Parking space recommendation method and device, electronic equipment and computer readable storage medium

By collecting environmental images through vehicle cameras, identifying obstacles and parking space information, and combining scoring processing and score fusion models, the system recommends parking spaces with lower difficulty levels, solving the problem of high parking difficulty in existing parking systems and improving the parking success rate.

CN117911980BActive Publication Date: 2026-11-10GUANGZHOU AUTOMOBILE GROUP CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202311829698.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-12-27
Publication Date
2026-11-10
Estimated Expiration
2043-12-27

AI Technical Summary

Technical Problem

Existing automatic parking systems face significant challenges in selecting optimal parking spaces, resulting in a low parking success rate.

Method used

By collecting environmental images through vehicle cameras, identifying obstacles and parking space information, and combining the parking space information for scoring, a fusion score is calculated using a score fusion model to recommend parking spaces with lower difficulty.

Benefits of technology

It improves the parking success rate, reduces the difficulty of parking, and enhances the intelligence and accuracy of the parking system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117911980B_ABST
    Figure CN117911980B_ABST
Patent Text Reader

Abstract

The application provides a parking space recommendation method and device, electronic equipment, computer readable storage medium and computer program product. The method comprises: collecting an environment image through a camera of a vehicle; performing obstacle identification on the environment image to obtain obstacle information of multiple parking spaces around the vehicle; wherein the obstacle information comprises an obstacle category and an obstacle direction; determining parking space information of the multiple parking spaces; wherein the parking space information comprises a parking space category and a parking space size; performing scoring processing on the parking space information and the obstacle information of each parking space to obtain multiple attribute scores of each parking space; performing score fusion processing on the multiple attribute scores of each parking space according to a score fusion model to obtain a fusion score of each parking space; and recommending a parking space according to the fusion score of each parking space. Through the application, the parking space can be recommended comprehensively according to the parking space information and the obstacle information, so that the recommended parking space has a lower parking difficulty and the parking success rate is effectively improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to vehicle technology, and more particularly to a parking space recommendation method, apparatus, electronic device, computer-readable storage medium, and computer program product. Background Technology

[0002] Parking refers to placing a vehicle in a parking space. With the development of vehicle technology and computer technology, automatic parking systems have emerged. Automatic parking systems are used to select the best parking space from multiple parking spaces around the vehicle and guide the user to park the vehicle in the best parking space.

[0003] The solutions provided by related technologies typically combine the vehicle's location, the distance between the vehicle and the parking lot, and the user's personalized needs (such as charging, shopping, etc.) to select an available parking space as the optimal parking space. However, the optimal parking space determined by this method may be difficult to park in, resulting in a low parking success rate. Summary of the Invention

[0004] This application provides a parking space recommendation method, device, electronic device, computer-readable storage medium, and computer program product, which can recommend parking spaces by comprehensively considering parking space information and obstacle information, making the recommended parking spaces easier to park and effectively improving the parking success rate.

[0005] The technical solution of this application is implemented as follows:

[0006] This application provides a parking space recommendation method, including:

[0007] The vehicle's cameras capture images of the environment.

[0008] Obstacle recognition is performed on the environmental image to obtain obstacle information for multiple parking spaces around the vehicle; wherein, the obstacle information includes obstacle type and obstacle location;

[0009] Determine the parking information of the plurality of parking spaces; wherein, the parking information includes the parking space category and the parking space size;

[0010] The scoring process is performed based on the parking space information and obstacle information of each parking space to obtain multiple attribute scores for each parking space;

[0011] The scores of multiple attributes for each parking space are fused according to the score fusion model to obtain the fused score for each parking space.

[0012] Parking spaces are recommended based on their integration score.

[0013] This application provides a parking space recommendation device, comprising:

[0014] The data acquisition module is used to acquire environmental images through the vehicle's cameras;

[0015] An obstacle recognition module is used to identify obstacles in the environmental image to obtain obstacle information for multiple parking spaces around the vehicle; wherein, the obstacle information includes obstacle category and obstacle location;

[0016] A parking space recognition module is used to determine the parking space information of the plurality of parking spaces; wherein, the parking space information includes the parking space category and the parking space size;

[0017] The scoring module is used to score each parking space based on its parking space information and obstacle information, and to obtain multiple attribute scores for each parking space.

[0018] The score fusion module is used to perform score fusion processing on the multiple attribute scores of each parking space according to the score fusion model to obtain the fused score of each parking space.

[0019] The recommendation module is used to recommend parking spaces based on the fusion score of each parking space.

[0020] This application provides an electronic device, including:

[0021] Memory, used to store executable instructions;

[0022] The processor, when executing executable instructions stored in the memory, implements the parking space recommendation method provided in this application.

[0023] This application provides a computer-readable storage medium storing executable instructions for implementing the parking space recommendation method provided in this application when executed by a processor.

[0024] This application provides a computer program product including executable instructions for implementing the parking space recommendation method provided in this application when executed by a processor.

[0025] This application has the following beneficial effects:

[0026] This application acquires environmental images using a vehicle's camera; it then performs obstacle recognition on these images to obtain obstacle information for multiple parking spaces around the vehicle, including obstacle type and location; simultaneously, it determines parking space information for multiple parking spaces, including parking space type and size. Based on the obtained parking space and obstacle information, a scoring process is performed on each parking space to obtain multiple attribute scores. These scores are then fused using a score fusion model to obtain a fused score for each parking space, accurately measuring its parking difficulty. Finally, parking spaces are recommended based on their fused scores, resulting in lower parking difficulty and effectively improving the parking success rate. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 This is a schematic diagram of the architecture of the parking space recommendation system provided in the embodiments of this application;

[0029] Figure 2 This is a schematic diagram of the structure of a terminal device provided in an embodiment of this application;

[0030] Figure 3A This is a flowchart illustrating a parking space recommendation method provided in an embodiment of this application;

[0031] Figure 3B This is another flowchart illustrating the parking space recommendation method provided in this application embodiment;

[0032] Figure 3C This is another flowchart illustrating the parking space recommendation method provided in this application embodiment;

[0033] Figure 4 This is a schematic diagram of the parking space categories provided in an embodiment of this application;

[0034] Figure 5 This is a schematic diagram of an obstacle scenario provided in an embodiment of this application;

[0035] Figure 6 This is a schematic diagram of the parking space dimensions and vehicle dimensions provided in the embodiments of this application. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0037] In the following description, references to "some embodiments" describe a subset of all possible embodiments; however, it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict. In the following description, the term "a plurality of" means at least two.

