Obstacle detection method and device for complex parking scene, equipment and storage medium

Abstract

The invention discloses an obstacle detection method and device for a complex parking lot scene, equipment and a storage medium, and relates to the technical field of intelligent driving, and the method comprises the steps: obtaining target sensor data, recognizing an obstacle of the parking lot scene according to the target sensor data, and obtaining a first recognition result; obtaining a target visual image of the parking lot scene, and analyzing the surface reflection characteristics and shape characteristics of the parking lot scene according to the target visual image to identify and distinguish false inverted images and obstacles, thereby obtaining a second identification result; and determining a parking strategy of the parking lot scene according to the first identification result and the second identification result. According to the invention, the accuracy of parking lot scene obstacle detection can be improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent driving technology, and in particular to an obstacle detection method, device, equipment and storage medium for complex parking scenarios. Background Art

[0002] With the vigorous development of intelligent driving technology, valet parking technology represented by low-speed scenarios has also been rapidly improved, bringing great convenience to users' travel. However, in the actual use of the valet parking function, water accumulation in the parking lot or the epoxy floor with severe reflections often produce reflections, causing the system to misjudge it as an obstacle, thereby causing unnecessary deceleration, braking, or even obstacle avoidance operations, affecting the valet parking experience and success rate. At present, the detection of water accumulation in the selected parking space has been realized in the automatic parking scenario. However, relying solely on cameras to detect and perceive the parking environment, the error detection rate of obstacles is relatively high. Therefore, how to improve the accuracy of obstacle detection in parking scenes is a problem that still needs to be solved.

[0003] The above contents are only used to assist in understanding the technical solution of the present application and do not constitute an admission that the above contents are prior art. Summary of the invention

[0004] The main purpose of this application is to provide an obstacle detection method, device, equipment and storage medium for complex parking scenarios, aiming to solve the technical problem of how to improve the accuracy of obstacle detection in parking scenarios.

[0005] To achieve the above objectives, the present application proposes an obstacle detection method for complex parking scenarios, the method comprising:

[0006] Acquiring target sensor data, and identifying obstacles in a parking scene according to the target sensor data to obtain a first identification result;

[0007] Acquire a target visual image of the parking scene, and analyze the surface reflective features and shape features of the parking scene according to the target visual image to identify and distinguish false reflections from obstacles, thereby obtaining a second recognition result;

[0008] A parking strategy for the parking scene is determined according to the first recognition result and the second recognition result.

[0009] In one embodiment, the step of acquiring target sensor data, identifying obstacles in the parking scene according to the target sensor data, and obtaining a first recognition result includes:

[0010] Acquire initial sensor data of multiple sensors, including cameras, ultrasonic radars, millimeter-wave radars, and lidars;

[0011] Performing time synchronization and spatial calibration on the initial sensor data to obtain target sensor data;

[0012] Determining the priority of the target sensor data according to the parking scenario;

[0013] According to the priority, the target sensor data is fused to identify the reflective features and obstacles of the parking scene to obtain a first recognition result.

[0014] In one embodiment, the step of determining the priority of the target sensor data according to the parking scenario includes:

[0015] Prioritize target sensor data based on the brightness of the parking scene;

[0016] The priority of the target sensor data is determined according to the distance between the parking scene and the multiple sensors.

[0017] In one embodiment, the step of acquiring a target visual image of a parking scene, and analyzing the surface reflective features and shape features of the parking scene according to the target visual image to identify and distinguish false reflections from obstacles to obtain a second recognition result includes:

[0018] Acquire an initial visual image of a parking scene, and preprocess the initial visual image to obtain a target visual image;

[0019] Extracting reflection features and water accumulation features of the parking scene according to the target visual image and obtaining pending obstacles;

[0020] Evaluating the obstacle to be determined to obtain a confidence score of the obstacle to be determined;

[0021] The obstacle is determined according to the confidence score of the obstacle to be determined, and a second recognition result is obtained.

[0022] In one embodiment, the step of extracting the reflection features and water accumulation features of the parking scene according to the target visual image and obtaining the pending obstacles includes:

[0023] detecting and extracting reflection features in a parking scene according to the symmetry of the target visual image;

[0024] Extracting water features in the parking scene according to the brightness detection result and the reflection detection result of the target visual image;

[0025] The obstacle to be determined is determined according to the target visual image, the reflection feature and the water accumulation feature.

