Method, device, equipment and storage medium for identifying an obliquely parked vehicle

Abstract

The present application relates to a method, apparatus, device, and storage medium for identifying obliquely parked vehicles. The method comprises: obtaining first perception fusion data collected by a first vehicle at a first moment, and determining at least one second vehicle that is obliquely parked relative to the first vehicle at the first moment based on the first perception fusion data; for each second vehicle, determining a first stationary vehicle from at least one second vehicle based on the relationship between the determined vehicle speed statistics of the second vehicle and a first preset speed statistical range; obtaining second perception fusion data collected by the first vehicle at a second moment; and updating the first stationary vehicle based on the feature information of the first stationary vehicle in the second perception fusion data and at least one third vehicle that is obliquely parked relative to the first vehicle determined based on the second perception fusion data to obtain a second stationary vehicle. The present application combines single-vehicle characteristics and group characteristics to construct stationary vehicles, and can be dynamically updated to improve recognition efficiency.

Description

Technical Field

[0001] The present application relates to the field of intelligent traffic information technology, and in particular to a method, device, equipment and storage medium for identifying an obliquely parked or stationary vehicle. Background Art

[0002] In intelligent driving systems, accurately identifying the motion state of surrounding vehicles is crucial for safe path planning and decision-making. As the core link between the perception and decision-making layers, the target selection module dynamically screens and classifies surrounding objects based on multimodal sensor data. This is particularly true in complex traffic scenarios, such as those with numerous stationary vehicles parked diagonally or sideways in adjacent lanes.

[0003] Traditional methods face two challenges: first, there are blind spots in perception caused by dense occlusion between target clusters, which leads to a significant decrease in the confidence of single-frame perception data; second, the interruption of feature association caused by mutual occlusion of multiple targets leads to cumulative errors in the motion state estimation based on continuous moment data, ultimately resulting in a much higher misjudgment rate for stationary targets than in open scenes. Summary of the Invention

[0004] Embodiments of the present application provide a method, apparatus, device, and storage medium for identifying an obliquely parked or stationary vehicle.

[0005] In a first aspect, an embodiment of the present application provides a method for identifying a vehicle that is parked at an angle, the method comprising:

[0006] Obtaining first perception fusion data collected by the first vehicle at a first moment, and determining, based on the first perception fusion data, at least one second vehicle that is parked obliquely relative to the first vehicle at the first moment;

[0007] For each second vehicle, determining a first stationary vehicle from the at least one second vehicle based on a relationship between the determined vehicle speed statistical information of the second vehicle and a first preset speed statistical range;

[0008] Acquire second perception fusion data collected by the first vehicle at a second moment; the second moment is continuous with the first moment;

[0009] Based on the feature information of the first stationary vehicle in the second perception fusion data and at least one third vehicle parked obliquely relative to the first vehicle determined based on the second perception fusion data, the first stationary vehicle is updated to obtain a second stationary vehicle.

[0010] According to the above-mentioned technical means, at least one second vehicle that meets the oblique stationary parking condition is identified in the first perception fusion data, and whether it constitutes a first stationary vehicle array is determined in combination with the statistical characteristics of the group, thereby improving the robustness of recognition. Furthermore, in subsequent moments, dynamic updates are performed based on the characteristic information of the first stationary vehicle and the newly appeared third vehicle, which can effectively deal with the problems of perception error accumulation and occlusion, reduce the misjudgment rate and improve recognition efficiency.

[0011] Furthermore, based on the first perception fusion data, at least one second vehicle that is parked obliquely relative to the first vehicle at the first moment is determined, including: determining each second vehicle that meets the oblique stationary parking condition in turn from the multiple vehicles included in the first perception fusion data; wherein the oblique stationary parking condition includes at least the following conditions: the lateral position of the vehicle is located in the effective area of ​​the adjacent lane of the first vehicle; the absolute value of the lateral speed of the vehicle is less than the preset lateral speed; the absolute value of the heading angle of the vehicle is within a preset angle range.

[0012] According to the above technical means, by setting clear oblique stationary parking conditions, including parameters such as lateral position, lateral speed and heading angle, potential obliquely parked vehicles can be quickly screened out within a single frame, reducing invalid calculations, improving processing efficiency, and enhancing the adaptability to real oblique parking scenarios.

[0013] Further, for each second vehicle, based on the relationship between the speed statistical information of the determined second vehicle and the first preset speed statistical range, a first stationary vehicle is determined from at least one second vehicle, including: for each second vehicle, in response to the number of the determined second vehicles being greater than the first preset number and the speed statistical information corresponding to the determined second vehicles being within the first preset speed statistical range, determining the determined second vehicle as the first stationary vehicle.

[0014] According to the above technical means, the first stationary vehicle is determined by judging the relationship between the vehicle speed statistical characteristics and the first preset speed statistical range. The use of group statistical characteristics significantly reduces the misjudgment rate caused by perceptual jitter and improves the stability of recognition.

[0015] Furthermore, the vehicle speed statistical information includes a lateral speed mean and a lateral speed variance, and the first preset speed statistical range includes a first speed mean range and a first speed variance range. For each second vehicle, based on the relationship between the determined vehicle speed statistical information of the second vehicle and the first preset speed statistical range, a first stationary vehicle is determined from at least one second vehicle, including: for each second vehicle, in response to the determined lateral speed mean of the second vehicle being within the first speed mean range and the lateral speed variance being within the first speed variance range, the determined second vehicle is determined as a first stationary vehicle.

[0016] Based on the above technical means, by introducing statistical features such as the mean and variance of lateral velocity and setting a preset range, multi-target collaborative analysis can be used to replace the traditional method that relies on single-target multi-frame timing judgment, significantly reducing the misjudgment rate caused by perceptual jitter and improving recognition stability.

[0017] Furthermore, based on the characteristic information of the first stationary vehicle in the second perception fusion data, the first stationary vehicle is updated to obtain the second stationary vehicle, including: determining first vehicle speed statistical information corresponding to the first stationary vehicle based on the characteristic information of the first stationary vehicle in the second perception fusion data; in response to the first vehicle speed statistical information not being within a second preset speed statistical range, determining a moving vehicle from the first stationary vehicle based on the characteristic information and deleting the moving vehicle.

[0018] According to the above technical means, by continuously monitoring the speed statistical information of the first stationary vehicle at consecutive moments, when it is detected that it deviates from the preset speed statistical range, the target that may change from stationary to moving can be eliminated in time to avoid misidentification as a stationary state, thereby improving the real-time performance and accuracy of the system.

[0019] Furthermore, the method also includes: in response to the first vehicle speed statistical information being within a second preset speed statistical range, or the number of the first stationary vehicles being less than the second preset number, at least one third vehicle parked obliquely relative to the first vehicle determined based on the second perception fusion data, updating the first stationary vehicle to obtain a second stationary vehicle.

[0020] According to the above technical means, by judging the entry of the third vehicle into the formation, it is possible to achieve dynamic maintenance of the target set while ensuring the stability of the existing targets, thereby improving the system's adaptability to complex scenarios.

[0021] Further, based on at least one third vehicle determined to be stationary and obliquely parked relative to the first vehicle by the second perception fusion data, updating the first stationary vehicle to obtain a second stationary vehicle includes:

[0022] sequentially determining, from the plurality of vehicles included in the second perception fusion data, each third vehicle that satisfies the oblique stationary parking condition; the third vehicle being a vehicle that is different from the first stationary vehicle in the second perception fusion data;

[0023] For each determined third vehicle, in response to vehicle speed statistical information corresponding to the determined third vehicle and the first stationary vehicle being within a first preset speed statistical range, the determined third vehicle is determined as a second stationary vehicle.

[0024] According to the above technical means, by performing a joint statistical analysis on the newly determined third vehicle and the first stationary vehicle, it is ensured that the newly added target meets the overall stationary characteristics, thereby achieving dynamic expansion of the target set and improving recognition accuracy.

[0025] Furthermore, the method further includes: setting a preset buffer time for the moving vehicle, and stopping identifying the moving vehicle within the preset buffer time.

[0026] According to the above technical means, by introducing a buffer mechanism, it is possible to avoid recognition instability caused by frequent state switching of moving vehicles in a short period of time, and improve the smoothness and robustness of system operation.

[0027] Furthermore, the method further includes: in response to the number of the second stationary vehicles being less than a third preset number, deleting the second stationary vehicles.

[0028] According to the above technical means, by setting a minimum number threshold of the second stationary vehicles, misjudgment caused by too few targets can be prevented, ensuring the stability of the system in low-confidence scenarios.

[0029] Furthermore, the method also includes: in response to the number of at least one second vehicle being less than or equal to a first preset number, or the vehicle speed statistical information corresponding to at least one second vehicle is not within the first preset speed statistical range, determining that the first vehicle is in the first scene information; the first scene information characterizes that there is no corresponding first stationary vehicle in the first perception fusion data.

[0030] According to the above technical means, by setting the judgment conditions for recognition failure, the diagonally parked stationary vehicle recognition process can be quickly exited to avoid resource waste and provide more accurate scene information for the downstream decision-making module.

