Intelligent parking space recognition navigation system and method based on artificial intelligence

Abstract

The invention relates to the technical field of parking space recognition navigation, and discloses an intelligent parking space recognition navigation system and method based on artificial intelligence, and the method comprises the steps: obtaining a parking lot image through an image obtaining module through a sensor network; the parking space recognition module recognizes the idle state of the parking space according to the parking lot image, so that a vehicle owner can know the parking space information in time, and the trouble of blindly finding the parking space is avoided; the state interaction module provides parking space selection according to the idle state of the parking space and generates a parking space selection result, so that the parking convenience and satisfaction are improved; the path planning module reduces the driving distance of the vehicle owner according to the parking space selection result and the current position of the vehicle owner, and reduces the congestion risk of the vehicle in the parking lot; the navigation control module provides navigation guidance according to the parking path, guides the vehicle owner to move to the target parking space and generates a vehicle owner feedback result; and the feedback adjustment module dynamically adjusts the parking space selection result according to the navigation guidance effect and the feedback result of the vehicle owner.

Description

Technical Field

[0001] The present invention relates to the technical field of parking space recognition and navigation, and in particular to an intelligent parking space recognition and navigation system and method based on artificial intelligence. Background Art

[0002] With the development of the smart car industry, research on intelligent parking recognition and navigation systems has gained increasing attention. These systems can monitor parking status in real time and quickly guide drivers to available spaces, significantly improving parking efficiency and reducing search time. By effectively managing vehicle flow within parking lots, they help reduce traffic congestion caused by searching for parking spaces and enhance the overall smoothness of urban traffic. Furthermore, they can reduce inefficient driving within parking lots, lower fuel consumption and exhaust emissions, and contribute positively to environmental protection. By providing convenient parking services, these systems enhance the user's parking experience and reduce parking stress and anxiety. However, most systems lack the solution to how to acquire parking lot images through sensor networks, identify parking space availability, provide parking space selection, determine the driver's parking path based on their current location, guide the driver to their selected available space, and generate feedback from the driver to dynamically adjust the path planning strategy.

[0003] For example, Chinese patent publication number CN111047898A discloses a real-time dynamic navigation system for designated parking spaces in an underground parking lot. The system comprises several license plate recognition and processing modules, a wired / wireless communication network, a central processing computer, an intelligent terminal, central processing computer information processing software, intelligent terminal information processing software, a database of license plate recognition module codes and their geographic locations, a database of dynamic correspondences between license plate numbers and intelligent terminals, an electronic parking space distribution map database, a parking space route guide line database, a parking space geographic location database, a vehicle schematic database, a voice prompt module database, and a database of partially enlarged guidance maps for parking lot intersections. Using the real-time electronic parking space distribution map, designated parking space guide lines, synchronized vehicle schematics, real-time voice prompts, and a real-time partially enlarged guidance map for parking lot intersections, drivers can quickly and directly reach their designated parking spaces. The system provides real-time and authentic guidance, providing a user-friendly experience similar to outdoor GPS navigation.

[0004] For example, Chinese patent publication CN114519938B discloses a positioning and navigation system and method based on parking space number image recognition. The system comprises a server and a client terminal, connected via server communication. The server includes a parking space image learning module, an image analysis module, a location recording module, and a location query module. The client includes a parking space recording module and a vehicle retrieval module. The parking space recording module records parking locations by capturing images with a camera or uploading them locally. The vehicle retrieval module navigates to the recorded parking location using a combination of indoor and outdoor positioning. This system offers lower implementation and maintenance costs, a stable system, high positioning accuracy, and the ability to achieve two-way navigation using indoor and outdoor navigation. It can also be integrated with a user's mobile terminal, making it more convenient and requiring less power.