[0026] In one embodiment, before the step of determining a parking strategy for the parking scene according to the first recognition result and the second recognition result, the step further includes:

[0027] When a water accumulation area, a reflective area, and an obstacle are identified in the parking scene, the water accumulation area, the reflective area, and the obstacle are marked in a dynamic map of the parking scene to obtain a marked area;

[0028] Historical data is obtained, and a water accumulation area, a reflective area, and an obstacle in a parking scene are determined based on the historical data and the marked area.

[0029] In one embodiment, the step of determining a parking strategy for a parking scene according to the first recognition result and the second recognition result includes:

[0030] When the false reflection is detected on the parking path of the parking scene, no braking response is performed;

[0031] When it is detected that there is the obstacle on the parking path of the parking scene, parking is completed by detouring to avoid the obstacle.

[0032] In addition, to achieve the above purpose, the present application also proposes an obstacle detection device for complex parking scenarios, the obstacle detection device for complex parking scenarios comprising:

[0033] A first recognition module, used to obtain target sensor data, and recognize obstacles in the parking scene according to the target sensor data to obtain a first recognition result;

[0034] A second recognition module is used to obtain a target visual image of the parking scene, and analyze the surface reflective features and shape features of the parking scene according to the target visual image to identify and distinguish false reflections from obstacles to obtain a second recognition result;

[0035] A determination module is used to determine a parking strategy for the parking scene according to the first recognition result and the second recognition result.

[0036] In addition, to achieve the above-mentioned objectives, the present application also proposes an obstacle detection device for complex parking scenarios, the device comprising: a memory, a processor, and a computer program stored on the memory and executable on the processor, the computer program being configured to implement the steps of the obstacle detection method for complex parking scenarios as described above.

[0037] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by the processor, the steps of the obstacle detection method for complex parking scenarios as described above are implemented.

[0038] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the steps of the obstacle detection method for complex parking scenarios as described above.

[0039] The present application provides an obstacle detection method for complex parking scenarios. The present application obtains target sensor data, and identifies obstacles in the parking scenario based on the target sensor data to obtain a first recognition result; obtains a target visual image of the parking scenario, and analyzes the surface reflective features and shape features of the parking scenario based on the target visual image to identify and distinguish false reflections from obstacles to obtain a second recognition result; and determines a parking strategy for the parking scenario based on the first recognition result and the second recognition result.

[0040] In summary, the present application reduces false detection of obstacles in specific environments and improves the accuracy of obstacle detection in parking scenarios by fusing multiple sensor data and improving visual algorithms. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0042] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0043] Figure 1 A flowchart of the first embodiment of the obstacle detection method for complex parking scenarios provided in this application;

[0044] Figure 2 A flow chart of the second embodiment of the obstacle detection method for complex parking scenarios provided in this application;

[0045] Figure 3 A flowchart diagram of the third embodiment of the obstacle detection method for complex parking scenarios of the present application;

[0046] Figure 4 A schematic diagram of a simplified process of an obstacle detection method for a complex parking scenario provided by an embodiment of the present application;

[0047] Figure 5 This is a schematic diagram of the module structure of an obstacle detection device for complex parking scenarios in an embodiment of the present application;

[0048] Figure 6Schematic diagram of the device structure of the hardware operating environment involved in the obstacle detection method for complex parking scenarios in the embodiment of the present application.

[0049] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0050] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0051] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0052] The main solution of the present application is to obtain target sensor data, and identify obstacles in the parking scene based on the target sensor data to obtain a first recognition result; obtain a target visual image of the parking scene, and analyze the surface reflective features and shape features of the parking scene based on the target visual image to identify and distinguish false reflections from obstacles to obtain a second recognition result; determine the parking strategy of the parking scene based on the first recognition result and the second recognition result.

[0053] At present, the detection of water accumulation in the selected parking space has been realized in the automatic parking scene. However, relying solely on cameras to detect and perceive the parking environment has a high rate of false detection of obstacles. Therefore, how to improve the accuracy of obstacle detection in parking scenes is a problem that still needs to be solved.

[0054] This application reduces the false detection of obstacles in specific environments and improves the accuracy of obstacle detection in parking scenarios by fusing multiple sensor data and improving visual algorithms.

[0055] Based on this, the embodiment of the present application provides an obstacle detection method for a complex parking scene, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the obstacle detection method for complex parking scenarios of the present application.