[0038] In the following description, the terms "first, second, third" are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first, second, third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0039] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0040] This application provides a parking space recommendation method, apparatus, electronic device, computer-readable storage medium, and computer program product. These methods integrate parking space information and obstacle information to recommend parking spaces, making it easier to park in the recommended spaces and effectively improving the parking success rate. The following describes exemplary applications of the electronic device provided in this application. This electronic device can be implemented as various types of terminal devices or as a server.

[0041] See Figure 1 , Figure 1 This is a schematic diagram of the architecture of the parking space recommendation system 100 provided in the embodiments of this application. The terminal device 400 is connected to the server 200 through the network 300, wherein the network 300 can be a wide area network or a local area network, or a combination of the two.

[0042] In some embodiments, taking the electronic device as an example of a terminal device, the parking space recommendation method provided in this application can be implemented by a terminal device. For example... Figure 1As shown, terminal device 400 can be a terminal device deployed in a vehicle. Terminal device 400 acquires environmental images through the vehicle's camera; performs obstacle recognition on the environmental images to obtain obstacle information for multiple parking spaces around the vehicle; wherein, the obstacle information includes obstacle category and obstacle location; determines parking space information for multiple parking spaces; wherein, the parking space information includes parking space category and parking space size; performs scoring processing based on the parking space information and obstacle information for each parking space to obtain multiple attribute scores for each parking space; performs score fusion processing on the multiple attribute scores for each parking space according to a score fusion model to obtain a fused score for each parking space; and recommends parking spaces based on the fused scores for each parking space. The method of recommending parking spaces is not limited; for example, the location of the recommended parking space can be displayed on the in-vehicle display screen so that the user (car owner) can drive into the parking space according to the location displayed on the in-vehicle display screen to achieve parking.

[0043] In some embodiments, the parking space recommendation method provided in this application can also be implemented collaboratively by a terminal device and a server. For example, after the terminal device 400 collects environmental images through the vehicle's camera, it sends the environmental images to the server 200; the server 200 performs obstacle recognition on the environmental images to obtain obstacle information for multiple parking spaces around the vehicle; wherein, the obstacle information includes obstacle category and obstacle location; the server 200 determines parking space information for multiple parking spaces; wherein, the parking space information includes parking space category and parking space size; the server 200 performs scoring processing based on the parking space information and obstacle information for each parking space to obtain multiple attribute scores for each parking space; the server 200 performs score fusion processing on the multiple attribute scores for each parking space according to a score fusion model to obtain a fusion score for each parking space; the server 200 recommends parking spaces based on the fusion score of each parking space, for example, it can notify the terminal device 400 based on the recommended parking space so that the terminal device 400 displays the location of the recommended parking space on the vehicle display screen.

[0044] In some embodiments, the terminal device 400 or server 200 can implement the parking space recommendation method provided in this application embodiment by running a computer program. For example, the computer program can be a native program or software module in an operating system; it can be a native application (APP), that is, a program that needs to be installed in the operating system to run; it can also be a mini-program, that is, a program that only needs to be downloaded to a browser environment to run; or it can be a mini-program that can be embedded in any APP. In short, the above-mentioned computer program can be any form of application, module or plugin.

[0045] In some embodiments, server 200 may be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms. Terminal device 400 may be an in-vehicle device, smartphone, tablet computer, laptop computer, desktop computer, smart TV, smartwatch, etc., but is not limited to these. Terminal devices and servers can be directly or indirectly connected via wired or wireless communication, which is not limited in this embodiment.

[0046] Taking the electronic device provided in this application as an example, which is a terminal device, it can be understood that in the case where the electronic device is a server, Figure 2 Some parts of the structure shown (such as the user interface, presentation module, and input processing module) can be omitted. See also Figure 2 , Figure 2 This is a schematic diagram of the structure of the terminal device 400 provided in the embodiments of this application. Figure 2 The terminal device 400 shown includes at least one processor 410, a memory 450, at least one network interface 420, and a user interface 430. The various components in the terminal device 400 are coupled together via a bus system 440. It is understood that the bus system 440 is used to implement communication between these components. In addition to a data bus, the bus system 440 also includes a power bus, a control bus, and a status signal bus. However, for clarity, ... Figure 2 The general labeled all buses as Bus System 440.

[0047] The processor 410 can be an integrated circuit chip with signal processing capabilities, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc.

[0048] User interface 430 includes one or more output devices 431 that enable the presentation of media content, including one or more speakers and / or one or more visual displays. User interface 430 also includes one or more input devices 432, including user interface components that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls.

[0049] The memory 450 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state storage, hard disk drives, optical disk drives, etc. The memory 450 may optionally include one or more storage devices physically located away from the processor 410.

[0050] The memory 450 may include volatile memory or non-volatile memory, or both. The non-volatile memory may be read-only memory (ROM), and the volatile memory may be random access memory (RAM). The memory 450 described in this application embodiment is intended to include any suitable type of memory.

[0051] In some embodiments, memory 450 is capable of storing data to support various operations, examples of which include programs, modules, and data structures or subsets or supersets thereof, as illustrated below.

[0052] Operating system 451 includes system programs for handling various basic system services and performing hardware-related tasks, such as the framework layer, core library layer, driver layer, etc., for implementing various basic business functions and handling hardware-based tasks;

[0053] The network communication module 452 is used to reach other computing devices via one or more (wired or wireless) network interfaces 420, exemplary network interfaces 420 including: Bluetooth, WiFi, and Universal Serial Bus (USB), etc.

[0054] Presentation module 453 is configured to enable the presentation of information (e.g., a user interface for operating peripheral devices and displaying content and information) via one or more output devices 431 (e.g., a display screen, a speaker, etc.) associated with user interface 430;

[0055] The input processing module 454 is used to detect and translate one or more user inputs or interactions from one or more input devices 432.

[0056] In some embodiments, the parking space recommendation device provided in this application can be implemented in software. Figure 2A parking space recommendation device 455 stored in memory 450 is shown. This device can be software in the form of programs and plug-ins, and includes the following software modules: a data acquisition module 4551, an obstacle recognition module 4552, a parking space recognition module 4553, a scoring module 4554, a score fusion module 4555, and a recommendation module 4556. These modules are logically connected and can therefore be arbitrarily combined or further separated according to their implemented functions. The functions of each module will be described below.