[0031] In a second aspect, an embodiment of the present application provides a device for identifying a vehicle that is parked at an angle, the device comprising:

[0032] an identification module, configured to obtain first perception fusion data collected by the first vehicle at a first moment, and determine, based on the first perception fusion data, at least one second vehicle that is parked obliquely relative to the first vehicle at the first moment;

[0033] a determining module configured to determine, for each second vehicle, a first stationary vehicle from the at least one second vehicle based on a relationship between the determined vehicle speed statistical information of the second vehicle and a first preset speed statistical range;

[0034] An acquisition module, configured to acquire second perception fusion data collected by the first vehicle at a second moment; the second moment being continuous with the first moment;

[0035] An updating module is used to update the first stationary vehicle based on the feature information of the first stationary vehicle in the second perception fusion data and at least one third vehicle that is parked obliquely relative to the first vehicle determined based on the second perception fusion data to obtain a second stationary vehicle.

[0036] In a third aspect, an embodiment of the present application provides a device for identifying a stationary vehicle parked at an angle, the device comprising: a processor, a memory, and a communication bus; the processor implements the above-mentioned method for identifying a stationary vehicle parked at an angle when executing a running program stored in the memory.

[0037] In a fifth aspect, an embodiment of the present application provides a computer-readable storage medium, characterized in that a computer program is stored thereon, and when the computer program is executed by a processor, the above-mentioned method for identifying an obliquely parked stationary vehicle is implemented.

[0038] Beneficial effects of the embodiments of the present application:

[0039] (1) The embodiment of the present application improves the robustness of recognition by identifying the second vehicle that meets the oblique parking condition in the first perception fusion data and judging whether it constitutes the first stationary vehicle in combination with the statistical characteristics of the group vehicle speed. Furthermore, in subsequent moments, dynamic updates are performed based on the characteristic information of the first stationary vehicle and the newly appeared third vehicle, which can effectively deal with the problems of perception error accumulation and occlusion, reduce the misjudgment rate and improve recognition efficiency.

[0040] (2) By introducing statistical features such as the mean and variance of the lateral velocity and setting a preset range, the embodiment of the present application can utilize a multi-target collaborative analysis approach to replace the traditional method that relies on single-target multi-frame timing judgment, significantly reducing the misjudgment rate caused by perceptual jitter and improving the stability of recognition.

[0041] (3) The embodiment of the present application continuously monitors the speed statistics of the first stationary vehicle at consecutive moments. When it is detected that the first stationary vehicle deviates from the preset speed statistics range, the target that may change from stationary to moving can be eliminated in a timely manner to avoid misidentification as a stationary state, thereby improving the real-time performance and accuracy of the system.

[0042] (4) By judging whether a third vehicle enters the formation, the embodiment of the present application can achieve dynamic maintenance of the target set while ensuring the stability of the existing targets, thereby improving the system's adaptability to complex scenarios.

[0043] (5) The embodiment of the present application performs a joint statistical analysis on the newly determined third vehicle and the first stationary vehicle to ensure that the newly added target meets the overall stationary characteristics, thereby achieving dynamic expansion of the target set and improving recognition accuracy.

[0044] (6) By introducing a buffer mechanism, the embodiment of the present application can avoid recognition instability caused by the frequent state switching of the moving vehicle in a short period of time, thereby improving the smoothness and robustness of the system operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A schematic flow chart of a method for identifying an obliquely parked vehicle provided in an embodiment of the present application;

[0046] Figure 2 A schematic structural diagram of an exemplary scenario of densely parked vehicles at an angle provided in an embodiment of the present application;

[0047] Figure 3 A flowchart of an exemplary vehicle exit method provided in an embodiment of the present application;

[0048] Figure 4 A schematic diagram of an exemplary process for determining a second stationary vehicle provided in an embodiment of the present application;

[0049] Figure 5 A flowchart of an exemplary horizontal and vertical planning method provided in an embodiment of the present application;

[0050] Figure 6 A schematic flow chart of an exemplary method for identifying an obliquely parked vehicle provided in an embodiment of the present application;

[0051] Figure 7 A schematic diagram of the structure of a device for identifying an obliquely parked vehicle provided in an embodiment of the present application;

[0052] Figure 8 A schematic structural diagram of a device for identifying an obliquely parked vehicle provided in an embodiment of the present application. DETAILED DESCRIPTION

[0053] The following will describe the embodiments of the present application with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand the other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be understood that the preferred embodiments are only for the purpose of illustrating the present application and are not intended to limit the scope of protection of the present application.

[0054] The target selection module of a vehicle's assisted driving system serves as the connecting link between the perception layer and the decision-making layer. Its core function is to dynamically screen multimodal targets, such as vehicles, pedestrians, and obstacles, identified by the perception fusion module based on their kinematic properties (including but not limited to position coordinates, velocity vectors, acceleration, and other spatiotemporal characteristics). By constructing a probabilistic motion model to predict the trajectories of the vehicle and targets, it ultimately selects key targets of interest and transmits them to the downstream planning and control module for lateral and longitudinal planning and vehicle control. Accurately determining the target's motion state is crucial during the target selection process, especially in complex scenarios with numerous diagonally parked, stationary vehicles in adjacent lanes. Traditional methods face two challenges: first, dense occlusions between target clusters create perception blind spots, significantly reducing the confidence level of single-frame perception data. Second, the interruption of feature association caused by mutual occlusion among multiple targets leads to cumulative errors in the motion state estimation based on continuous-time data, ultimately resulting in a much higher misjudgment rate for stationary targets than in open scenarios.

[0055] To address the problem of identifying target motion states, related technologies have proposed a dynamic target screening mechanism based on lateral position pipeline prediction. This method establishes a pipeline affiliation between the target's actual position and the predicted lateral position to identify whether the target is crossing, crossing diagonally, transitioning from motion to stillness, or stationary. However, in practical applications of high-density diagonal parking scenarios in adjacent lanes, this method has the following limitations: its lateral position pipeline model is designed based on ideal observation conditions. When the target spacing is less than 2.5 meters (a typical vehicle distance characteristic of diagonal parking scenarios), dense occlusions will cause the standard deviation of the lateral position perception value to increase, exceeding the pipeline tolerance threshold. At this time, the trajectory prediction pipeline of the low-speed moving target overlaps with the attribute fluctuation range of the stationary target, and the system cannot effectively distinguish the essential differences in the motion of the two types of targets. Furthermore, the proposed general screening framework does not consider scene specificity and lacks the ability to identify high-density diagonal parking scenarios.

[0056] In view of the above technical problems, an embodiment of the present application provides a method for identifying a vehicle parked at an angle, which is implemented by an identification device for a vehicle parked at an angle. Figure 1 A flow chart of a method for identifying an obliquely parked vehicle provided in an embodiment of the present application is shown as follows: Figure 1 As shown, the process includes the following steps S101 to S104:

[0057] Step S101: Acquire first perception fusion data collected by a first vehicle at a first moment, and determine, based on the first perception fusion data, at least one second vehicle that is parked obliquely relative to the first vehicle at the first moment.

[0058] In the embodiments of this application, in an intelligent driving system, the first vehicle generally refers to a vehicle currently operating an assisted driving function. The first vehicle may be equipped with sensing devices such as onboard cameras and radars to collect real-time information about the surrounding environment. The first sensing fusion data may be fused data output by the onboard cameras, radars, and other sensing devices, which may include, but is not limited to, characteristic information such as the vehicle's position coordinates, velocity vector, and acceleration.

[0059] In the embodiment of the present application, each second vehicle identified by the device for identifying an obliquely parked stationary vehicle refers to a second vehicle that is parked obliquely relative to the first vehicle. For example, Figure 2 As shown, there are multiple stationary second vehicles 21 parked in the adjacent lane to which the first vehicle 20 is traveling. The identification device for diagonally parked stationary vehicles can identify the second vehicle 21 that is parked diagonally relative to the first vehicle. The basis for identification may be that the lateral position of the second vehicle is within the effective area of ​​the adjacent lane of the first vehicle, the lateral speed is close to zero, and the heading angle is within a preset angle range. This indicates that the second vehicle may be in a diagonally parked or sideways state, and a vehicle in this state may interfere with the longitudinal decision of the first vehicle.

[0060] In the embodiments of this application, the first fusion perception data is the fused perception data collected by the perception device at the first moment. Because intelligent driving systems need to monitor and respond to the surrounding environment in real time, each fusion perception data corresponds to a specific time point and is used for subsequent target detection and recognition processing. For example, when the first vehicle is traveling at 60 km / h, dozens of frames of fusion perception data may be collected per second. Each fusion perception data point serves as an important basis for the recognition device to control the vehicle when the vehicle is parked at an angle.

[0061] In an embodiment of the present application, the first perception fusion data is the perception fusion data collected by the perception device on the first vehicle at the first moment, and the previous perception fusion data did not identify the stationary vehicle, that is, no corresponding stationary vehicle was obtained.

[0062] In the embodiments of this application, the first perception fusion data can be collected by onboard cameras, radar, and lidar. Image processing algorithms then extract the vehicle's kinematic properties, including parameters such as lateral position, lateral velocity, and heading angle. Analysis of these parameters enables a preliminary determination of which vehicles may be parked at an angle. This allows for rapid screening of potential parked vehicles, reducing subsequent computing resource consumption and improving response efficiency.