[0056] In this embodiment, the obstacle detection method for complex parking scenarios includes steps S10 to S30:

[0057] Step S10: acquiring target sensor data, and identifying obstacles in the parking scene according to the target sensor data to obtain a first identification result;

[0058] It should be noted that the execution subject of this embodiment may be a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device capable of realizing the above functions, an obstacle detection device for complex parking scenes, etc. The following takes the obstacle detection device for complex parking scenes as an example to illustrate this embodiment and the following embodiments.

[0059] It should be noted that the sensors may include cameras, ultrasonic radars, millimeter-wave radars, and lidars. By acquiring and fusing data from different sensors, they can more accurately perceive the parking environment under various conditions, effectively reduce the probability of misjudgment of a single visual sensor in scenes such as reflection and water accumulation, and identify obstacles in parking scenes.

[0060] It is understandable that by fusing multiple sensors, the accuracy of vehicle system perception can be enhanced; the redundant perception information obtained by multiple sensors can be used to set the priority of different sensors in combination with parking scenarios; and by prioritizing redundant perception information, the "credibility" of obstacles can be evaluated to enhance reliability.

[0061] Step S20: acquiring a target visual image of the parking scene, and analyzing the surface reflective features and shape features of the parking scene according to the target visual image to identify and distinguish false reflections from obstacles, thereby obtaining a second recognition result;

[0062] It is understandable that the target visual image can improve the system's understanding and recognition capabilities of complex parking scene (such as reflection, water, cone barrel reflection, etc.), and more accurately distinguish between reflection, water and real obstacles, avoiding unnecessary response behaviors such as deceleration and braking caused by false detection. By improving the visual obstacle detection algorithm, adding the corresponding feature library, and combining the extraction features of reflection, the recognition accuracy of non-obstacles such as reflection and water can be improved.

[0063] Step S30: Determine a parking strategy for the parking scene according to the first recognition result and the second recognition result.

[0064] It is understandable that the first recognition result and the second recognition result are two ways of identifying obstacles in parking scenes. Combining these two recognition results can accurately identify false reflections and obstacles in parking scenes, thereby determining parking strategies.

[0065] In a feasible manner, before the step of determining the parking strategy of the parking scene according to the first recognition result and the second recognition result, the step further includes:

[0066] When a water accumulation area, a reflective area, and an obstacle are identified in the parking scene, the water accumulation area, the reflective area, and the obstacle are marked in a dynamic map of the parking scene to obtain a marked area;

[0067] Historical data is obtained, and a water accumulation area, a reflective area, and an obstacle in a parking scene are determined based on the historical data and the marked area.

[0068] It is understandable that when an area of ​​the parking scene is first identified as a waterlogged or reflective area, the system will temporarily mark it and speed up the detection frequency in a short period of time to confirm whether it is indeed a stable feature. If an area is continuously identified as a fixed feature in multiple set detections, the area will be marked as a "high weight" area and added to the dynamic map as a long-term obstacle. If the marked area is no longer detected as a waterlogged or reflective area for a period of time, the weight of the marked area will be reduced, and the mark may eventually be cancelled to ensure the real-time nature of the map. At the same time, the changing trend of dynamic features is recorded according to historical environmental data, such as certain weather conditions or light intensity, specific areas are prone to waterlogging or reflection. Through this data accumulation, the system marks and updates the dynamic map in advance under similar conditions to improve the system response efficiency. The parking environment map is updated in real time through real-time positioning and map construction technology, marking highly reflective ground, waterlogged points, etc., and generating a specific obstacle map of the parking area, so that the system can accumulate more information from multiple parking records in the same parking lot, and the system can give priority to historical data and preset information when detecting similar environments.

[0069] It is understandable that when water accumulation or reflective areas are detected, the updated water accumulation and reflective area annotations can be synchronized to the main map. If the same area is passed multiple times and the environment is stable, the marking information of the area can be uploaded to the cloud for reference by other vehicles of the same model to further improve the dynamic map of the parking lot. In addition, marked areas that no longer appear can be cleaned up regularly to avoid redundant information in the dynamic map affecting system performance. Set a timestamp, and when a marked area has not been detected for a long time, its weight will be automatically reduced and eventually removed, thereby ensuring the real-time and accuracy of map information.