[0057] The parking space recommendation method provided in this application will be described in conjunction with exemplary applications and implementations of the electronic devices provided in the embodiments of this application.

[0058] See Figure 3A , Figure 3A This is a flowchart illustrating a parking space recommendation method provided in an embodiment of this application, which will be combined with... Figure 3A The steps shown are explained.

[0059] In step 101, environmental images are captured using the vehicle's camera.

[0060] Here, images of the vehicle's surrounding environment are captured by cameras deployed (installed) on the vehicle. For ease of distinction, the images captured by the cameras are named environmental images.

[0061] This application does not limit the type of camera. For example, environmental images can be collected using a fisheye camera on a vehicle. Since the fisheye camera has a large field of view, the amount of information in the collected environmental images is also large, and more effective information can be extracted.

[0062] In step 102, obstacle recognition is performed on the environmental image to obtain obstacle information for multiple parking spaces around the vehicle; the obstacle information includes obstacle type and obstacle location.

[0063] Here, obstacle recognition is performed on the environmental image to obtain obstacle information for multiple parking spaces around the vehicle. The obstacle information includes the obstacle type and the obstacle location. The obstacle type is not limited to pillars, walls, pedestrians, vehicles, cones, water-filled barriers, parking lines, etc. The obstacle location refers to the position of the obstacle relative to the parking space, such as being located to the left or right of the parking space.

[0064] The embodiments of this application do not limit the method of obstacle recognition. For example, it can be semantic segmentation. Semantic segmentation refers to labeling each pixel in the environmental image according to semantics (i.e. obstacle category), so that different categories of obstacles can be distinguished in the environmental image. Semantic segmentation can be understood as a pixel-level classification task.

[0065] In some embodiments, after performing obstacle recognition on the environmental image to obtain obstacle information for multiple parking spaces around the vehicle, the parking space recommendation method further includes: projecting the obstacle orientation in the obstacle information of each parking space onto a bird's-eye view image; wherein the bird's-eye view image is obtained by converting the environmental image.

[0066] Here, when a fisheye camera is used to acquire environmental images, the resulting environmental image is a spherical image because its imaging model approximates a spherical projection model. Therefore, the environmental image can be converted into a bird's-eye view image. Correspondingly, the obstacle locations identified from the environmental image are projected onto the bird's-eye view image, which is a planar image, thus providing a more accurate description of obstacle locations. Based on this, subsequent scoring processing is performed based on the obstacle locations in the bird's-eye view image, rather than in the environmental image.

[0067] It is worth noting that fisheye cameras rely on artificially introducing a large amount of distortion to obtain wide-angle image information. Therefore, before converting environmental images into bird's-eye view images, distortion correction processing can be performed on the environmental images to improve the accuracy of the converted bird's-eye view images.

[0068] In step 103, parking information for multiple parking spaces is determined; wherein, the parking information includes parking space type and parking space size.

[0069] Here, parking information for multiple parking spaces around the vehicle is determined. This information includes the type and dimensions of the parking space. The types of parking spaces include perpendicular parking spaces, horizontal parking spaces, angled parking spaces, etc. They can also be classified as line parking spaces or open spaces based on whether they have parking lines. Line parking spaces are those with parking lines, while open spaces are those without. Parking space dimensions include the width and length of the parking space.

[0070] In some embodiments, the above-mentioned determination of parking space information for multiple parking spaces can be achieved by performing any of the following processes: converting an environmental image into a bird's-eye view image and performing parking space recognition on the bird's-eye view image to obtain parking space information for multiple parking spaces; or detecting parking space information for multiple parking spaces using ultrasonic waves.

[0071] Here are two example methods for determining parking space information.

[0072] 1) When the camera is a fisheye camera, the captured environmental image is a spherical image. Therefore, the environmental image can be converted into a bird's-eye view image, and parking space recognition can be performed on the bird's-eye view image to obtain parking space information for multiple parking spaces around the vehicle. Since the bird's-eye view image is a two-dimensional planar image, the accuracy of parking space recognition can be improved. It is worth noting that the parking space information here is obtained through image recognition, and the identified parking spaces are those with parking lines; this parking space information can be called line parking space information.

[0073] 2) Parking information for multiple parking spaces around the vehicle is obtained through ultrasonic detection. This can be achieved using ultrasonic sensors mounted on the side of the vehicle, although the sensor placement is not limited to this. It's important to note that this parking information is obtained in three-dimensional space through ultrasonic detection. The identified parking spaces may or may not have parking lines; this information can be referred to as spatial parking information.

[0074] Either of the two methods can be used, or they can be used simultaneously, which can improve the flexibility of determining parking space information. When both methods are used simultaneously, if both line parking space information and spatial parking space information are detected for the same parking space, the line parking space information shall prevail, because line parking space information is usually more accurate.

[0075] In step 104, a scoring process is performed based on the parking space information and obstacle information of each parking space to obtain multiple attribute scores for each parking space.

[0076] Here, a scoring process is performed based on the parking space information and obstacle information for each parking space, resulting in multiple attribute scores for each space. These attribute scores describe the parking difficulty from different perspectives. For example, there are three types of attribute scores: the first corresponds to the parking space size, the second to the parking space category, and the third to obstacle information.

[0077] This application does not limit the scoring method in its embodiments. For example, the attribute score can be determined by a preset scoring function, where the independent variable of the scoring function is the information involved in the scoring process and the dependent variable is the attribute score; or the attribute score can be determined by a preset mapping relationship, where the mapping relationship refers to the mapping relationship between the information involved in the scoring process and the attribute score.

[0078] In step 105, the scores of multiple attributes of each parking space are fused according to the score fusion model to obtain the fused score of each parking space.

[0079] Here, the scores of multiple attributes of each parking space are fused according to the score fusion model to obtain the fused score of each parking space. The fused score is used to describe the parking difficulty of the parking space in a holistic and comprehensive way. The higher the fused score, the lower the parking difficulty.

[0080] In step 106, parking spaces are recommended based on the fusion score of each parking space.