[0063] Step S102: For each second vehicle, based on the determined relationship between the vehicle speed statistical information of the second vehicle and a first preset speed statistical range, determine a first stationary vehicle from at least one second vehicle.

[0064] In the embodiments of the present application, considering that under the conditions of the same sensor and the same perception fusion algorithm, the lateral velocity sequence of multiple stationary targets in the frames (perception fusion data) collected at the same time should be equivalent to the multi-frame random sampling of the lateral velocity of a single stationary target within a certain period of time, therefore, multi-target single-frame joint judgment can be adopted to replace the traditional single-target multi-frame time series judgment, and the problem of single-target perception jitter can be overcome by the group statistical characteristics. That is, after the preliminary judgment of the second vehicle that is parked obliquely relative to the first vehicle is made based on the first perception fusion data in step S101, a further judgment will be made based on the relationship between the determined vehicle speed statistical information of the second vehicle and the first preset speed statistical range to see whether it meets the group statistical characteristics.

[0065] In the embodiments of the present application, the first preset speed statistical range is a set of statistical indicators used to determine whether the identified second vehicles constitute a stable, stationary group. These characteristics reflect the stability of the second vehicles' lateral motion within a given frame. If the mean lateral velocity of multiple second vehicles is close to zero and the variance is small, this indicates that the lateral motion of these second vehicles varies little and has a high probability of being stationary. In other words, if the speed statistical information of the identified second vehicles (e.g., the mean lateral velocity or the variance) falls within the first preset speed statistical range, the identified second vehicles are determined to be the first stationary vehicles corresponding to the first perception fusion data.

[0066] Step S103: Acquire second perception fusion data collected by the first vehicle at the second moment.

[0067] In the embodiments of this application, the second perception fusion data refers to the next frame of data collected by the perception device at the moment immediately following the first moment of collection of the first perception fusion data. The first perception fusion data and the second perception fusion data are consecutive frames collected at consecutive moments, i.e., the first moment is continuous with the second moment. Because the intelligent driving system needs to continuously monitor the surrounding environment, each frame is temporally continuous with the previous frame.

[0068] For example, a first vehicle travels at a constant speed, continuously collecting multiple frames of sensory fusion data. Each frame of sensory fusion data may contain multiple vehicles in adjacent lanes. The device detecting diagonally parked vehicles must compare the preceding and following data to determine whether these vehicles remain stationary. Considering that most stationary vehicles in two consecutive frames will remain stationary, the second sensory fusion data corresponding to the first stationary vehicle is constructed based on the first sensory fusion data, resulting in more efficient identification of diagonally parked vehicles.

[0069] Step S104: Based on the feature information of the first stationary vehicle in the second perception fusion data and at least one third vehicle parked obliquely relative to the first vehicle determined based on the second perception fusion data, update the first stationary vehicle to obtain a second stationary vehicle.

[0070] In an embodiment of the present application, after obtaining the first stationary vehicle corresponding to the first perception fusion data, for the second perception fusion data, it is possible to first determine whether the first vehicle speed statistical information of the first stationary vehicle is still within the second preset speed statistical range. If not, the first stationary vehicle is adjusted based on the characteristic information of the first stationary vehicle in the second perception fusion data, and the speed statistical information corresponding to the adjusted first stationary vehicle is within the second preset speed statistical range.

[0071] Exemplarily, the first stationary vehicles corresponding to the first perception fusion data include vehicle 1, vehicle 2, vehicle 3, vehicle 4, and vehicle 5. Then, in the second perception fusion data, the first vehicle speed statistical information corresponding to vehicle 1, vehicle 2, vehicle 3, vehicle 4, and vehicle 5 is not within the second preset speed statistical range, indicating that the motion state of one of vehicles 1, vehicle 2, vehicle 3, vehicle 4, and vehicle 5 has changed, that is, a certain vehicle may turn from stationary to moving. Therefore, the vehicle with the largest lateral speed among vehicles 1, vehicle 2, vehicle 3, vehicle 4, and vehicle 5 can be deleted from the first stationary vehicle to obtain the second stationary vehicle.

[0072] In an embodiment of the present application, for a third vehicle that is different from the third vehicle in the first perception fusion data and the second perception fusion data, it is determined whether the third vehicle is a stationary vehicle based on the characteristic information of the third vehicle in the second perception fusion data. If so, it is determined as the second stationary vehicle.

[0073] In an embodiment of the present application, based on the first stationary vehicle corresponding to the first perception fusion data, it is only necessary to determine whether the vehicle speed statistical information of the first stationary vehicle corresponding to the first perception fusion data has changed in the second perception fusion data. If there is no change, the first stationary vehicle is directly determined to be the second stationary vehicle, and the third vehicle in the second perception fusion data is continued to be confirmed to be stationary. In this way, there is no need to determine whether each vehicle is a diagonally parked stationary vehicle as in the first perception fusion data, which can improve the efficiency of diagonal stationary parking recognition.

[0074] In an embodiment of the present application, by monitoring the relationship between the vehicle speed statistical information corresponding to the first stationary vehicle and the second preset speed statistical range, the stationary vehicles included in the first stationary vehicle corresponding to the first perception fusion data are enabled to leave the formation, and the newly appearing stationary vehicles in the second perception fusion data are enabled to enter the formation. This allows for dynamic maintenance of the stationary vehicles, ensuring that the stationary vehicles at different times are all obliquely parked stationary vehicles and conform to the group stationary characteristics.

[0075] Thus, the diagonally parked stationary vehicle identification method provided in an embodiment of the present application obtains first sensory fusion data and identifies a second vehicle parked diagonally therefrom. It then identifies the first stationary vehicle based on the relationship between a first preset speed statistical range and the speed statistical information of the identified second vehicle. It then updates the second sensory fusion data based on the first stationary vehicle and the newly identified third vehicle. This method utilizes information correlation and a dynamic update mechanism between consecutive moments to effectively improve the accuracy and stability of diagonally parked stationary vehicle identification. In particular, in scenarios with high-density diagonal parking in adjacent lanes, the introduction of group statistical characteristics instead of the traditional single-target time-series determination method significantly reduces the false braking rate and optimizes the overall performance of the intelligent driving system.

[0076] In some embodiments, the device for identifying an obliquely parked stationary vehicle may further perform the following steps when executing the above-mentioned step S101 of "determining at least one second vehicle that is obliquely parked stationary relative to the first vehicle at the first moment based on the first perception fusion data": determining each second vehicle that meets the oblique stationary parking condition in turn from the multiple vehicles included in the first perception fusion data; wherein the oblique stationary parking condition includes at least the following conditions: the lateral position of the vehicle is located in the effective area of ​​the adjacent lane of the first vehicle; the absolute value of the lateral speed of the vehicle is less than the preset lateral speed; the absolute value of the heading angle of the vehicle is within a preset angle range.

[0077] In the embodiments of this application, sequential determination means that the diagonally parked vehicle identification device sequentially evaluates each second vehicle that meets the diagonally parked condition among the multiple vehicles included in the first perception fusion data. This ensures stable processing of large amounts of target data and avoids data loss or misjudgment due to parallel processing. In practical applications, this operation is typically implemented by looping through the target list, with each vehicle's motion state independently evaluated.

[0078] Exemplarily, the first perception fusion data contains multiple vehicles: vehicle a, vehicle b, vehicle c, vehicle d, ..., vehicle p. For each of these vehicles, the characteristic information in the first perception fusion data is sequentially determined to determine whether it satisfies the diagonal stationary parking condition. Exemplarily, the diagonal stationary parking condition includes at least the following conditions: the vehicle's lateral position is within the valid area of ​​the adjacent lane of the first vehicle; the absolute value of the vehicle's lateral velocity is less than a preset lateral velocity; and the absolute value of the vehicle's heading angle is within a preset angle range.

[0079] In an embodiment of the present application, the lateral position of the vehicle may be the center of the rear axle of the vehicle, or the center of the rear of the vehicle, or the center point of the vehicle. For example, the lateral position of the vehicle may be selected based on actual needs and application scenarios, and this application does not limit this.

[0080] In an embodiment of the present application, the effective area of ​​the adjacent lane may be the area position of the adjacent lane of the first vehicle's driving lane. The effective area of ​​the adjacent lane may be the area within a preset range centered on the first vehicle, for example, the adjacent lane area less than 10m away from the first vehicle. Of course, it may also be the adjacent lane area less than 15m away from the first vehicle, or other adjacent lane areas at a distance from the first vehicle. For example, the effective area of ​​the adjacent lane may be set based on actual needs and application scenarios, and this application does not limit this.

[0081] In an embodiment of the present application, the preset lateral speed can be 0.4 m / s, or 0.38 m / s. Of course, it can be determined based on the perception fusion error in the actual scenario. If the perception fusion error is large, the preset lateral speed can be set larger. If the perception fusion error is small, the preset lateral speed can be set smaller. It can be set based on actual needs.

[0082] In the embodiment of the present application, the preset angle range may be [30°, 150°]. Of course, it may also be other angle ranges, which can be set based on actual needs and application scenarios.

[0083] For example, the oblique stationary stopping condition may be: the lateral position is located in the valid area of ​​the adjacent lane, the absolute value of the lateral speed is less than 0.4 m / s, and the absolute value of the heading angle is in the range of [30°, 150°].