[0070] In a feasible manner, the step of determining a parking strategy for a parking scene according to the first recognition result and the second recognition result includes:

[0071] When the false reflection is detected on the parking path of the parking scene, no braking response is performed;

[0072] When it is detected that there is the obstacle on the parking path of the parking scene, parking is completed by detouring to avoid the obstacle.

[0073] It is understandable that for objects that are judged to be reflections or water accumulation, the braking response can be ignored or reduced. A probability-based judgment mechanism is used to determine whether it is necessary to stop or fine-tune the path; if the identified obstacle is on the parking path but the confidence level is low, dynamic path adjustment (such as detour) can be used to avoid potential interference without blindly slowing down or stopping.

[0074] This embodiment obtains target sensor data, and identifies obstacles in the parking scene based on the target sensor data to obtain a first recognition result; obtains a target visual image of the parking scene, and analyzes the surface reflective features and shape features of the parking scene based on the target visual image to identify and distinguish false reflections from obstacles to obtain a second recognition result; and determines a parking strategy for the parking scene based on the first recognition result and the second recognition result.

[0075] In summary, this embodiment reduces the false detection of obstacles in specific environments and improves the accuracy of obstacle detection in parking scenarios by fusing multiple sensor data and improving visual algorithms.

[0076] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can be referred to the above introduction, and will not be repeated in the following. Figure 2 , step S10 includes steps S101 to S104:

[0077] Step S101: Acquire initial sensor data of multiple sensors, where the multiple sensors include cameras, ultrasonic radars, millimeter-wave radars, and laser radars;

[0078] It should be noted that the redundant perception information that can be obtained through multiple sensor data increases the accuracy of obstacle judgment.

[0079] Step S102: performing time synchronization and space calibration on the initial sensor data to obtain target sensor data;

[0080] It is understandable that after obtaining multiple sensor data, time synchronization and space calibration are required. Time synchronization can be specifically achieved by synchronizing multiple sensor data through timestamps, uniformly processing them in the central control unit, and calibrating the sampling timestamps of the sensors using a high-precision synchronous clock; and using filters to smooth the detection data of different sensors to reduce the errors caused by time differences. Spatial calibration can be achieved by converting multiple sensor data into the same coordinate system, such as the vehicle's body coordinate system, and fusing the data of cameras and millimeter-wave radars, ultrasonic radars, and lidars, so that obstacle information from different perspectives can be superimposed and analyzed.

[0081] Step S103: determining the priority of the target sensor data according to the parking scenario;

[0082] It is understandable that when identifying suspected obstacles, the priority of the sensor can be adjusted dynamically. The priority setting can be determined based on the reliability of the sensor in the current parking scenario. For example, when there is sufficient light, the camera data is prioritized, while in reflective or water-logged scenarios, the millimeter-wave radar and lidar data are prioritized. Determining different priorities for sensors in different parking scenarios can improve the accuracy of obstacle detection in parking scenarios.

[0083] In a feasible manner, the step of determining the priority of the target sensor data according to the parking scenario includes:

[0084] Prioritize target sensor data based on the brightness of the parking scene;

[0085] The priority of the target sensor data is determined according to the distance between the parking scene and the multiple sensors.

[0086] It is understandable that different weights are assigned to data from different sensors, and the weights are adjusted dynamically according to the scene. For example, in low-light scenes, the weight of radar perception data should be increased accordingly, and the weight of visual data perceived by the camera should be relatively reduced. Layered processing is performed according to the effective distance detected by the sensor and the reliability of the depth information. For example, lidar data is used at long distances, camera and millimeter-wave radar data are integrated at medium distances, and ultrasonic data is used first at close distances.

[0087] Step S104: According to the priority, the target sensor data is fused to identify the reflective features and obstacles of the parking scene to obtain a first recognition result.