[0081] Here, parking spaces are recommended based on the integration score of each parking space. The recommended parking spaces can be located by outputting the location of the parking space, such as displaying the location on the in-vehicle display screen, or by announcing the location of the parking space via voice. In addition to outputting the location of the parking space, at least one of the following can also be output: parking space category, parking space size, and obstacle information, so that users (such as car owners) can fully understand the relevant information of the recommended parking spaces.

[0082] It is worth noting that the parking space information determined in step 103 may include the location of the parking space.

[0083] In some embodiments, the above-mentioned recommendation of parking spaces based on the fusion score of each parking space can be implemented in such a way as by performing any of the following processes: recommending the parking space with the highest fusion score; or recommending multiple parking spaces in descending order of fusion score.

[0084] Here, two methods for recommending parking spaces are provided.

[0085] 1) Recommend parking spaces with the highest integration score. The parking space with the highest integration score is the easiest to park in among all the parking spaces around the vehicle. Therefore, only parking spaces with the highest integration score can be recommended to make it easier for users to park their vehicles in the parking spaces with the highest integration score.

[0086] In some embodiments, while recommending the parking space with the highest fusion score, a parking strategy can be planned based on the location of the parking space with the highest fusion score and the vehicle's location. The vehicle is then controlled to park in the parking space with the highest fusion score according to the parking strategy. Besides manual parking by the user, this application embodiment can also achieve automatic parking. For example, a parking strategy can be planned based on the location of the parking space with the highest fusion score and the vehicle's location. The parking strategy includes a parking path from the vehicle's location to the parking space with the highest fusion score, and vehicle control operations (such as steering, accelerator, and brake) to be performed at each location along the parking path. Then, the vehicle is controlled to park in the parking space with the highest fusion score according to the parking strategy, thereby improving the intelligence and accuracy of automatic parking.

[0087] 2) Recommend multiple parking spaces based on their integration scores, from highest to lowest. Here, multiple parking spaces can be recommended based on their integration scores, for example, the higher the integration score of a parking space, the earlier it will be displayed on the in-vehicle display screen. This method provides users with more choices, allowing them to select parking spaces according to their needs and actual circumstances.

[0088] like Figure 3AAs shown, this embodiment of the application comprehensively considers parking space information and obstacle information, so that the calculated fusion score can comprehensively and accurately represent the parking difficulty of the parking space, thereby making the parking space recommended based on the fusion score a parking space with less parking difficulty, which can effectively improve the parking success rate.

[0089] In some embodiments, see Figure 3B , Figure 3B This is a flowchart illustrating a parking space recommendation method provided in an embodiment of this application. Figure 3A Before step 105 shown ( Figure 3B (Taking step 101 as an example), in step 201, you can also obtain the scores of multiple sample attributes of the sample parking space and the sample fusion score of the sample parking space.

[0090] Here, the sample parking space refers to the parking space used as a sample. The scores of various sample attributes of the sample parking space can be obtained in a similar manner to steps 101 to 104. The sample fusion score of the sample parking space can be obtained by manual annotation.

[0091] In step 202, the scores of multiple sample attributes of the sample parking space are fused according to the score fusion model to obtain the fused scores to be compared.

[0092] Here, the scores of multiple sample attributes of the sample parking spaces are fused according to the score fusion model to obtain the fused score. For easy differentiation, the fused score obtained here is named the fused score to be compared.

[0093] In step 203, a loss value is determined based on the sample fusion score and the fusion score to be compared, and a score fusion model is trained based on the loss value.

[0094] Here, the sample fusion score is the expected result, and the fusion score to be compared is the predicted result. Therefore, the loss value is determined based on the sample fusion score and the fusion score to be compared, and the type of loss function is not limited. Then, the score fusion model is trained based on the loss value to improve the score fusion effect of the trained model.

[0095] It is worth noting that there can be multiple sample parking spaces, and steps 201 to 203 above can be executed multiple times to improve the training effect.

[0096] In some embodiments, the above-mentioned score fusion processing of multiple sample attribute scores of sample parking spaces according to the score fusion model can be achieved in the following way: weighted summation of multiple sample attribute scores of sample parking spaces according to the weight parameters in the score fusion model to obtain the fusion score to be compared; the above-mentioned training of the score fusion model based on the loss value can be achieved in the following way: backpropagation is performed in the score fusion model according to the loss value, and the weight parameters in the score fusion model are updated along the gradient descent direction during the backpropagation process.

[0097] Here, the score fusion model includes weight parameters. The essence of score fusion processing can be a weighted summation based on these weight parameters. For example, the weighted summation of multiple sample attribute scores for a parking space is performed using the weight parameters in the score fusion model to obtain the fusion score to be compared. The weight parameters include the weights corresponding to each of the multiple sample attribute scores. For instance, if the sample attribute scores for a parking space include sample attribute score 1, sample attribute score 2, and sample attribute score 3, then the weight parameters include weight 1 for sample attribute score 1, weight 2 for sample attribute score 2, and weight 3 for sample attribute score 3. The final calculated fusion score to be compared is: Sample attribute score 1 × Weight 1 + Sample attribute score 2 × Weight 2 + Sample attribute score 3 × Weight 3. It is worth noting that the score fusion model can also include bias parameters (bias terms). In this case, after performing a weighted summation of the multiple sample attribute scores for the parking space using the weight parameters in the score fusion model, the result of the weighted summation is summed with the bias parameters to obtain the fusion score to be compared. Among them, the bias parameter can improve the expressiveness and flexibility of the model, while effectively preventing the model from getting overfitted during training.

[0098] When the fractional fusion model includes weight parameters, after obtaining the loss value, backpropagation (BP) is performed within the fractional fusion model based on the loss value, and the weight parameters in the fractional fusion model are updated along the gradient descent direction during backpropagation. Stochastic gradient descent (SGD) or other gradient descent algorithms can be used; there are no restrictions. If the fractional fusion model also includes bias parameters, these bias parameters can also be updated along the gradient descent direction during backpropagation.