[0084] In an embodiment of the present application, the oblique stationary parking condition covers comprehensive judgment criteria in three dimensions: the lateral position, lateral speed, and heading angle of the second vehicle. Through the joint verification of these three conditions, the second vehicle that is obliquely stationary relative to the first vehicle can be determined from the multiple vehicles included in the first perception fusion data. That is to say, for each of the multiple vehicles, it is determined whether it is the second vehicle based on the oblique stationary parking condition. In this way, the obliquely parked second vehicle that may be obstructed in the adjacent lane can be identified, thereby improving the robustness of the intelligent driving system in complex traffic environments.

[0085] In this way, by combining multiple dimensions of judgment conditions such as the effective area of ​​the adjacent lane, the absolute value of the lateral speed, and the heading angle range, each second vehicle that is stationary at an angle in the adjacent lane can be identified more accurately, thereby reducing the occurrence of false braking incidents and further improving the safety and user experience of the intelligent driving system.

[0086] In some embodiments, when executing the above-mentioned step S102, the identification device for an obliquely parked stationary vehicle may further include the following steps: for each second vehicle, in response to the number of determined second vehicles being greater than a first preset number and the vehicle speed statistical information corresponding to the determined second vehicles being within a first preset speed statistical range, the determined second vehicle is determined as a first stationary vehicle.

[0087] In embodiments of the present application, vehicle speed statistics may include parameters such as the mean lateral velocity and the variance of the lateral velocity, reflecting the overall motion distribution of these second vehicles. For example, if there are a large number of vehicles parked diagonally in an adjacent lane, the intelligent driving system will collect characteristic information (lateral velocity) of all identified second vehicles and, based on this characteristic information, determine the speed statistics of the identified second vehicles (mean lateral velocity and variance), thereby determining whether these second vehicles exhibit consistent stationary characteristics, i.e., whether they are within a first preset speed statistical range.

[0088] In an embodiment of the present application, the first preset number can be 5, 6, or other values. Generally, based on the central limit theorem, when the number is greater than 5, the distribution of the lateral velocity approaches a normal distribution, which can effectively reduce the probability of misjudgment.

[0089] In an embodiment of the present application, the first preset number is 5. Then, when vehicle 6 is determined to be the second vehicle, it is determined whether the speed statistical information of the 6 determined vehicles is within the first preset speed statistical range. If the speed statistical information of the 6 determined vehicles is within the first preset speed statistical range, vehicles 1 to 6 are determined to be the first stationary vehicles. When vehicle 7 is further determined to be the second vehicle, it is determined whether the speed statistical information of the 7 determined vehicles is within the first preset speed statistical range. If so, vehicle 7 is determined to be the first stationary vehicle. If not, vehicle 8 is determined.

[0090] In the embodiments of the present application, the mean lateral velocity and the lateral velocity variance have a complementary relationship. The mean lateral velocity is used to determine the overall motion trend, while the variance is used to determine the consistency of motion. Using these two together, it is possible to more accurately determine whether the first stationary vehicle is stationary. For example, when the mean lateral velocity is close to zero and the variance is small, it indicates that the lateral motion of these second vehicles is very stable, consistent with the characteristics of a group stationary state.

[0091] For example, if the average lateral velocity of multiple identified second vehicles is 0.2 m / s, but the actual values ​​fluctuate greatly, such as 1.0 m / s, 0.0 m / s, 0.0 m / s, 0.0 m / s, and 0.0 m / s, then the lateral velocity variance is large, indicating that the movement differences of these second vehicles in the lateral direction are large, and it is difficult to determine that they are in a unified stationary state.

[0092] In some embodiments, the vehicle speed statistical information includes a lateral speed mean and a lateral speed variance, and the first preset speed statistical range includes a first speed mean range and a first speed variance range; when the device for identifying an obliquely parked stationary vehicle executes the above-mentioned step S102, it may further include the following steps: for each second vehicle, in response to the determined lateral speed mean of the second vehicle being within the first speed mean range and the lateral speed variance being within the first speed variance range, the determined second vehicle is determined to be a first stationary vehicle.

[0093] In the embodiment of the present application, the lateral velocity mean is set to measure the overall lateral movement trend of the target vehicle. If the mean is close to zero, it means that the vehicle has almost no movement in the lateral direction; and the lateral velocity variance is used to measure the lateral velocity difference between each vehicle. If the variance is small, it means that the lateral velocity distribution between vehicles is relatively concentrated, and there is no obvious movement abnormality. For example, the lateral velocity of the second vehicle identified by the recognition device of the obliquely parked stationary vehicle from the first perception fusion data is 0.1m / s, 0.05m / s, 0.08m / s, 0.03m / s, and 0.07m / s, respectively. Its lateral velocity mean is about 0.066m / s, and the lateral velocity variance is about 0.000584m 2 / s 2 , are all within the first preset speed statistical range, so it can be determined that these second vehicles are first stationary vehicles. Among them, the first preset speed mean range in the first preset speed statistical range is [-0.2m / s, 0.2m / s], and the first speed variance range is less than 0.1m 2 / s 2 , i.e. 0 to 0.1m 2 / s 2 .

[0094] Exemplarily, for the sixth identified second vehicle, it will be determined whether the mean lateral speed of the six identified second vehicles is within the first speed mean range, and whether the lateral speed variance is within the first speed variance range. If the mean lateral speed of the six identified second vehicles is within the first speed mean range, and the lateral speed variance is within the first speed variance range, the second vehicles 1 to 6 are determined to be the first stationary vehicles. For the case where the determined vehicle 7 is the second vehicle, it will continue to be determined whether the speed statistical information of the seven identified vehicles is within the first preset speed statistical range. If so, vehicle 7 is determined to be the first stationary vehicle.

[0095] For example, if the lateral speeds of the five identified second vehicles are 0.1 m / s, 0.2 m / s, 0.05 m / s, 0.15 m / s, and 0.08 m / s, respectively, then their lateral speed average is approximately 0.116 m / s, which is within the first speed average range. This indicates that the lateral movement of these second vehicles is relatively stable and close to a stationary state. Further judgment of the lateral speed variance is required. If the lateral speed variance is also within the first speed variance range, it is determined to be the first stationary vehicle.

[0096] By incorporating vehicle speed statistics—the mean and variance of lateral speed—into a decision-making process, multi-target collaborative judgment can be achieved based on sensory fusion data, overcoming the misjudgment problem caused by sensor jitter in traditional methods. This effectively improves the accuracy of stationary vehicle recognition, thereby preventing false braking and ultimately enhancing the safety and user experience of intelligent driving systems.

[0097] In some embodiments, when the device for identifying an obliquely parked stationary vehicle executes the above step S104 of "updating the first stationary vehicle based on the feature information of the first stationary vehicle in the second perception fusion data to obtain the second stationary vehicle", as follows Figure 3 As shown, the following steps S301 and S302 may be included:

[0098] Step S301: Determine first vehicle speed statistical information corresponding to the first stationary vehicle based on feature information of the first stationary vehicle in the second perception fusion data.

[0099] In an embodiment of the present application, the characteristic information of the first stationary vehicle includes the lateral velocity of the first stationary vehicle. By calculating vehicle speed statistics based on the lateral velocity of each vehicle in the first stationary vehicle, first vehicle speed statistics corresponding to the first stationary vehicle at the second moment can be obtained, such as the mean and variance of the lateral velocity.

[0100] In an embodiment of the present application, the first stationary vehicle is a stationary vehicle in the first perception fusion data, but the characteristic information in the second perception fusion data may indicate a change in the motion state of a certain first stationary vehicle. At this time, the identification device of the obliquely parked stationary vehicle can determine whether the first stationary vehicle corresponding to the first perception fusion data is still within the first preset speed statistical range based on the characteristic information in the second perception fusion data. If so, continue to determine whether the third vehicle different from the first stationary vehicle in the second perception fusion data is a stationary vehicle. If not, execute step S302.

[0101] Step S302: In response to the first vehicle speed statistical information not being within a second preset speed statistical range, a moving vehicle is determined from the first stationary vehicles based on the characteristic information and is deleted.

[0102] In an embodiment of the present application, if the first vehicle speed statistical information is not within the second preset speed statistical range, it means that the lateral speed of a certain vehicle among the first stationary vehicles presents an abnormal value that is different from the normal value of the lateral speed distribution of the stationary vehicles. In this way, the certain stationary vehicle among the first stationary vehicles is most likely a starting target that is transitioning from stationary to moving. In this way, the vehicle with the largest lateral speed among the first stationary vehicles can be deleted.

[0103] For example, if the first perception fusion data corresponds to six first stationary vehicles, and if the first vehicle speed statistics are not within the second preset speed statistical range, the vehicle with the highest lateral speed among the first stationary vehicles is deleted. At this point, there are five first stationary vehicles. The diagonally parked vehicle identification device will continue to determine the relationship between the first vehicle speed statistics corresponding to the remaining five and the second preset speed statistical range, and then determine whether to stop deleting the first stationary vehicles. If the first vehicle speed statistics are within the second preset speed statistical range, the first stationary vehicles are stopped from being deleted.

[0104] In the embodiments of the present application, the second preset speed statistical range is also a set of statistical indicators used to determine whether the target vehicles constitute a stable stationary group. The second preset speed statistical range may or may not be consistent with the first preset speed statistical range and can be set based on actual needs and application scenarios, and this application does not impose any restrictions on this.