[0088] It is understandable that, according to the priority, the target sensor data is fused to obtain redundant information of the parking scene and improve the accuracy of obstacle detection. By configuring a flexible redundancy mechanism, when the object attributes (such as distance, material) detected by a sensor do not match the detection results of other sensors, the system can re-evaluate to enhance the reliability of the detection results; when a sensor is detected to be faulty or abnormal, the redundancy mechanism can be enabled, such as when the camera is damaged, the millimeter-wave radar and ultrasonic radar take over; for reflections or reflections, the consistency of the data can be checked by cross-comparison of sensors. For example, if the camera detects an obstacle, but the millimeter-wave radar and ultrasonic radar do not detect it, the obstacle is marked as low credibility. In addition, image analysis can be used to identify areas such as water surface reflections and glass reflections in parking scenes using color and texture, and image processing algorithms (such as edge detection) can be used to extract the features of reflective objects. According to the sensor data, reflective scenes, water accumulation scenes and ordinary scenes are identified, and the fusion strategies under different environments are optimized respectively, so that different scenes such as water accumulation and reflections can be accurately identified during parking.

[0089] This embodiment obtains initial sensor data of multiple sensors, including cameras, ultrasonic radars, millimeter-wave radars, and lidars; performs time synchronization and spatial calibration on the initial sensor data to obtain target sensor data; determines the priority of the target sensor data according to the parking scene; and fuses the target sensor data according to the priority to identify reflective features and obstacles of the parking scene to obtain a first recognition result.

[0090] In summary, this embodiment accurately perceives parking scenes by fusing multiple types of sensor data, effectively reduces the probability of misjudgment of a single visual sensor in scenes such as reflections and water accumulation, and improves the accuracy of obstacle detection in parking scenes.

[0091] Based on the first embodiment of the present application, in the third embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can be referred to the above introduction, and will not be repeated in the following. Figure 3 , step S20 includes steps S201 to S204:

[0092] Step S201: acquiring an initial visual image of a parking scene, and preprocessing the initial visual image to obtain a target visual image;

[0093] It should be noted that the preprocessing method of the initial visual image includes using Canny edge detection and Sobel algorithm to extract significant edges in the image, removing weak interference edges to reduce the impact of reflections and water reflections; applying median filtering or Gaussian filtering to remove noise in the image to make the edges clearer and continuous. Using adaptive histogram equalization to adjust the brightness of the initial visual image, especially in strong light or shadow areas, can enhance dark details and make the features of water and reflections easier to identify; using correction to optimize the contrast of the initial visual image, so that the system can maintain good recognition accuracy under strong light and low light conditions.

[0094] Step S202: extracting reflection features and water accumulation features of the parking scene according to the target visual image and obtaining obstacles to be determined;

[0095] It should be noted that, based on the target visual image, features of the parking scene can be extracted to obtain pending obstacles.

[0096] In a feasible manner, the step of extracting the reflection features and water accumulation features of the parking scene according to the target visual image and obtaining the pending obstacles includes:

[0097] detecting and extracting reflection features in a parking scene according to the symmetry of the target visual image;

[0098] Extracting water features in the parking scene according to the brightness detection result and the reflection detection result of the target visual image;

[0099] The obstacle to be determined is determined according to the target visual image, the reflection feature and the water accumulation feature.

[0100] It is understandable that feature extraction is performed by combining visual geometric features and extracting the difference information between accumulated water, reflections and real objects. For example, reflections generally appear as low-transparency, top-to-bottom symmetrical structures. Image symmetry detection and depth estimation can be used to determine whether they are reflections. Feature extraction algorithms based on symmetry detection can be used, such as calculating the axis of symmetry in the image, mirror pixel similarity, etc. Accumulated water often presents the characteristics of local highlights and smooth surfaces, which can be determined by brightness threshold detection and reflective feature analysis. In addition, the texture of the accumulated water area is usually less, which can be further confirmed by texture feature extraction. By using a binocular camera or a structured light camera to estimate the depth in the scene, the depth information is combined to determine the difference between distant objects (such as ground reflections) and closer real obstacles (such as vehicles or pillars), so as to avoid reflections being mistaken for real obstacles.

[0101] Step S203: Evaluate the obstacle to be determined to obtain a confidence score of the obstacle to be determined;

[0102] It is understandable that for the detected pending obstacles, the confidence score of the pending obstacles is calculated. Low-confidence objects may be reflections or water accumulation areas, which can be further judged whether they are real obstacles; a risk judgment model based on Bayesian estimation is established. For water accumulation or reflections that appear repeatedly in a specific area, the model determines its low risk through historical data and makes dynamic weight adjustments. In addition, with the accumulation of data on reflective scenes in parking lots, an incremental learning method is used to allow the algorithm to continuously adapt to new scenes in actual use and optimize the recognition effect of reflective or water accumulation scenes; automated annotation tools are used to collect samples that are misdetected by the system during parking (such as the misjudgment of water accumulation reflections), and the algorithm is continuously trained and improved in an offline environment.