[0099] In some embodiments, the sample fusion scores of the sample parking spaces include multiple types, and different sample fusion scores correspond to different score fusion strategies. The above-mentioned weighted summation of multiple sample attribute scores of the sample parking spaces based on the weight parameters in the score fusion model can be implemented in the following way: The weighted summation of multiple sample attribute scores of the sample parking spaces based on the weight parameters corresponding to the first score fusion strategy in the score fusion model can be implemented to obtain the fusion score to be compared corresponding to the first score fusion strategy; wherein, the first score fusion strategy represents any one of the multiple score fusion strategies. The above-mentioned determination of the loss value based on the sample fusion scores and the fusion scores to be compared can be implemented in the following way: The loss value corresponding to the first score fusion strategy is determined based on the sample fusion score corresponding to the first score fusion strategy and the fusion score to be compared corresponding to the first score fusion strategy.

[0100] Here, various score fusion strategies can be set, such as aggressive and conservative, without limitation. Based on this premise, the sample fusion scores of the sample parking spaces corresponding to the various score fusion strategies can be obtained; for the score fusion model, it includes weight parameters corresponding to the various score fusion strategies (it may also include bias parameters, but here we only use weight parameters as an example).

[0101] For each score fusion strategy, the loss value can be calculated separately. For ease of understanding, we will take any score fusion strategy (named the first score fusion strategy) as an example. The scores of various sample attributes of the sample parking spaces are weighted and summed according to the weight parameters corresponding to the first score fusion strategy in the score fusion model to obtain the fused score to be compared corresponding to the first score fusion strategy. Then, based on the sample fusion score corresponding to the first score fusion strategy and the fused score to be compared corresponding to the first score fusion strategy, the loss value corresponding to the first score fusion strategy is determined.

[0102] Then, the loss values ​​corresponding to the various score fusion strategies can be fused, and a score fusion model can be trained based on the resulting fused loss value. For example, backpropagation can be performed on the score fusion model based on the fused loss value, and the weight parameters corresponding to the various score fusion strategies can be updated along the gradient descent direction during backpropagation. The method of loss value fusion is not limited; for example, it can be a summation process.

[0103] By using the above methods, the trained score fusion model can support multiple score fusion strategies simultaneously, thereby improving its applicability to different application scenarios.

[0104] exist Figure 3B middle, Figure 3AThe step 105 shown can be updated to step 204, in which the multiple attribute scores of each parking space are fused according to the trained score fusion model to obtain the fused score of each parking space.

[0105] During the training of the score fusion model, the scores of multiple attributes for each parking space are fused according to the trained score fusion model to obtain the fused score for each parking space. For example, if the trained score fusion model includes weight parameters, the scores of multiple attributes for each parking space are weighted and summed according to the weight parameters in the trained score fusion model to obtain the fused score for each parking space; if the trained score fusion model also includes bias parameters, the scores of multiple attributes for each parking space are weighted and summed according to the weight parameters in the trained score fusion model, and then the weighted summation result for each parking space is summed with the bias parameters to obtain the fused score for each parking space.

[0106] In some embodiments, the trained score fusion model includes weight parameters corresponding to various score fusion strategies. The above-mentioned score fusion processing of multiple attribute scores for each parking space based on the trained score fusion model to obtain a fused score for each parking space can be achieved in the following manner: The multiple attribute scores for each parking space are weighted and summed according to the weight parameters corresponding to the second score fusion strategy in the trained score fusion model to obtain a fused score for each parking space corresponding to the second score fusion strategy; wherein, the second score fusion strategy represents the score fusion strategy corresponding to the vehicle among the various score fusion strategies.

[0107] Here, given that the trained score fusion model includes weight parameters (and may also include bias parameters, but we will only use weight parameters as an example) for each of the multiple score fusion strategies, the score fusion strategy corresponding to the vehicle can be determined from among the multiple score fusion strategies (for ease of distinction, the determined score fusion strategy is named the second score fusion strategy). For example, the second score fusion strategy can be determined from among the multiple score fusion strategies by analyzing the vehicle's historical parking habits; or, the names of multiple score fusion strategies can be displayed on the vehicle's in-vehicle display screen, and the score fusion strategy selected by the user can be determined as the second score fusion strategy, thus supporting the user's free choice.

[0108] Then, based on the weight parameters corresponding to the second score fusion strategy in the trained score fusion model, the multiple attribute scores of each parking space are weighted and summed to obtain the fused score for each parking space corresponding to the second score fusion strategy. Through this method, even when multiple score fusion strategies exist, it can be guaranteed that the calculated fused score corresponds to the second score fusion strategy, thus improving the accuracy of the calculated fused score.

[0109] like Figure 3B As shown, this application embodiment improves the score fusion effect of the score fusion model through model training, thereby improving the accuracy of the fusion score predicted for each parking space based on the trained score fusion model.

[0110] In some embodiments, see Figure 3C , Figure 3C This is a flowchart illustrating a parking space recommendation method provided in an embodiment of this application. Figure 3A Step 104 shown can be implemented through steps 301 to 303, which will be explained in conjunction with each step.

[0111] In step 301, the vehicle size and the size difference between each parking space are determined, and a first attribute score for each parking space is determined based on the size difference; wherein, the first attribute score for each parking space is positively correlated with the size difference.

[0112] Here, the attribute score for each parking space includes three types: the first attribute score, the second attribute score, and the third attribute score, which will be explained separately.

[0113] For example, the vehicle dimensions and the size difference between each parking space can be determined, i.e., size difference = parking space size - vehicle dimensions. Then, a first attribute score is determined for each parking space based on the size difference. The first attribute score for each parking space is positively correlated with the size difference; that is, the larger the size difference, the easier it is to park, and the higher the first attribute score.

[0114] The definitions of vehicle size and parking space size are consistent. For example, vehicle size can refer to the longest side of the quadrilateral formed by the vehicles, and parking space size can refer to the longest side of the quadrilateral formed by the parking spaces; or, vehicle size can refer to the shortest side of the quadrilateral formed by the vehicles, and parking space size can refer to the shortest side of the quadrilateral formed by the parking spaces.

[0115] In step 302, a query process is performed in the parking space category mapping relationship according to the parking space category of each parking space to obtain the second attribute score of each parking space; wherein, the parking space category mapping relationship represents the mapping relationship between the parking space category and the second attribute score.

[0116] Here, a parking space category mapping relationship can be preset. This parking space category mapping relationship represents the mapping relationship between the parking space category and the second attribute score. For example, the second attribute score corresponding to a perpendicular parking space is greater than the second attribute score corresponding to a diagonal parking space; the second attribute score corresponding to a line parking space is greater than the second attribute score corresponding to a space parking space.