[0105] For example, the second preset speed statistical range includes a second speed mean range and a second speed variance range. The second speed mean range may be [-0.25 m / s, 0.25 m / s], and the second speed variance range may be 0 to 0.12 m / s. 2 / s 2 Here, the second preset speed statistical range is slightly larger than the first preset speed statistical range, which can provide a buffer for the deleted first stationary vehicle to avoid misidentification.

[0106] In this way, by continuously monitoring the statistical characteristics of stationary vehicles at consecutive moments, when it is detected that they deviate from the preset range, targets that may have changed from stationary to moving can be eliminated in a timely manner to avoid misidentification as stationary, thereby improving the real-time performance and accuracy of the system.

[0107] In some embodiments, the device for identifying an obliquely parked stationary vehicle may further perform the following steps: in response to the first vehicle speed statistical information being within a second preset speed statistical range, or the number of first stationary vehicles being less than a second preset number, at least one third vehicle parked obliquely relative to the first vehicle determined based on the second perception fusion data is updated to obtain a second stationary vehicle.

[0108] In an embodiment of the present application, if the first vehicle speed statistical information is within the second preset speed statistical range, the first stationary vehicle is updated based on the third vehicle parked obliquely relative to the first vehicle determined in the second perception fusion data to obtain the second stationary vehicle.

[0109] In an embodiment of the present application, if the number of the first stationary vehicles is less than the second preset number, the first stationary vehicle is updated based on the third vehicle parked obliquely relative to the first vehicle determined in the second perception fusion data to obtain the second stationary vehicle.

[0110] In an embodiment of the present application, the second preset number may be 4, 3 or other numbers, and may be set based on actual needs and application scenarios, which is not limited in this application.

[0111] For example, if the number of first stationary vehicles corresponding to the first perception fusion data is 5 and the second preset number is 4, then when the first vehicle speed statistical information is not within the second preset speed statistical range, the vehicle with the largest lateral speed among the first stationary vehicles will be deleted. At this time, there are 4 first stationary vehicles, and the identification device for diagonally parked stationary vehicles will continue to judge the relationship between the first vehicle speed statistical information corresponding to the remaining 4 vehicles and the second preset speed statistical range. If the first vehicle speed statistical information corresponding to the remaining 4 vehicles is not within the second preset speed statistical range, the vehicle with the largest lateral speed among the first stationary vehicles will continue to be deleted. There are 3 first stationary vehicles, which satisfies the requirement that the number of first stationary vehicles is less than the second preset number. In this case, the first stationary vehicle will also be updated based on the third vehicle that is diagonally parked relative to the first vehicle determined in the second perception fusion data.

[0112] In this way, by introducing a new third vehicle to make entry judgments, the dynamic maintenance of the target set can be achieved while ensuring the stability of the existing targets, thereby improving the system's adaptability to complex scenarios.

[0113] In some embodiments, when the device for identifying an obliquely parked stationary vehicle performs the above step of "updating the first stationary vehicle to obtain a second stationary vehicle based on at least one third vehicle parked obliquely relative to the first vehicle determined by the second perception fusion data", as follows: Figure 4 As shown, the following steps S401 and S402 may be included:

[0114] Step S401: Determine in sequence each third vehicle that meets the oblique stationary parking condition from the plurality of vehicles included in the second perception fusion data; the third vehicle is a vehicle that is different from the first stationary vehicle in the second perception fusion data.

[0115] In an embodiment of the present application, the identification device for an obliquely parked stationary vehicle determines each third vehicle that meets the oblique stationary parking condition from the multiple vehicles included in the second perception fusion data in turn. For the first stationary vehicle corresponding to the first perception fusion data, the identification device for an obliquely parked stationary vehicle does not make a judgment here, and it is either just deleted or the first stationary vehicle. Therefore, what is judged here is the third vehicle that is different from the first stationary vehicle in the first perception fusion data.

[0116] In an embodiment of the present application, a third vehicle that meets the diagonal stationary parking condition is sequentially determined for multiple vehicles included in the second perception fusion data. For example, if the second perception fusion data includes vehicles 7, 8, and 9, which are distinct from the first stationary vehicle, the third vehicle is sequentially determined based on the feature information corresponding to each of these vehicles in the second perception fusion data to determine whether they meet the diagonal stationary parking condition. If all of these vehicles meet the condition, then vehicles 7, 8, and 9 are considered the third vehicle.

[0117] Step S402: For each determined third vehicle, in response to the vehicle speed statistical information corresponding to the determined third vehicle and the first stationary vehicle being within a first preset speed statistical range, determine the determined third vehicle as a second stationary vehicle.

[0118] For example, if the third vehicle 7 satisfies the oblique stationary parking condition, and after executing the above-mentioned step S302, the first stationary vehicles are vehicles 1 to 4, then it will be determined whether the speed statistical information corresponding to vehicles 1 to 4 and the third vehicle 7 is within the first preset speed statistical range. If so, the third vehicle 7 is determined as the second stationary vehicle. If not, the judgment of the third vehicle 8 that meets the oblique stationary parking condition is continued until all the third vehicles obtained in step S401 are judged and the second stationary vehicle corresponding to the second perception fusion data is obtained.

[0119] In this way, by introducing a collaborative judgment mechanism based on group statistical characteristics, efficient verification of new targets can be achieved without relying on multi-frame data, reducing computational complexity and improving recognition efficiency. At the same time, this method can also enhance the recognition ability of real moving vehicles and reduce the occurrence of false braking.

[0120] In some embodiments, the device for identifying obliquely parked stationary vehicles may further perform the following step: in response to the number of the second stationary vehicles being less than a third preset number, deleting the second stationary vehicles.

[0121] In an embodiment of the present application, if, after executing step S104, the number of the second stationary vehicles is still less than the third preset number, then the first vehicle in the current second perception fusion data does not belong to a dense diagonally parked vehicle scene, i.e., there are relatively few diagonally parked vehicles. There is no need to separately determine the number of diagonally parked vehicles, and the current scene can be released. The third preset number can be 3, 4, or another number. The relationship between the third preset number, the first preset number, and the second preset number can also be set based on actual needs and application scenarios, and this application does not impose any limitations on this.

[0122] In this way, by setting a minimum number threshold for the second stationary vehicle, misjudgment due to too few targets can be prevented, ensuring the stability of the system in low-confidence scenarios.

[0123] In some embodiments, the device for identifying an obliquely parked stationary vehicle may further perform the following steps: in response to the number of at least one second vehicle being less than or equal to a first preset number, or the vehicle speed statistical information corresponding to at least one second vehicle being not within the first preset speed statistical range, determining that the first vehicle is in the first scene information; the first scene information characterizing that there is no corresponding first stationary vehicle in the first perception fusion data.

[0124] In an embodiment of the present application, if the number of all determined at least one second vehicle is less than or equal to the first preset number, it means that the number of second vehicles parked diagonally relative to the first vehicle is small, and the scene in which the first vehicle is located is a non-dense diagonally parked stationary vehicle scene. In this way, there is no need to determine the first stationary vehicle corresponding to the first perception fusion data.

[0125] In an embodiment of the present application, the first preset number can be set to 5, that is, when the number of at least one second vehicle identified is less than 5, it is considered that the conditions for forming a stationary vehicle array are not met. The first preset number can be adjusted according to the actual sensor accuracy and road scene characteristics to adapt to the distribution of diagonally parked vehicles with different densities.

[0126] In an embodiment of the present application, when the number of at least one identified second vehicle exceeds a first preset number, but the speed statistics of the at least one identified second vehicle do not fall within the first preset speed statistical range, this indicates that the current distribution of vehicles in the adjacent lane is insufficient to form a stable array of diagonally parked, stationary vehicles. In this case, the diagonally parked, stationary vehicle identification device determines that the first vehicle's environment does not constitute a dense diagonally parked, stationary vehicle scene, and thus enters the first scene information state. This judgment logic helps avoid misidentifying normally moving vehicles as stationary vehicles in low-density or uncertain motion situations, thereby improving the robustness and accuracy of the system.

[0127] Thus, determining that the first vehicle is in the first scenario indicates that there are not a large number of diagonally parked vehicles in the adjacent lane, eliminating the need for special processing logic for a stationary array. This conclusion not only avoids emergency braking events caused by misjudgments but also saves computing resources and improves overall system efficiency.

[0128] In some embodiments, the device for identifying an obliquely parked stationary vehicle may further perform the following steps: setting a preset buffer time for a moving vehicle, and stopping identifying the moving vehicle within the preset buffer time.

[0129] In the embodiments of the present application, the preset buffer time refers to a period of time set after a vehicle is determined to be a moving vehicle in order to avoid misjudgment due to brief perception errors or environmental interference. During this period, further identification and status updates of the moving vehicle will be suspended to ensure the stability of the judgment results. The preset buffer time is typically set based on the complexity of the scene and the response speed of the system. For example, it can be set to between 50ms and 200ms. During the buffer period, the moving vehicle will no longer be dynamically tracked and will not be included in the subsequent target screening process. Only when the buffer time expires will the recognition process re-enter.