[0005] The above patents have the problems raised by this background technology: the above-mentioned real-time dynamic parking navigation system uses a real-time electronic parking space distribution map, guide lines for designated parking spaces, a schematic diagram of vehicles moving synchronously with the vehicle, real-time voice prompts, and a real-time partial zoom guidance map of the parking lot intersection to drive the vehicle to the designated parking space quickly and directly; the above-mentioned positioning navigation system records the parking location by taking pictures with a camera or uploading images locally, and navigates to the recorded parking location in combination with indoor and outdoor positioning. The above two patents do not solve the problem of how to obtain parking lot images through a sensor network to identify the vacancy status of parking spaces, provide parking space selection, obtain the owner's parking path in combination with the owner's current location, and navigate the owner to the selected vacant parking space, and generate feedback results from the owner to dynamically adjust the path planning strategy. To solve this problem, the present invention proposes an intelligent parking space recognition and navigation system and method based on artificial intelligence. Summary of the Invention

[0006] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of this application to avoid obscuring the purpose of this section, the abstract and the title of the invention, and such simplifications or omissions should not be used to limit the scope of the present invention.

[0007] In view of the above-mentioned problems existing in the existing intelligent parking space recognition and navigation system and method based on artificial intelligence, the present invention is proposed.

[0008] Therefore, the object of the present invention is to provide an intelligent parking space recognition and navigation system and method based on artificial intelligence.

[0009] To solve the above technical problems, the present invention provides an intelligent parking space recognition and navigation system based on artificial intelligence, comprising:

[0010] Image acquisition module, parking space recognition module, state interaction module, path planning module, navigation control module and feedback adjustment module;

[0011] The image acquisition module is used to acquire parking lot images through a sensor network;

[0012] The parking space recognition module is used to recognize the vacant state of a parking space based on the parking lot image;

[0013] The state interaction module is used to provide parking space selection according to the vacancy status of the parking space and generate a parking space selection result;

[0014] The path planning module is used to obtain a parking path through a path planning strategy based on the parking space selection result and the current location of the car owner;

[0015] The navigation control module is used to provide navigation guidance according to the parking path, guide the owner to move to the target parking space, and generate feedback results from the owner;

[0016] The feedback adjustment module is used to dynamically adjust the parking space selection result according to the effect of the navigation guidance and the feedback result of the car owner.

[0017] As a preferred solution of the intelligent parking space recognition and navigation system based on artificial intelligence described in the present invention, the strategy for identifying the vacant state of a parking space includes:

[0018] The captured parking lot image is preprocessed and divided into regions to generate a regional image. Feature extraction is performed on the regional image to obtain regional features. Classification is performed based on the regional features to obtain the vacancy status of the parking space.

[0019] The function expression of the idle state of a parking space is as follows:

[0020]

[0021] Where q represents the vacant state of parking spaces in the parking lot, 1 represents an empty parking space, and 0 represents an occupied parking space. j )>T represents the region feature p of the jth region image j The probability of being classified as a free parking space is greater than the state threshold T, else means that the ... j )>T.

[0022] As a preferred solution of the artificial intelligence-based intelligent parking space recognition and navigation system of the present invention, the strategy for providing parking space selection based on the vacancy status of parking spaces and generating parking space selection results includes:

[0023] Record vacant parking spaces and their attributes, including parking space location, parking space size, and parking space safety. Assign a weight coefficient to each parking space attribute, calculate the attribute value of each parking space attribute, obtain a parking space selection value based on the weight coefficient and the attribute value of each parking space attribute, and generate a parking space selection result based on the parking space selection value.

[0024] The function expression of the parking space selection value is as follows:

[0025]

[0026] In the formula, C represents the parking space selection value, B represents the set of available parking spaces, A represents the set of parking space attributes for each available parking space, and w a represents the weight coefficient of parking space attribute a, and h(b,a) represents the attribute value of parking space b in parking space attribute a.