[0117] After obtaining the parking space category for each parking space, a query is performed in the parking space category mapping relationship based on the parking space category for each parking space to obtain the second attribute score that has a mapping relationship with the parking space category for each parking space.

[0118] In step 303, the obstacle information of each parking space is queried in the obstacle information mapping relationship to obtain the third attribute score of each parking space; wherein, the obstacle information mapping relationship represents the mapping relationship between obstacle information and third attribute score.

[0119] Similarly, an obstacle information mapping relationship can be preset, which represents the mapping relationship between obstacle information and the third attribute score. For example, the third attribute score corresponding to obstacle information that represents that there are no obstacles on the left and right sides of the parking space is larger than the third attribute score corresponding to obstacle information that represents that there is a car on the left side of the parking space.

[0120] After obtaining the obstacle information for each parking space, the obstacle information mapping relationship is queried based on the obstacle information for each parking space to obtain the third attribute score that has a mapping relationship with the obstacle information of each parking space.

[0121] In subsequent steps, the first attribute score, second attribute score, and third attribute score of each parking space can be fused according to the score fusion model to obtain the fused score of each parking space, and parking spaces can be recommended based on the fused score of each parking space.

[0122] like Figure 3C As shown, this application embodiment predefines three types of attribute scores and their corresponding calculation methods, thereby taking into account both parking space information and obstacle information for each parking space. It is worth noting that this does not constitute a limitation on the embodiments of this application; for example, there may be more or fewer types of attribute scores, and the calculation methods corresponding to the attribute scores may also vary.

[0123] The following will describe an exemplary application of the embodiments of this application in a real-world application scenario. For ease of understanding, the description will be presented in the form of steps.

[0124] Step 1: Sensing parking space information and obstacle information.

[0125] 1) Obstacle information recognition.

[0126] The vehicle's surround-view fisheye camera captures environmental images, which are then semantically segmented to extract obstacle information (or obstacle semantic information) around each parking space. This obstacle information includes obstacle type and obstacle location. Obstacle types include pillars, walls, pedestrians, vehicles, traffic cones, water-filled barriers, and parking lines. Obstacle location indicates whether the obstacle is located to the left or right of the parking space.

[0127] It is worth noting that, in addition to identifying obstacle information, the embodiments of this application can also identify more information related to the vehicle's surrounding environment, such as the width of the parking aisle.

[0128] 2) Obstacle information projection.

[0129] Obstacle information (mainly obstacle orientation) extracted from the environmental image is projected onto the bird's-eye view to obtain obstacle information described in the bird's-eye view for subsequent steps.

[0130] 3) Identification of parking space information.

[0131] The environmental image is distorted and converted into a planar bird's-eye view (BEV) image. Parking space information, specifically parking spaces with parking lines, is then identified within the BEV image. In this embodiment, the parking space information may include the location, type, and size of the parking space.

[0132] 4) Spatial parking space information recognition.

[0133] Ultrasonic detection is performed using ultrasonic sensors on the side of the vehicle. Based on the obtained direct echo (DE) and cross echo (CE) values, the spatial parking space area is fitted, and the height value (i.e., the z-axis value) is removed to obtain spatial parking space information described in a two-dimensional plane. The parking space information may or may not have parking space lines.

[0134] It is worth noting that the line parking space information recognition method and the spatial parking space information recognition method can be used individually or simultaneously. If both line parking space information and spatial parking space information are recognized for a certain parking space, the line parking space information shall prevail, because line parking space information is usually more accurate.

[0135] By identifying parking space information along lines and / or identifying parking space information in spatial areas, parking space information for each parking space can be obtained for subsequent steps.

[0136] Step 2: Constructing parking space attribute scores.

[0137] In this embodiment, the quality of a parking space depends primarily on the parking space itself and its external environment. Therefore, by combining the parking space information and obstacle information obtained in step one, an attribute score is constructed for the parking space. The attribute score mainly refers to the parking space category, parking space size, and obstacle information. The obstacle categories referred to may include pillars, walls, and vehicles.

[0138] For ease of understanding, the embodiments of this application provide, as follows: Figure 4 The diagram shows the different types of parking spaces. Figure 4 It shows five types of parking spaces: vertical parking spaces, horizontal parking spaces, diagonal parking spaces, vertical space parking spaces, and horizontal space parking spaces. Figure 4 The target parking space in the context refers to any one of the identified parking spaces.

[0139] For ease of understanding, the embodiments of this application provide, as follows: Figure 5 The diagram illustrates the following obstacle scenarios: ① No obstacles on either side; ② Car on the left, no obstacle on the right; ③ Car on the right, no obstacle on the left; ④ Cars on both sides; ⑤ Pillar / wall on the left, no obstacle on the right; ⑥ Pillar / wall on the left, car on the right; ⑦ No obstacle on the left, pillar / wall on the right; ⑧ Car on the left, pillar / wall on the right; ⑨ Pillar / wall on both sides. Each obstacle scenario corresponds to one type of obstacle information.

[0140] Step 3: Parking space attribute score scoring method.

[0141] Parking space attribute scores can be divided into two categories: the first category is the attribute score corresponding to the parking space size based on objective standards (corresponding to the first attribute score mentioned above); the second category is related to user habits, such as the attribute score corresponding to the parking space category (corresponding to the second attribute score mentioned above) and the attribute score corresponding to obstacle information (corresponding to the third attribute score mentioned above). An example scoring method for parking space attribute scores is as follows:

[0142] 1) Adaptive vehicle body parameter scoring method (suitable for parking space size).

[0143]

[0144] In this formula, S (x) The first attribute score is represented by x, where x represents the parking space size, d represents the vehicle size, and Δ represents the safety threshold. For example, x represents the parking space width in vertical / diagonal parking spaces and the parking space length in horizontal parking spaces; d represents the vehicle width in vertical / diagonal parking spaces and the vehicle length in horizontal parking spaces; the value of Δ may differ between vertical / diagonal and horizontal parking spaces, and can be set according to the actual application scenario. Figure 6 The example of a parking space on the Sino-Israeli line illustrates the possible values ​​of x and d.

[0145] 2) User habit mapping method (expert rating method).

[0146] For example, a mapping relationship between parking space categories and second attribute scores can be established in advance to achieve scoring of parking space categories; a mapping relationship between obstacle information and third attribute scores can be established in advance to achieve scoring of obstacle information.