[0130] In an embodiment of the present application, after the moving vehicle is deleted from the first stationary vehicle in the above step S302, a buffer time is set for the moving vehicle. Therefore, for the third vehicle determined to be stationary at an angle relative to the first vehicle based on the second perception fusion data, if there is a vehicle within the preset buffer time among the third vehicles, no judgment is made, or, for the vehicle within the preset buffer time in the second perception fusion data, no identification is performed, that is, it is not judged whether it is stationary at an angle relative to the first vehicle.

[0131] In this way, by introducing a buffer mechanism, the recognition instability caused by the frequent state switching of the moving vehicle in a short period of time can be avoided, and the smoothness and robustness of the operation can be improved.

[0132] In one embodiment of this application, a method for identifying diagonally parked stationary vehicles based on multi-target collaborative analysis is proposed. This method aims to address the problem of motion state misjudgment caused by the decay of perception attribute confidence in scenarios with a high density of diagonally parked stationary vehicles in adjacent lanes. By constructing a collaborative judgment mechanism (a single-target and multi-target joint judgment mechanism) for a dynamic array of stationary vehicles, this method effectively suppresses the misjudgment of stationary vehicles caused by traditional single-target independent judgment methods, while avoiding the omission of actual moving vehicles. This diagonally parked stationary vehicle recognition method can be integrated into the target selection module of an intelligent driving system, serving as a motion state correction unit between the perception fusion layer and the planning and decision-making layer. After obtaining the stationary vehicle corresponding to each frame, the stationary vehicle can be sent to the planning and decision-making layer.

[0133] For example, Figure 5 As shown, an exemplary implementation of a data processing method is provided, including the following steps S501 to S505:

[0134] Step S501: The sensor obtains target and road environment information.

[0135] Here, sensors (such as lidar, visual sensors, etc.) obtain feature information (perception fusion data) corresponding to the target (vehicle) and road environment information (obstacle information).

[0136] Step S502: The perception fusion algorithm processes the target information.

[0137] Here, the acquired perception data corresponding to the target is sent to the recognition device of the obliquely parked vehicle, and the target information is processed using the perception fusion algorithm.

[0138] Step S503: Target state judgment key target selection.

[0139] Here, based on target information, the target status is determined, and key targets are selected. For example, for a single target, the initial screening criteria are determined based on whether the target meets the diagonal stationary stop condition. For multiple targets, key target selection is based on a coordinated multi-target determination (whether the speed statistics of the multiple targets are within a first or second preset speed statistical range).

[0140] Step S504: Processing road environment information.

[0141] Here, the road environment information acquired by the sensor, such as obstacles, pedestrians, etc., can be processed by the environmental information processing module of the identification device of the diagonally parked stationary vehicle, or the processed road environment information can be sent to the planning decision-making layer independently of the identification device of the diagonally parked stationary vehicle.

[0142] Step S505: Horizontal and vertical planning decision and control.

[0143] Here, planning, decision-making and control are performed based on the processed road environment information obtained in step S504 and the key target obtained in step S503.

[0144] The implementation of the above steps S502 and S503 may include the following steps S1 to S4:

[0145] For example, S1 pre-screens all vehicle targets output by the current perception fusion layer and sets the initial screening conditions (corresponding to the oblique stationary parking conditions discussed above) as follows:

[0146] (1) Target lateral position Adjacent lane effective area;

[0147] (2) Absolute value of target lateral velocity ( Set the actual perception fusion error according to the scenario, see the preset lateral speed discussed above);

[0148] (3) Target heading angle (For the typical angle range of closely parked vehicles, please refer to the preset angle range discussed above. and is the angle);

[0149] The n targets that meet the conditions constitute the stationary vehicle candidate set ,in, For the first vehicle that meets the initial screening conditions, For the second vehicle that meets the initial screening conditions, For the Vehicles that meet the initial screening criteria, For the vehicles that meet the initial screening criteria.

[0150] S2. Dynamic stationary array modeling and collaborative verification:

[0151] Calculate the statistical characteristics of the stationary vehicle candidate set s (corresponding to the vehicle speed statistics discussed above): mean lateral speed , and the lateral velocity variance ;in, For the The lateral velocity of a target (vehicle).

[0152] Based on the statistical characteristics, the array generation conditions (corresponding to the conditions for determining the stationary vehicle corresponding to the perception fusion data) can be set as follows:

[0153] (1) ; ( is the first speed mean range)

[0154] (2) ( According to the actual perception fusion error setting in the scene, is the first speed variance range;

[0155] (3) Number of candidate set targets (recommend = 5, to ensure statistical significance);

[0156] When the conditions are met, a dynamic stationary vehicle array S=s is created, and all targets in the array are stationary targets (corresponding to the first stationary vehicle discussed above).

[0157] Considering that, under the same sensor and perception fusion algorithm, the lateral velocity sequence of multiple stationary targets in the same frame should be equivalent to the multi-frame random sampling of the lateral velocity of a single stationary target over a certain period of time, a multi-target single-frame joint judgment can be used to replace the traditional single-target multi-frame time series judgment. This overcomes the problem of single-target perception jitter by leveraging the statistical characteristics of the group. Based on the central limit theorem, when n ≥ 5, the distribution of target lateral velocity approaches a normal distribution, effectively reducing the probability of misjudgment.

[0158] S3. After the stationary vehicle array S is created, the array dynamic maintenance mechanism is used to determine whether old targets are allowed to exit the array and whether new targets are allowed to enter the array:

[0159] The exit judgment establishes a motion state transfer model. When the array statistical characteristics (corresponding to the vehicle speed statistical information discussed above) deviate significantly (not within the second preset speed statistical range), the target that changes from stationary to moving will be removed from the array. The entry judgment uses a double verification mechanism to include new stationary targets in the stationary vehicle array to ensure the consistency of the newly added targets.

[0160] The conditions for going out (second preset speed statistical range) are:

[0161] (1) ;

[0162] (2) ( and is the hysteresis coefficient to prevent the number of array targets from fluctuating);

[0163] When the conditions for exiting the array are met, the target with the largest lateral speed in the array is eliminated (assuming the speed is positive in the direction of the vehicle).

[0164] The entry conditions are: 1) Basic condition verification, where the target meets the same initial screening criteria; 2) Group compatibility verification, where the statistical characteristics of the array's lateral velocity (corresponding to the vehicle speed statistics discussed above) after the target is pre-added to array S meet the array generation conditions. The key to entry and exit operations lies in distinguishing true moving targets. The core logic is that when a true moving target is introduced into an array of stationary vehicles, the distribution consistency of the target's perceptual attributes and those of the array targets is disrupted. Therefore, the presence of a moving target in the array can be determined by monitoring the statistical characteristics of the array target's lateral velocity.

[0165] S4. The life cycle of the array needs to be managed. ( is the hysteresis coefficient to prevent scene switching oscillation), the array is destroyed, the motion status flags of all targets in the array are reset, and the scene exit event is triggered.

[0166] The advantages that can be achieved by steps S1 to S4 are: first, single-frame processing replaces multi-frame caching, which significantly reduces memory usage; second, compared with the multi-frame caching judgment method, the computational complexity is reduced from O(kn) to O(n) (k is the time window length); third, multi-target collaborative judgment and recognition of the scene can reduce the false braking rate without introducing missed braking.

[0167] In another embodiment of the present application, Figure 6 As shown, the exemplary method for identifying an obliquely parked vehicle may include the following steps S601 to S605:

[0168] Step S601: Whether there is a stationary vehicle array.

[0169] Here, at the beginning of each frame, it is determined whether the current frame already has a stationary vehicle array S (corresponding to the first stationary vehicle in the first perception fusion data discussed above). If not, the process proceeds to step S602; if the stationary vehicle array already exists, the process proceeds to step S603.

[0170] Step S602: Whether the creation conditions are met.

[0171] Here, step S602 includes step S6021 and step S6024, wherein:

[0172] Step S6021: Determine whether the first target meets the conditions.

[0173] Here, we determine whether the first target (corresponding to the first vehicle in the first sensory fusion data discussed above) meets the following conditions (corresponding to the oblique stationary parking conditions discussed above): its lateral position is in the adjacent lane, its absolute lateral velocity is less than 0.4 m / s (corresponding to the preset lateral velocity discussed above), and its absolute heading angle is within the range [30°, 150°] (corresponding to the preset angle range discussed above). If so, the target is added to the candidate set s.

[0174] Step S6022: Determine whether the number of targets included in the candidate set is greater than 5.

[0175] Here, the system determines whether the number of targets (vehicles) in the current candidate set s (corresponding to the number of at least one second vehicle discussed above) is greater than 5 (corresponding to the first preset number discussed above); if the number of targets is greater than 5, proceed to step S6023; otherwise, return to step S6021 and continue processing the next target.

[0176] Step S6023: Calculate the mean and variance of the lateral velocity.

[0177] Here, the lateral velocity mean and lateral velocity variance of the potential stationary vehicle array target (corresponding to the determined vehicles that meet the oblique stationary parking conditions corresponding to the current perception fusion data discussed above) are calculated. If the lateral velocity mean is within the threshold range [-0.2m / s, 0.2m / s] (corresponding to the first velocity mean range discussed above), and the lateral velocity variance is less than 0.1m 2 / s 2 (corresponding to the first speed variance range discussed above), the candidate set s is converted into a stationary vehicle array S, and step S604 (new target enters the array) is entered at the same time.