[0027] As a preferred solution of the intelligent parking space recognition and navigation system based on artificial intelligence described in the present invention, the function expression of the attribute value of parking space b in parking space attribute a is as follows:

[0028]

[0029] h(b,a2)=c(b) / c max ;

[0030]

[0031] In the formula, h(b,a1) represents the attribute value of parking space b at parking space position a1, d max represents the distance threshold, d(b) represents the actual distance between parking space b and the current location of the owner, h(b,a2) represents the attribute value of parking space b in parking space size a2, c(b) represents the actual parking space size of parking space b, c max represents the parking space size threshold, h(b,a3) represents the attribute value of parking space b in parking space safety a3, β1 represents the quantity weight coefficient, u represents the number of cameras in the parking lot, u max represents the number threshold of cameras in the parking lot, β2 represents the brightness weight coefficient, l represents the current brightness of the parking lot image, and l max represents the brightness threshold of the parking lot image, β3 represents the safety weight coefficient, and sm represents the safety factor of the parking space surrounding environment.

[0032] As a preferred solution of the artificial intelligence-based intelligent parking space recognition and navigation system described in the present invention, the path planning strategy includes:

[0033] The target parking space is set according to the parking space selection result, the distance of the owner's current position is set to 0, the distance of the vacant parking space in the parking lot is set to infinity, and the vacant parking spaces in the parking lot are set to an array. When the array is not empty, the first vacant parking space is extracted from the array. If the first vacant parking space is the target parking space, the loop process is exited;

[0034] If the first vacant parking space is not the target parking space, find the second vacant parking space adjacent to the first vacant parking space, calculate the distance from the owner's current position to the second vacant parking space, and record it as the difference distance. If the difference distance is less than the distance of the second vacant parking space, update the distance to the second vacant parking space and update the array synchronously until the target parking space is found. By backtracking the vacant parking spaces, a parking path from the owner's current position to the target parking space is generated.

[0035] As a preferred solution of the artificial intelligence-based intelligent parking space recognition and navigation system described in the present invention, the logic for providing navigation guidance based on the parking path, guiding the driver to move to the target parking space, and generating driver feedback results includes:

[0036] After the car owner generates a parking space selection result, the system determines the car owner's current location and provides navigation guidance based on the parking path. The car owner is guided to the target parking space through voice broadcast, and the car owner's feedback data on the navigation guidance is collected. The collected feedback data is analyzed to generate the car owner's feedback results.

[0037] As a preferred solution of the artificial intelligence-based intelligent parking space recognition and navigation system described in the present invention, the strategy for acquiring parking lot images through the sensor network includes:

[0038] The camera height and angle are configured according to the parking lot layout, the lidar scanning frequency is adjusted according to the traffic volume in the parking lot, and infrared sensors are deployed in the parking lot. The parking lot image is captured through a sensor network combining cameras, lidar, and infrared sensors, and the camera exposure time is adjusted according to the brightness of the parking lot image.

[0039] The function expression for adjusting the scanning frequency of the lidar according to the traffic flow in the parking lot is as follows:

[0040] f=f0×(1+λ×(r-r0));

[0041] Where f represents the scanning frequency of the lidar after adjustment, f0 represents the standard scanning frequency of the lidar, λ represents the adjustment coefficient, r represents the current traffic flow in the parking lot, and r0 represents the traffic flow threshold.

[0042] An intelligent parking space recognition and navigation method based on artificial intelligence, comprising:

[0043] S1, obtain parking lot images through sensor network;

[0044] S2, identifying the vacant status of parking spaces based on parking lot images;

[0045] S3. Providing parking space selection based on the availability of parking spaces and generating a parking space selection result;

[0046] S4. Obtain a parking path through a path planning strategy based on the parking space selection result and the current location of the vehicle owner;

[0047] S5. Provide navigation guidance based on the parking path, guide the car owner to move to the target parking space, and generate feedback results for the car owner;

[0048] S6. Dynamically adjust the parking space selection result based on the effect of the navigation guidance and the feedback from the car owner.