[0103] Step S204: determining the obstacle according to the confidence score of the obstacle to be determined, and obtaining a second recognition result.

[0104] It is understandable that when the confidence score of the pending obstacle reaches a certain value, the pending obstacle can be considered to be an obstacle, otherwise the pending obstacle can be determined to be a reflection, a reflective object, or an object of accumulated water.

[0105] This embodiment obtains an initial visual image of a parking scene, and preprocesses the initial visual image to obtain a target visual image; extracts reflection features and water accumulation features of the parking scene according to the target visual image and obtains a pending obstacle; evaluates the pending obstacle to obtain a confidence score of the pending obstacle; determines the obstacle according to the confidence score of the pending obstacle to obtain a second recognition result.

[0106] To summarize, this embodiment improves the automatic parking system's ability to understand and recognize complex parking scenes in parking lots (such as reflections, stagnant water, reflections of cones, etc.) by optimizing the visual perception algorithm. By improving the visual algorithm, it can more accurately distinguish between reflections, stagnant water and real obstacles, thereby improving the accuracy of obstacle detection in parking scenes.

[0107] For example, in order to help understand the implementation process of the obstacle detection method for complex parking scenes obtained by combining this embodiment with the above-mentioned embodiment 1, please refer to Figure 4 , Figure 4A brief flow chart of an obstacle detection method for a complex parking scene is provided. Specifically, there are three modules for water accumulation and reflective ground detection, module 1: sensor perception and redundancy module, module 2: visual perception algorithm optimization module, and module 3: parking lot map real-time update module. Among them, the steps of module 1 are sensor data synchronization, data fusion, sensor priority determination, sensor redundancy perception, environmental feature extraction and classification. The steps of module 2 are image preprocessing, feature extraction and depth estimation, risk assessment, decision mechanism, and algorithm update. The steps of module 3 are map layer marking, weight adjustment, environmental historical data backtracking, real-time map update and synchronization, and regular map cleaning and reset. Water accumulation and reflective ground are identified by modules 1 and 2 to reduce false detection, and the detection results are provided to module 3 as historical data for use. Module 3 assists modules 1 and 2 in detection. Therefore, the sensor fusion and redundant perception module and the visual perception algorithm optimization module are used to realize the recognition of accumulated water and reflective ground, reduce false detection, and the parking lot map real-time update module collects and integrates the map information of accumulated water and reflective ground, dynamically updates the map information of the parking area, reduces false detection in specific environments, and improves the accuracy of obstacle detection in parking scenes.

[0108] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the obstacle detection method for complex parking scenarios of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0109] This application also provides an obstacle detection device for complex parking scenarios, please refer to Figure 5 , the obstacle detection device for complex parking scenes includes:

[0110] A first recognition module 10 is used to obtain target sensor data, and recognize obstacles in the parking scene according to the target sensor data to obtain a first recognition result;

[0111] The second recognition module 20 is used to obtain a target visual image of the parking scene, and analyze the surface reflective features and shape features of the parking scene according to the target visual image to identify and distinguish false reflections from obstacles to obtain a second recognition result;

[0112] The determination module 30 is used to determine a parking strategy for the parking scene according to the first recognition result and the second recognition result.

[0113] This embodiment obtains target sensor data, and identifies obstacles in the parking scene based on the target sensor data to obtain a first recognition result; obtains a target visual image of the parking scene, and analyzes the surface reflective features and shape features of the parking scene based on the target visual image to identify and distinguish false reflections from obstacles to obtain a second recognition result; and determines a parking strategy for the parking scene based on the first recognition result and the second recognition result.

[0114] In summary, this embodiment reduces the false detection of obstacles in specific environments and improves the accuracy of obstacle detection in parking scenarios by fusing multiple sensor data and improving visual algorithms.

[0115] In one embodiment, the first recognition module 10 is also used to obtain initial sensor data of multiple sensors, including cameras, ultrasonic radars, millimeter-wave radars, and lidars; perform time synchronization and spatial calibration on the initial sensor data to obtain target sensor data; determine the priority of the target sensor data according to the parking scene; and fuse the target sensor data according to the priority to identify reflective features and obstacles of the parking scene to obtain a first recognition result.

