The emission medium of the distance detection device has a diffusion angle. Before determining an available parking space based on the distance between adjacent point clusters, the method includes: For any of the aforementioned point clusters, determine the target acquisition points that deviate from the true boundary contour of the obstacle due to the diffusion angle of the emission medium and the shape of the obstacle; Remove the target acquisition point from the point cluster.

14. The method according to claim 13, characterized in that, For any of the aforementioned point clusters, determining the target acquisition points that deviate from the true boundary contour of the obstacle due to the diffusion angle of the emission medium and the shape of the obstacle includes: For any given cluster of points, determine the obstacle distance of each collection point in the cluster corresponding to the vehicle position at the current time, and the average obstacle distance corresponding to all collection points in the cluster. The collection points in the point cluster whose obstacle distance exceeds the average obstacle distance and reaches a preset error threshold are determined as the target collection points.

15. An electronic device, characterized in that, include: A memory on which computer programs are stored; A processor for executing the computer program in the memory to implement the method of any one of claims 1-14.

16. A vehicle, characterized in that, Includes the electronic device as described in claim 15.

17. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the method described in any one of claims 1-14.

Images