[0049] A computer device includes a memory for storing instructions; and a processor for executing the instructions, so that the computer device implements an intelligent parking space recognition and navigation method based on artificial intelligence.

[0050] A computer-readable storage medium stores a computer program, which, when executed, implements an intelligent parking space recognition and navigation method based on artificial intelligence.

[0051] The beneficial effects of the present invention are as follows: the present invention obtains parking lot images through a sensor network through an image acquisition module; the parking space recognition module recognizes the vacant status of parking spaces based on the parking lot images, ensuring that car owners can timely understand parking space information and avoid the trouble of blindly searching for parking spaces; the state interaction module provides parking space selection according to the vacant status of parking spaces and generates parking space selection results, so that car owners can easily view and select parking spaces that suit them, thereby improving the convenience and satisfaction of parking; the path planning module obtains a parking path through a path planning strategy based on the parking space selection results and the current location of the car owner, thereby reducing the driving distance of the car owner and reducing the risk of vehicle congestion in the parking lot; the navigation control module provides navigation guidance according to the parking path, guides the car owner to move to the target parking space, and generates car owner feedback results, providing the car owner with detailed navigation guidance, and guiding the car owner to move quickly and accurately to the selected vacant parking space; the feedback adjustment module dynamically adjusts the parking space selection results according to the effect of the navigation guidance and the car owner feedback results, so that the system can adapt to different parking lot environments and car owner needs, and continuously improve recognition accuracy, navigation accuracy and user experience.

[0052] By reducing the ineffective driving time and distance of vehicles in parking lots, it helps to reduce fuel consumption and exhaust emissions, has positive significance for environmental protection, facilitates urban parking management, and helps optimize parking resource allocation and management decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be derived from these drawings without inventive work. Among them:

[0054] Figure 1 This is a system structure diagram of an artificial intelligence-based intelligent parking space recognition and navigation system of the present invention;

[0055] Figure 2 This is a resolution comparison chart of an artificial intelligence-based intelligent parking space recognition and navigation system of the present invention;

[0056] Figure 3 A strategy diagram for generating parking space selection results of an artificial intelligence-based intelligent parking space recognition and navigation system of the present invention;

[0057] Figure 4 This is a flow chart of an artificial intelligence-based intelligent parking space recognition and navigation method of the present invention. DETAILED DESCRIPTION

[0058] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0059] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can also be implemented in other ways different from those described herein. Those skilled in the art can make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0060] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0061] Example 1

[0062] In this embodiment, a system structure diagram of an intelligent parking space recognition and navigation system based on artificial intelligence is provided, such as Figure 1 As shown, an artificial intelligence-based intelligent parking space recognition and navigation system includes an image acquisition module, a parking space recognition module, a state interaction module, a path planning module, a navigation control module and a feedback adjustment module.

[0063] The image acquisition module is used to acquire parking lot images through the sensor network.

[0064] Strategies for acquiring parking lot images through sensor networks include:

[0065] The camera height and angle are configured according to the parking lot layout, the lidar scanning frequency is adjusted according to the traffic flow in the parking lot, and infrared sensors are deployed in the parking lot. The parking lot image is captured through a sensor network combining cameras, lidar and infrared sensors, and the camera exposure time is adjusted according to the brightness of the parking lot image.

[0066] The function expression for adjusting the scanning frequency of the lidar according to the traffic flow in the parking lot is as follows:

[0067] f=f0×(1+λ×(r-r0));

[0068] Where f represents the scanning frequency of the lidar after adjustment, f0 represents the standard scanning frequency of the lidar, λ represents the adjustment coefficient, r represents the current traffic flow in the parking lot, and r0 represents the traffic flow threshold.