[0116] In one embodiment, the first recognition module 10 is further configured to determine the priority of the target sensor data according to the brightness of the parking scene; and determine the priority of the target sensor data according to the distance between the parking scene and the multiple sensors.

[0117] In one embodiment, the second recognition module 20 is further used to obtain an initial visual image of the parking scene, and preprocess the initial visual image to obtain a target visual image; extract reflection features and water accumulation features of the parking scene according to the target visual image and obtain a pending obstacle; evaluate the pending obstacle to obtain a confidence score of the pending obstacle; determine the obstacle according to the confidence score of the pending obstacle to obtain a second recognition result.

[0118] In one embodiment, the second recognition module 20 is further used to detect and extract reflection features in the parking scene according to the symmetry of the target visual image; extract water accumulation features in the parking scene according to the brightness detection results and reflection detection results of the target visual image; and determine the pending obstacles according to the target visual image, the reflection features and the water accumulation features.

[0119] In one embodiment, the determination module 30 is also used to mark the water accumulation area, the reflective area and the obstacle in the dynamic map of the parking scene to obtain a marked area when the water accumulation area, the reflective area and the obstacle are identified in the parking scene; obtain historical data, and judge the water accumulation area, the reflective area and the obstacle of the parking scene based on the historical data and the marked area.

[0120] In one embodiment, the determination module 30 is further configured to not perform a braking response when the false reflection is detected on the parking path of the parking scene;

[0121] When it is detected that there is the obstacle on the parking path of the parking scene, parking is completed by detouring to avoid the obstacle.

[0122] The obstacle detection device for complex parking scenes provided by the present application adopts the obstacle detection method for complex parking scenes in the above-mentioned embodiment, which can solve the technical problem of how to improve the accuracy of obstacle detection in parking scenes. Compared with the prior art, the beneficial effects of the obstacle detection device for complex parking scenes provided by the present application are the same as the beneficial effects of the obstacle detection method for complex parking scenes provided by the above-mentioned embodiment, and the other technical features of the obstacle detection device for complex parking scenes are the same as the features disclosed in the above-mentioned embodiment method, which will not be repeated here.

[0123] The present application provides an obstacle detection device for complex parking scenarios, and the obstacle detection device for complex parking scenarios includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the obstacle detection method for complex parking scenarios in the above-mentioned embodiment one.

[0124] Reference below Figure 6 , which shows a schematic diagram of the structure of an obstacle detection device suitable for implementing a complex parking scene in an embodiment of the present application. The obstacle detection device for a complex parking scene in an embodiment of the present application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 6The obstacle detection device for the complex parking scenario shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0125] like Figure 6 As shown, the obstacle detection device for complex parking scenes may include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the obstacle detection device for complex parking scenes are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems may be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 1009. The communication device 1009 may allow the obstacle detection device for complex parking scenarios to communicate wirelessly or wired with other devices to exchange data. Although the figure shows an obstacle detection device for complex parking scenarios with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or have alternatively.

[0126] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0127] The obstacle detection device for complex parking scenes provided by the present application adopts the obstacle detection method for complex parking scenes in the above-mentioned embodiment, which can solve the technical problem of how to improve the accuracy of obstacle detection in parking scenes. Compared with the prior art, the beneficial effects of the obstacle detection device for complex parking scenes provided by the present application are the same as the beneficial effects of the obstacle detection method for complex parking scenes provided by the above-mentioned embodiment, and the other technical features of the obstacle detection device for complex parking scenes are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0128] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0129] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0130] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the obstacle detection method for complex parking scenarios in the above-mentioned embodiment.

[0131] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0132] The computer-readable storage medium may be included in the obstacle detection device for complex parking scenarios; or may exist independently without being assembled into the obstacle detection device for complex parking scenarios.

[0133] The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by the obstacle detection device for complex parking scenes, the obstacle detection device for complex parking scenes: obtains target sensor data, and identifies obstacles in the parking scene based on the target sensor data to obtain a first recognition result; obtains a target visual image of the parking scene, and analyzes the surface reflective features and shape features of the parking scene based on the target visual image to identify and distinguish false reflections from obstacles to obtain a second recognition result; determines the parking strategy for the parking scene based on the first recognition result and the second recognition result.