[0178] Step S6024: Determine whether the creation conditions are met.

[0179] Here, if all targets have completed the above processing steps and have not yet entered step S6023, step S607 is executed.

[0180] Step S603: Go out into battle.

[0181] Here, S603 includes S6031 and S6032:

[0182] Step S6031: Calculate the mean and variance of the lateral velocity.

[0183] Here, the lateral velocity mean and lateral velocity variance (corresponding to the first preset velocity statistical range discussed above) of the stationary vehicle array S target (corresponding to the first stationary vehicle discussed above) are calculated to determine whether the lateral velocity mean is in the interval [-0.25m / s, 0.25m / s] (corresponding to the second velocity mean range discussed above) and whether the lateral velocity variance is less than 0.12m 2 / s 2 (Corresponding to the second speed variance range discussed above).

[0184] Step S6032: Determine whether it is necessary to go out into battle.

[0185] Here, if the condition is met (corresponding to the preset speed statistical range discussed above), it means that the lateral speed of a target in the array (corresponding to a vehicle in the first stationary vehicle discussed above) shows an abnormal value that is different from the normal value of the lateral speed distribution of stationary vehicles (the largest value is removed). This target is very likely a starting target that has transitioned from stationary to moving, and the process proceeds to S6033.

[0186] If the conditions are not met, it means that there is no obvious abnormality in the statistical values ​​of all targets in the array, the lateral speed is still within the normal value range of the lateral speed distribution of a stationary vehicle, and there is no target in the array that needs to be removed from the array. At this time, step S604 is entered (new target enters the array).

[0187] Step S6033: Eliminate the target with the largest lateral speed.

[0188] Here, the target with the highest lateral velocity in array S (assuming the velocity is positive in the direction of the vehicle) is removed from the array. If the controller detects that the number of targets in array S is greater than 3 (corresponding to the second preset number discussed above), the controller returns to S6031. If the number of targets currently detected in array S is less than or equal to 3, S604 is executed (the new target is added to the array).

[0189] Note that a cooldown period should be set for the target object that exits the array in the current step. During this cooldown period, the target object will not be able to re-enter the array.

[0190] Step S604: Enter the formation.

[0191] Here, step S604 includes step S6041 and step S6042:

[0192] Step S6041: Determine whether the new target meets the conditions.

[0193] Here, we determine whether the new target meets the following conditions (corresponding to the oblique stationary stop conditions discussed above): the lateral position of the new target is in the adjacent lane, the absolute value of the lateral velocity of the new target is less than 0.4 m / s, and the absolute value of the heading angle of the new target is in the range of [30°, 150°].

[0194] If the detected target meets the determination conditions of a stationary target, the system determines that the target is a stationary target and transfers the processing flow to step S6042.

[0195] If the detection result does not meet the preset judgment condition, the target does not belong to the type of diagonally parked stationary vehicle and therefore cannot enter the parking array. At this time, the control process returns to step S6041 to continue to judge whether the next new target is a diagonally parked stationary vehicle.

[0196] Step S6042: pre-add the target and verify.

[0197] Here, the target is pre-added to the array S, and the lateral velocity mean and lateral velocity variance (corresponding to the velocity statistics discussed above) of the array target (corresponding to the current perception fusion data discussed above and the determined vehicle that meets the oblique stationary parking condition) are calculated to see whether they meet the conditions: the lateral velocity mean is within the threshold range [-0.2m / s, 0.2m / s] (corresponding to the first velocity mean range discussed above), and the lateral velocity variance is less than 0.1m 2 / s 2 (Corresponding to the first speed variance range discussed above).

[0198] If the conditions are met, it is determined that the currently detected target is a stationary target, the currently detected target is put into the array, and step S6041 is repeated to determine the next new target.

[0199] If the condition is not met, it means that the target is most likely not a stationary target, so it cannot be included in the stationary vehicle array, and then step S6041 is repeated to determine whether the next newly identified target is a stationary target.

[0200] After completing the above processing for all new targets, proceed to step S605.

[0201] Step S605: Is n greater than 3?

[0202] Here, the number of targets in the stationary vehicle array S is determined. If the target number n is less than or equal to 3 (corresponding to the third preset number discussed above), step S607 is executed; if the target number is greater than 3, step S606 is executed.

[0203] Step S606: Output the stationary vehicle array.

[0204] Here, the stationary car array S is saved until the next frame calculation, and all targets in the array are marked as stationary cars and output.

[0205] Step S607: The vehicle is not parked at an angle.

[0206] Here, the array S is destroyed, the correlation calculation of the current frame is completed, and no target is obtained as a stationary car mark.

[0207] In this way, based on the target stationary vehicle, the longitudinal response strategy can be optimized for scenarios where there are a large number of stationary vehicles parked diagonally (including sideways) in adjacent lanes, reducing false braking and improving user experience.

[0208] An embodiment of the present application provides a method for identifying obliquely parked stationary vehicles, which includes obtaining first perception fusion data collected by a first vehicle at a first moment, and determining at least one second vehicle that is obliquely parked stationary relative to the first vehicle at the first moment based on the first perception fusion data; for each second vehicle, determining a first stationary vehicle from at least one second vehicle based on the relationship between the determined vehicle speed statistical information of the second vehicle and a first preset speed statistical range; obtaining second perception fusion data collected by the first vehicle at a second moment; and updating the first stationary vehicle based on the feature information of the first stationary vehicle in the second perception fusion data and at least one third vehicle that is obliquely parked stationary relative to the first vehicle determined based on the second perception fusion data, thereby obtaining a second stationary vehicle. The embodiment of the present application combines single-vehicle features and group features to construct a first stationary vehicle, which can improve the robustness of recognition, and, in subsequent moments, dynamically updates based on the feature information of the first stationary vehicle and newly occurring stationary states to improve recognition efficiency.

[0209] The embodiment of the present application provides a device 7 for identifying a vehicle parked at an angle, such as Figure 7 As shown, including:

[0210] an identification module 71 configured to obtain first perception fusion data collected by the first vehicle at a first moment, and determine, based on the first perception fusion data, at least one second vehicle that is parked obliquely relative to the first vehicle at the first moment;

[0211] a determination module 72 configured to determine, for each second vehicle, a first stationary vehicle from the at least one second vehicle based on a relationship between the determined vehicle speed statistical information of the second vehicle and a first preset speed statistical range;

[0212] An acquisition module 73 is configured to acquire second perception fusion data collected by the first vehicle at a second moment, where the second moment is continuous with the first moment;

[0213] The updating module 74 is configured to update the first stationary vehicle based on the feature information of the first stationary vehicle in the second perception fusion data and at least one third vehicle that is parked obliquely relative to the first vehicle determined based on the second perception fusion data to obtain a second stationary vehicle.

[0214] In one embodiment of the present application, the identification module 71 is further used to determine, based on the first perception fusion data, at least one second vehicle that is parked obliquely relative to the first vehicle at a first moment, including: determining each second vehicle that meets the oblique stationary parking condition in turn from the multiple vehicles included in the first perception fusion data; wherein the oblique stationary parking condition includes at least the following conditions: the lateral position of the vehicle is located in the effective area of ​​the adjacent lane of the first vehicle; the absolute value of the lateral speed of the vehicle is less than the preset lateral speed; the absolute value of the heading angle of the vehicle is within a preset angle range.

[0215] In one embodiment of the present application, the determination module 72 is further configured to, for each second vehicle, determine, based on a relationship between the determined vehicle speed statistical information of the second vehicle and the first preset speed statistical range, a first stationary vehicle from at least one second vehicle, including: for each second vehicle, in response to the number of determined second vehicles being greater than the first preset number and the vehicle speed statistical information corresponding to the identified second vehicle being within the first preset speed statistical range, determining the determined second vehicle as the first stationary vehicle.

[0216] In one embodiment of the present application, the vehicle speed statistical information includes a lateral speed mean and a lateral speed variance, and the first preset speed statistical range includes a first speed mean range and a first speed variance range; the determination module 72 is further used to, for each second vehicle, determine the determined second vehicle as a first stationary vehicle in response to the determined second vehicle's lateral speed mean being within the first speed mean range and the lateral speed variance being within the first speed variance range.

[0217] In one embodiment of the present application, the determination module 72 is further used to determine first vehicle speed statistical information corresponding to the first stationary vehicle based on characteristic information of the first stationary vehicle in the second perception fusion data; in response to the first vehicle speed statistical information not being within the second preset speed statistical range, the moving vehicle is determined from the first stationary vehicle based on the characteristic information and deleted.

[0218] In one embodiment of the present application, the updating module 74 is further configured to update the first stationary vehicle to obtain a second stationary vehicle in response to the first vehicle speed statistical information being within a second preset speed statistical range or the number of the first stationary vehicles being less than the second preset number, based on at least one third vehicle that is stationary and parked obliquely relative to the first vehicle as determined by the second perception fusion data.

[0219] In one embodiment of the present application, the update module 74 is further used to determine, in sequence, each third vehicle that meets the oblique stationary parking condition from the multiple vehicles included in the second perception fusion data; the third vehicle is a vehicle that is different from the first stationary vehicle in the second perception fusion data; for each determined third vehicle, in response to the vehicle speed statistical information corresponding to the determined third vehicle and the first stationary vehicle being within the first preset speed statistical range, the determined third vehicle is determined as the second stationary vehicle.