[0069] It should be explained that the standard scanning frequency f0 of the lidar is the standard scanning frequency setting under normal or low traffic conditions, and is an initial value that is not affected by real-time traffic flow; the adjustment coefficient λ is used to express the degree of influence of traffic flow changes on the scanning frequency of the lidar, reflecting the sensitivity to changes in traffic flow; the current traffic flow r in the parking lot is the traffic flow data obtained by the camera, usually expressed as the number of vehicles passing through a certain area in the parking lot per unit time; the traffic flow threshold r0 is usually a reference value, representing the normal range or critical point of traffic flow. According to the setting of safety requirements, when the traffic flow is too much, the critical value of the scanning frequency needs to be increased.

[0070] The function expression for adjusting the camera exposure time according to the brightness of the parking lot image is as follows:

[0071] t=t0×(l0 / l);

[0072] Where t represents the exposure time of the camera after adjustment, t0 represents the basic exposure time of the camera, l0 represents the target brightness of the parking lot image, and l represents the current brightness of the parking lot image.

[0073] The function expression of the current brightness of the parking lot image is as follows:

[0074]

[0075] Where l represents the current brightness of the parking lot image, n represents the total number of pixels in the parking lot image, and I(x i ,y i ) represents the coordinate (x i ,yi )’s brightness value.

[0076] It should be explained that the total number of pixels n in the parking lot image represents the resolution of the parking lot image, which is usually the product of the parking lot image's width and height. The brightness value of each pixel in the parking lot image can be provided by the parking lot image file itself, usually expressed as a grayscale value. The above formula is designed to measure the overall lighting conditions of the parking lot image. By taking the average brightness value of all pixels, it can effectively reflect the current brightness level of the parking lot image.

[0077] A sensor network including cameras, lidars, and infrared sensors is deployed in the parking lot. The camera installation angle needs to be selected according to the parking lot layout to maximize the field of view coverage. At the same time, high-resolution cameras are used in key areas (such as entrances and exits) to enable clear vehicle recognition. Figure 2 As shown in the figure, (a) represents a parking lot scene captured by a low-resolution camera, and (b) represents a parking lot scene captured by a high-resolution camera. For the lidar, the scanning frequency needs to be adjusted according to the traffic flow in the parking lot. When the traffic flow is high, the scanning frequency needs to be increased to capture more dynamic information and ensure that the lidar can cover the entire parking lot to avoid blind spots. At night or in low-light environments, the sensitivity of the infrared sensor needs to be increased to ensure that it can effectively monitor vehicles. At the same time, the monitoring range of the infrared sensor needs to be adjusted according to actual needs to ensure that it can cover important areas of the parking lot. The adaptive adjustment of the sensor network through the above two formulas can achieve efficient parking lot image acquisition and environmental perception, which is helpful for the subsequent identification of the vacancy status of parking spaces.

[0078] The parking space recognition module is used to identify the vacant status of parking spaces based on parking lot images.

[0079] Strategies for identifying the vacancy status of parking spaces include:

[0080] The captured parking lot image is preprocessed and divided into regions to generate a regional image. Feature extraction is performed on the regional image to obtain regional features. Classification is performed based on the regional features to obtain the vacancy status of the parking space.

[0081] The function expression of the idle state of a parking space is as follows:

[0082]

[0083] Where q represents the vacant state of parking spaces in the parking lot, 1 represents an empty parking space, and 0 represents an occupied parking space. j )>T represents the region feature p of the jth region image jThe probability of being classified as a free parking space is greater than the state threshold T, else means that the ... j )>T.

[0084] Among them, the j-th region image feature p j The functional expression of the probability of being classified as a parking space vacant is as follows:

[0085] P(q=1|p j )=e z1,j / (e z0,j +e z1,j );

[0086] Where, P(q=1|p j ) represents the regional feature p of the jth region image j The probability of being classified as a vacant parking space, e z1,j The output result of the image of the jth area where the parking space is available is represented by e z0,j Represents the output result of parking space occupancy in the jth area image.