[0134] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0135] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0136] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0137] The readable storage medium provided in the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned obstacle detection method for complex parking scenes, and can solve the technical problem of how to improve the accuracy of obstacle detection in parking scenes. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in the present application are the same as the beneficial effects of the obstacle detection method for complex parking scenes provided in the above-mentioned embodiments, and will not be elaborated here.

[0138] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the obstacle detection method for complex parking scenarios as described above.

[0139] The computer program product provided by the present application can solve the technical problem of how to improve the accuracy of obstacle detection in parking scenes. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as the beneficial effects of the obstacle detection method for complex parking scenes provided by the above embodiment, which will not be elaborated here.

[0140] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. An obstacle detection method for complex parking scenes, characterized in that: The method includes: Acquiring target sensor data, and identifying obstacles in a parking scene according to the target sensor data to obtain a first identification result; Acquire a target visual image of the parking scene, and analyze the surface reflective features and shape features of the parking scene according to the target visual image to identify and distinguish false reflections from obstacles, thereby obtaining a second recognition result; A parking strategy for the parking scene is determined according to the first recognition result and the second recognition result.

2. The method according to claim 1, characterized in that The step of acquiring target sensor data, and identifying obstacles in the parking scene according to the target sensor data to obtain a first recognition result includes: Acquire initial sensor data of multiple sensors, including cameras, ultrasonic radars, millimeter-wave radars, and lidars; Performing time synchronization and spatial calibration on the initial sensor data to obtain target sensor data; Determining the priority of the target sensor data according to the parking scenario; According to the priority, the target sensor data is fused to identify the reflective features and obstacles of the parking scene to obtain a first recognition result.

3. The method according to claim 2, characterized in that The step of determining the priority of the target sensor data according to the parking scenario comprises: Prioritize target sensor data based on the brightness of the parking scene; The priority of the target sensor data is determined according to the distance between the parking scene and the multiple sensors.

4. The method according to claim 1, characterized in that The step of acquiring a target visual image of the parking scene, and analyzing the surface reflective features and shape features of the parking scene according to the target visual image to identify and distinguish false reflections from obstacles to obtain a second recognition result comprises: Acquire an initial visual image of a parking scene, and preprocess the initial visual image to obtain a target visual image; Extracting reflection features and water accumulation features of the parking scene according to the target visual image and obtaining pending obstacles; Evaluating the obstacle to be determined to obtain a confidence score of the obstacle to be determined; The obstacle is determined according to the confidence score of the obstacle to be determined, and a second recognition result is obtained.

5. The method according to claim 4, characterized in that The step of extracting the reflection features and the water accumulation features of the parking scene according to the target visual image and obtaining the pending obstacles comprises: detecting and extracting reflection features in a parking scene according to the symmetry of the target visual image; Extracting water features in the parking scene according to the brightness detection result and the reflection detection result of the target visual image; The obstacle to be determined is determined according to the target visual image, the reflection feature and the water accumulation feature.

6. The method according to claim 1, characterized in that Before the step of determining a parking strategy for the parking scene according to the first recognition result and the second recognition result, the method further includes: When a water accumulation area, a reflective area, and an obstacle are identified in the parking scene, the water accumulation area, the reflective area, and the obstacle are marked in a dynamic map of the parking scene to obtain a marked area; Historical data is obtained, and a water accumulation area, a reflective area, and an obstacle in a parking scene are determined based on the historical data and the marked area.

7. The method according to claim 1, characterized in that The step of determining a parking strategy for a parking scene according to the first recognition result and the second recognition result comprises: When the false reflection is detected on the parking path of the parking scene, no braking response is performed; When it is detected that there is the obstacle on the parking path of the parking scene, the parking is completed by detouring to avoid the obstacle.

8. An obstacle detection device for complex parking scenes, characterized in that: The device comprises: A first recognition module, used to obtain target sensor data, and recognize obstacles in the parking scene according to the target sensor data to obtain a first recognition result; A second recognition module is used to obtain a target visual image of the parking scene, and analyze the surface reflective features and shape features of the parking scene according to the target visual image to identify and distinguish false reflections from obstacles to obtain a second recognition result; A determination module is used to determine a parking strategy for the parking scene according to the first recognition result and the second recognition result.

9. An obstacle detection device for complex parking scenarios, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the obstacle detection method for a complex parking scene as claimed in any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the obstacle detection method for a complex parking scene according to any one of claims 1 to 7 is implemented.

Images