[0147] The values ​​of attribute scores (such as second attribute scores or third attribute scores) can be as shown in Table 1.

[0148] Table 1 shows the possible values ​​of the attribute scores.

[0149] The value of attribute score Scale level 1 Low 3 middle 5 high 2,4 Between adjacent scale medians

[0150] Step 4: Obtain the sample fusion score matrix R. Where:

[0151]

[0152] In this formula, U represents the sample attribute score (i.e., the attribute score of the sample parking space); V represents the weight matrix (corresponding to the weight parameters mentioned above); represents the bias term (bias parameter); i represents the number of samples (number of sample parking spaces); k represents the type of sample attribute score, for example, it can be 3; j represents the type of score fusion strategy, which can be set according to the actual application scenario. For example, the score fusion strategy can include two types: aggressive and conservative, and j can be 2.

[0153] The purpose of this application's embodiments is to find a suitable weight matrix V such that... With R i×j The differences between them should be minimized.

[0154] Step 5: Predict the fusion score. The formula for predicting the fusion score is as follows:

[0155]

[0156] in, This represents the fusion score predicted for the i-th sample parking space according to the j-th score fusion strategy (corresponding to the fusion score to be compared above); u ik v represents the score of the k-th sample attribute for the i-th sample parking space; jk This represents the weight of the k-th sample attribute score in the weight parameters corresponding to the j-th score fusion strategy.

[0157] Step 6: Solve for v using SGD j .

[0158] The objective function is as follows:

[0159]

[0160] Where f is the loss value, u i Let U be a row matrix of U. Based on this, the gradient can be calculated using SGD:

[0161]

[0162] After calculating the gradient, update v along the gradient descent direction. j This yields the final weight matrix V.

[0163] Step 7: Calculate the fusion score of multiple parking spaces around the vehicle, and select the parking space with the highest fusion score for recommendation.

[0164] Through steps one through three above, we can obtain the various attribute scores of multiple parking spaces around the vehicle. For ease of explanation, we will use u' to represent the various attribute scores of a certain parking space. Then, we can obtain the fusion score of that parking space as follows:

[0165]

[0166] This application embodiment identifies parking space and obstacle information through visual and ultrasonic perception. It considers multiple aspects such as successful route planning, user alighting convenience, and parking space size, using adaptive vehicle parameter scoring and expert scoring to quantify evaluation indicators and obtain quantifiable attribute scores. The sample fusion score matrix is ​​decomposed using Funk SVD, and an objective function for optimization is defined. The weight matrix is ​​solved using the SGD algorithm. Thus, the fusion score for each parking space can be calculated using the weight matrix and the attribute scores of each parking space. By sorting the fusion scores and recommending the parking space with the highest fusion score to the user, the parking success rate can be improved. Furthermore, since the calculation of the fusion score takes alighting convenience into account, the intelligence of parking space recommendations and the user experience are enhanced.

[0167] The following continues to describe the exemplary structure of the parking space recommendation device 455 provided in the embodiments of this application as a software module. In some embodiments, such as Figure 2 As shown, the software modules in the parking space recommendation device 455 stored in the memory 450 may include: a data acquisition module 4551, used to acquire environmental images through the vehicle's camera; an obstacle recognition module 4552, used to perform obstacle recognition on the environmental images to obtain obstacle information for multiple parking spaces around the vehicle; wherein, the obstacle information includes obstacle category and obstacle location; a parking space recognition module 4553, used to determine parking space information for multiple parking spaces; wherein, the parking space information includes parking space category and parking space size; a scoring module 4554, used to perform scoring processing based on the parking space information and obstacle information of each parking space to obtain multiple attribute scores for each parking space; a score fusion module 4555, used to perform score fusion processing on the multiple attribute scores of each parking space according to the score fusion model to obtain a fused score for each parking space; and a recommendation module 4556, used to recommend parking spaces based on the fused score of each parking space.

[0168] In some embodiments, the parking space recommendation device 455 further includes a training module, which is used to acquire multiple sample attribute scores of sample parking spaces and sample fusion scores of sample parking spaces; perform score fusion processing on the multiple sample attribute scores of sample parking spaces according to the score fusion model to obtain a fusion score to be compared; determine a loss value based on the sample fusion score and the fusion score to be compared, and train the score fusion model based on the loss value; the score fusion module 4555 is also used to perform score fusion processing on the multiple attribute scores of each parking space according to the trained score fusion model to obtain a fusion score for each parking space.

[0169] In some embodiments, the training module is further configured to: perform weighted summation of multiple sample attribute scores of the sample parking spaces according to the weight parameters in the score fusion model to obtain the fusion score to be compared; perform backpropagation in the score fusion model according to the loss value, and update the weight parameters in the score fusion model along the gradient descent direction during the backpropagation process.

[0170] In some embodiments, the scoring module 4554 is further configured to: determine the size difference between the vehicle size and the size of each parking space, and determine a first attribute score for each parking space based on the size difference; wherein the first attribute score for each parking space is positively correlated with the size difference; perform query processing in the parking space category mapping relationship based on the parking space category of each parking space to obtain a second attribute score for each parking space; wherein the parking space category mapping relationship represents the mapping relationship between the parking space category and the second attribute score; and perform query processing in the obstacle information mapping relationship based on the obstacle information of each parking space to obtain a third attribute score for each parking space; wherein the obstacle information mapping relationship represents the mapping relationship between the obstacle information and the third attribute score.

[0171] In some embodiments, the parking space recognition module 4553 is further configured to perform any of the following processes: converting an environmental image into a bird's-eye view image and performing parking space recognition on the bird's-eye view image to obtain parking space information of multiple parking spaces; or detecting parking space information of multiple parking spaces by ultrasonic waves.

[0172] In some embodiments, the obstacle recognition module 4552 is further configured to: project the obstacle orientation from the obstacle information of each parking space onto a bird's-eye view image; wherein the bird's-eye view image is obtained by converting an environmental image.

[0173] In some embodiments, the recommendation module 4556 is further configured to perform any of the following processes: recommend the parking space with the highest fusion score; recommend multiple parking spaces in descending order of fusion score.