[0220] In an embodiment of the present application, the updating module 74 is further configured to delete the second stationary vehicle in response to the number of the second stationary vehicles being less than a third preset number.

[0221] In one embodiment of the present application, the update module 74 is further used to determine that the first vehicle is in the first scene information in response to the number of at least one second vehicle being less than or equal to the first preset number, or the vehicle speed statistical information corresponding to at least one second vehicle is not within the first preset speed statistical range; the first scene information represents that there is no corresponding first stationary vehicle in the first perception fusion data.

[0222] In one embodiment of the present application, the updating module 74 is further configured to set a preset buffer time for the moving vehicle and stop identifying the moving vehicle within the preset buffer time.

[0223] An embodiment of the present application provides a device for identifying obliquely parked stationary vehicles. The device obtains first perception fusion data collected by a first vehicle at a first moment, and based on the first perception fusion data, determines at least one second vehicle that is obliquely parked stationary relative to the first vehicle at the first moment; for each second vehicle, the device determines a first stationary vehicle from at least one second vehicle based on the relationship between the determined vehicle speed statistical information of the second vehicle and a first preset speed statistical range; obtains second perception fusion data collected by the first vehicle at a second moment; updates the first stationary vehicle based on the feature information of the first stationary vehicle in the second perception fusion data, and at least one third vehicle that is obliquely parked stationary relative to the first vehicle determined based on the second perception fusion data, to obtain a second stationary vehicle. Using the above implementation scheme, the embodiment of the present application combines single-vehicle characteristics and group characteristics to construct a first stationary vehicle, which can improve the robustness of the determination. Moreover, in subsequent moments, the device dynamically updates the first stationary vehicle based on the feature information of the first stationary vehicle and the newly occurring stationary state, thereby improving the recognition efficiency.

[0224] The embodiment of the present application provides a device 8 for identifying a car parked at an angle. In practical applications, based on the same disclosed concept of the above embodiment, as shown in FIG. Figure 8 As shown, the device 8 for identifying an obliquely parked vehicle according to the embodiment of the present application includes: a processor 81, a memory 82 and a communication bus 83;

[0225] A communication bus 83 is used to implement communication between the processor 81 and the memory 82;

[0226] The processor 81 is configured to execute the computer program stored in the memory 82 to implement the above-mentioned method for identifying a stationary vehicle parked at an angle.

[0227] An embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. The computer-readable storage medium stores one or more programs, which can be executed by one or more processors. The computer program implements the above-mentioned method for identifying an obliquely parked stationary vehicle.

[0228] An embodiment of the present application provides a computer program product, including a computer program or instructions. When the computer program or instructions are executed by the processor 81, the above-mentioned method for identifying an obliquely parked stationary vehicle is implemented.

[0229] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage and optical storage) containing computer-usable program code.

[0230] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0231] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0232] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0233] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the above scope of protection.

Claims

1. A method for identifying a vehicle parked at an angle, characterized in that: The method comprises: Obtaining first perception fusion data collected by a first vehicle at a first moment, and determining, based on the first perception fusion data, at least one second vehicle that is parked obliquely relative to the first vehicle at the first moment; For each second vehicle, determining a first stationary vehicle from the at least one second vehicle based on a relationship between the determined vehicle speed statistical information of the second vehicle and a first preset speed statistical range, wherein the vehicle speed statistical information is a lateral speed mean or a lateral speed variance; Acquire second perception fusion data collected by the first vehicle at a second moment, where the second moment is continuous with the first moment; Based on the feature information of the first stationary vehicle in the second perception fusion data and at least one third vehicle parked obliquely relative to the first vehicle determined based on the second perception fusion data, the first stationary vehicle is updated to obtain a second stationary vehicle.

2. The method for identifying a vehicle parked at an angle according to claim 1, wherein: The determining, based on the first perception fusion data, at least one second vehicle that is parked obliquely relative to the first vehicle at the first moment includes: sequentially determining each second vehicle that meets an oblique stationary parking condition from a plurality of vehicles included in the first perception fusion data; Among them, the oblique stationary parking conditions include at least the following conditions: the lateral position of the vehicle is located in the effective area of ​​the adjacent lane of the first vehicle; the absolute value of the lateral speed of the vehicle is less than the preset lateral speed; the absolute value of the heading angle of the vehicle is within the preset angle range.

3. The method for identifying a vehicle parked at an angle according to claim 2, wherein: The method of determining, for each of the second vehicles, a first stationary vehicle from the at least one second vehicle based on the determined relationship between the vehicle speed statistical information of the second vehicle and a first preset speed statistical range includes: For each second vehicle, in response to the number of the determined second vehicles being greater than a first preset number and the vehicle speed statistical information corresponding to the determined second vehicles being within the first preset speed statistical range, the determined second vehicle is determined as the first stationary vehicle.

4. The method for identifying a vehicle parked at an angle according to claim 1, wherein: The vehicle speed statistical information includes a lateral speed mean and a lateral speed variance, and the first preset speed statistical range includes a first speed mean range and a first speed variance range. The determining, for each second vehicle, of a first stationary vehicle from the at least one second vehicle based on a relationship between the determined vehicle speed statistical information of the second vehicle and the first preset speed statistical range comprises: For each of the second vehicles, in response to a determined lateral velocity mean of the second vehicle being within a first velocity mean range and a lateral velocity variance being within the first velocity variance range, the determined second vehicle is determined as the first stationary vehicle.

5. The method for identifying a vehicle parked at an angle according to claim 1, wherein: The updating of the first stationary vehicle based on the feature information of the first stationary vehicle in the second perception fusion data and at least one third vehicle parked obliquely relative to the first vehicle determined based on the second perception fusion data to obtain a second stationary vehicle includes: determining first vehicle speed statistical information corresponding to the first stationary vehicle based on feature information of the first stationary vehicle in the second perception fusion data; In response to the first vehicle speed statistical information not being within a second preset speed statistical range, determining a moving vehicle from the first stationary vehicles based on the characteristic information, and updating the first stationary vehicle based on the moving vehicle to obtain a second stationary vehicle; sequentially determining, from the plurality of vehicles included in the second perception fusion data, each of the third vehicles that meets the oblique stationary parking condition; the third vehicle being a vehicle that is different from the first stationary vehicle in the second perception fusion data; For each of the determined third vehicles, in response to the vehicle speed statistical information corresponding to the determined third vehicle and the second stationary vehicle being within the first preset speed statistical range, the corresponding third vehicle is determined to be the second stationary vehicle.

6. The method for identifying a vehicle parked at an angle according to claim 5, characterized in that: The method further comprises: In response to the first vehicle speed statistical information being within the second preset speed statistical range, or the number of the first stationary vehicles being less than the second preset number, the first stationary vehicle is updated based on the at least one third vehicle that is stationary and diagonally parked relative to the first vehicle, determined based on the second perception fusion data, to obtain the second stationary vehicle.

7. The method for identifying a vehicle parked at an angle according to claim 5, wherein: The method further comprises: A preset buffer time is set for the moving vehicle, and the recognition of the moving vehicle is stopped within the preset buffer time.

8. The method for identifying an obliquely parked vehicle according to any one of claims 1 to 7, characterized in that: The method further comprises: In response to the number of the second stationary vehicles being less than a third preset number, the second stationary vehicles are deleted.

9. The method for identifying a vehicle parked at an angle according to claim 1, wherein: The method further comprises: In response to the number of the at least one second vehicle being less than or equal to a first preset number, or the vehicle speed statistical information corresponding to the at least one second vehicle being not within the first preset speed statistical range, determining that the first vehicle is in the first scene information; The first scene information represents that the first stationary vehicle does not correspond to the first perception fusion data.

10. A device for identifying a vehicle parked at an angle, comprising: an identification module, configured to obtain first perception fusion data collected by a first vehicle at a first moment, and determine, based on the first perception fusion data, at least one second vehicle that is parked obliquely relative to the first vehicle at the first moment; a determining module configured to, for each second vehicle, determine a first stationary vehicle from the at least one second vehicle based on a relationship between determined vehicle speed statistical information of the second vehicle and a first preset speed statistical range, wherein the vehicle speed statistical information is a lateral speed mean or a lateral speed variance; An acquisition module, configured to acquire second perception fusion data collected by the first vehicle at a second moment; the second moment is continuous with the first moment; An updating module is configured to update the first stationary vehicle based on the feature information of the first stationary vehicle in the second perception fusion data and at least one third vehicle that is parked obliquely relative to the first vehicle and is determined based on the second perception fusion data to obtain a second stationary vehicle.

11. A device for identifying a vehicle parked at an angle, comprising: processor, memory, and communication bus; The communication bus is used to realize the communication connection between the processor and the memory; The processor is configured to execute the computer program stored in the memory to implement the method for identifying an obliquely parked stationary vehicle according to any one of claims 1 to 9.

12. A computer-readable storage medium storing one or more computer programs, wherein the one or more computer programs can be executed by one or more processors to implement the method for identifying an obliquely parked stationary vehicle according to any one of claims 1 to 9.

Images