[0174] This application provides a computer program product or computer program that includes executable instructions stored in a computer-readable storage medium. A processor of an electronic device reads the executable instructions from the computer-readable storage medium and executes the executable instructions, causing the electronic device to perform the parking space recommendation method described in this application.

[0175] This application provides a computer-readable storage medium storing executable instructions, wherein the executable instructions are stored and, when executed by a processor, will cause the processor to execute the parking space recommendation method provided in this application.

[0176] In some embodiments, the computer-readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash memory, magnetic surface memory, optical disk, or CD-ROM; or it may be a variety of devices including one or any combination of the above-mentioned memories.

[0177] In some embodiments, executable instructions may take the form of a program, software, software module, script, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a standalone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0178] As an example, executable instructions may, but do not necessarily, correspond to files in a file system. They may be stored as part of a file that holds other programs or data, for example, in one or more scripts in a Hyper Text Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple collaborating files (e.g., a file that stores one or more modules, subroutines, or code sections).

[0179] As an example, executable instructions can be deployed to execute on a single computing device, or on multiple computing devices located in one location, or on multiple computing devices distributed across multiple locations and interconnected via a communication network.

[0180] The above are merely embodiments of this application and are not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, and improvements made within the spirit and scope of this application are included within the scope of protection of this application.

Claims

1. A parking space recommendation method, characterized in that, include: The vehicle's cameras capture images of the environment. Obstacle recognition is performed on the environmental image to obtain obstacle information for multiple parking spaces around the vehicle; wherein, the obstacle information includes obstacle type and obstacle location; Determine the parking information of the plurality of parking spaces; wherein, the parking information includes the parking space category and the parking space size; Scoring is performed on each parking space based on its parking space information and obstacle information to obtain multiple attribute scores for each parking space. This scoring process includes: determining the vehicle size and the size difference between each parking space, and determining a first attribute score for each parking space based on the size difference; wherein the first attribute score for each parking space is positively correlated with the size difference; and querying the obstacle information of each parking space in an obstacle information mapping relationship to obtain a third attribute score for each parking space; wherein the obstacle information mapping relationship represents the mapping relationship between obstacle information and the third attribute score. The scores of multiple attributes for each parking space are fused according to the score fusion model to obtain the fused score for each parking space. Parking spaces are recommended based on their integration score.

2. The method according to claim 1, characterized in that, Before performing score fusion processing on the multiple attribute scores of each parking space according to the score fusion model to obtain the fused score for each parking space, the method further includes: Obtain the scores of multiple sample attributes of the sample parking space and the sample fusion score of the sample parking space; The scores of multiple sample attributes of the sample parking spaces are fused according to the score fusion model to obtain the fused scores to be compared. The loss value is determined based on the sample fusion score and the fusion score to be compared, and the score fusion model is trained based on the loss value; The step involves fusing multiple attribute scores for each parking space using a score fusion model to obtain a fused score for each parking space, including: The multi-attribute scores of each parking space are fused using the trained score fusion model to obtain the fused score for each parking space.

3. The method according to claim 2, characterized in that, The step of performing score fusion processing on the multiple sample attribute scores of the sample parking spaces according to the score fusion model to obtain the fused score to be compared includes: The scores of various sample attributes of the sample parking spaces are weighted and summed according to the weight parameters in the score fusion model to obtain the fusion score to be compared. The step of training the score fusion model based on the loss value includes: Backpropagation is performed in the fractional fusion model based on the loss value, and the weight parameters in the fractional fusion model are updated along the gradient descent direction during the backpropagation process.

4. The method according to claim 1, characterized in that, The scoring process, which involves evaluating each parking space based on its information and obstacles, to obtain multiple attribute scores for each parking space, also includes: Based on the parking space category of each parking space, a query process is performed in the parking space category mapping relationship to obtain the second attribute score of each parking space; wherein, the parking space category mapping relationship represents the mapping relationship between the parking space category and the second attribute score.

5. The method according to claim 1, characterized in that, Determining the parking space information of the plurality of parking spaces includes: Perform any of the following processes: The environmental image is converted into a bird's-eye view image, and parking space identification is performed on the bird's-eye view image to obtain the parking space information of the multiple parking spaces; The parking information of the multiple parking spaces is detected by ultrasonic waves.

6. The method according to claim 1, characterized in that, After performing obstacle recognition on the environmental image to obtain obstacle information for multiple parking spaces around the vehicle, the method further includes: The obstacle location in the obstacle information of each parking space is projected onto the bird's-eye view image; wherein the bird's-eye view image is obtained by converting the environmental image.

7. The method according to any one of claims 1 to 6, characterized in that, The method of recommending parking spaces based on the fusion score of each parking space includes: Perform any of the following processes: Recommend the parking space with the highest fusion score; The multiple parking spaces are recommended in descending order of their fusion scores.

8. A parking space recommendation device, characterized in that, include: The data acquisition module is used to acquire environmental images through the vehicle's cameras; An obstacle recognition module is used to identify obstacles in the environmental image to obtain obstacle information for multiple parking spaces around the vehicle; wherein, the obstacle information includes obstacle category and obstacle location; A parking space recognition module is used to determine the parking space information of the plurality of parking spaces; wherein, the parking space information includes the parking space category and the parking space size; The scoring module is used to score each parking space based on its parking space information and obstacle information, obtaining multiple attribute scores for each parking space. This scoring process includes: determining the vehicle size and the size difference between the dimensions of each parking space, and determining a first attribute score for each parking space based on the size difference; wherein the first attribute score for each parking space is positively correlated with the size difference; and querying the obstacle information of each parking space in an obstacle information mapping relationship to obtain a third attribute score for each parking space; wherein the obstacle information mapping relationship represents the mapping relationship between obstacle information and the third attribute score. The score fusion module is used to perform score fusion processing on the multiple attribute scores of each parking space according to the score fusion model to obtain the fused score of each parking space. The recommendation module is used to recommend parking spaces based on the fusion score of each parking space.

9. An electronic device, characterized in that, include: Memory, used to store executable instructions; A processor, when executing executable instructions stored in the memory, implements the parking space recommendation method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It stores executable instructions for implementing the parking space recommendation method according to any one of claims 1 to 7 when executed by a processor.

Citation Information

Patent Citations

  • Automatic parking recommendation method, electronic equipment and automobile

    CN110304051A

  • Method and apparatus for guiding a recommended parking location

    KR1020180069174A