[0087] It should be explained that: e represents the base of the natural logarithm, which is usually used for exponential operations, z1,j represents the regional feature of the j-th regional image that is a vacant parking space, and z0,j represents the regional feature of the j-th regional image that is an occupied parking space. According to the exponential function, the regional feature values ​​of vacant or occupied parking spaces are converted into probability values, which represent the probability of the regional feature of the given j-th regional image being judged as a vacant or occupied parking space.

[0088] In the specific application, the parking lot image is preprocessed, including denoising, scaling and normalization. The regions are divided into the left, center and right sides of the parking lot image to generate regional images. The regional images are input into the convolutional neural network for feature extraction. The convolutional neural network automatically learns the key features of the regional images through multi-layer convolution and pooling operations. The extracted regional features are input into the fully connected layer for classification to determine the vacancy status of parking spaces in the parking lot. In the convolutional neural network, the basic calculation of regional feature extraction is expressed as:

[0089] F m =σ(w m *F m-1 +ε m );

[0090] Where, F m represents the feature map of the mth layer, σ(·) represents the activation function, and w m represents the convolution kernel of the mth layer, * represents the convolution operation, F m-1 represents the feature map of the m-1th layer, ε m represents the bias term.

[0091] A state threshold is set and the vacant status of parking spaces in the parking lot is identified based on the output probability. When a misidentification or missed identification is detected during re-inspection, relevant re-inspection samples are collected and the learning rate of the convolutional neural network is adjusted based on the relevant re-inspection samples. By adjusting the learning rate of the convolutional neural network, the extraction of regional features and the classification of the vacant status of parking spaces are optimized, and new regional images are used for retraining. The above process can effectively improve the recognition accuracy of vacant parking spaces in the parking lot and adapt to the ever-changing environment.

[0092] The status interaction module is used to provide parking space selection according to the vacancy status of the parking space and generate a parking space selection result.

[0093] When there are vacant parking spaces in the parking lot, parking space selection is provided, and the strategy for generating parking space selection results is as follows: Figure 3 As shown, specifically including:

[0094] Record the vacant parking spaces and the attributes of the vacant parking spaces, where the parking space attributes include parking space location, parking space size, and parking space safety. Assign a weight coefficient to each parking space attribute, calculate the attribute value of each parking space attribute, obtain the parking space selection value based on the weight coefficient and the attribute value of each parking space attribute, and generate the parking space selection result based on the parking space selection value.

[0095] The function expression of the parking space selection value is as follows:

[0096]

[0097] In the formula, C represents the parking space selection value, B represents the set of available parking spaces, A represents the set of parking space attributes for each available parking space, and w a represents the weight coefficient of parking space attribute a, and h(b,a) represents the attribute value of parking space b in parking space attribute a.

[0098] What needs to be explained is: the weight coefficient w of parking space attribute a a It reflects the importance car owners attach to the attributes of different parking spaces.

[0099] Among them, the function expressions of the attribute value of parking space b on the corresponding parking space attribute a are as follows:

[0100]

[0101] h(b,a2)=c(b) / c max ;

[0102]

[0103] In the formula, h(b,a1) represents the attribute value of parking space b at parking space position a1, d maxrepresents the distance threshold, d(b) represents the actual distance between parking space b and the current location of the owner, h(b,a2) represents the attribute value of parking space b in parking space size a2, c(b) represents the actual parking space size of parking space b, c max represents the parking space size threshold, h(b,a3) represents the attribute value of parking space b in parking space safety a3, β1 represents the quantity weight coefficient, u represents the number of cameras in the parking lot, u max represents the number threshold of cameras in the parking lot, β2 represents the brightness weight coefficient, l represents the current brightness of the parking lot image, and l max represents the brightness threshold of the parking lot image, β3 represents the safety weight coefficient, and sm represents the safety factor of the parking space surrounding environment.

[0104] It should be explained that the closer parking space b is to the current location of the car owner, the more convenient the parking space is. max It is expressed as the maximum distance acceptable to the car owner; the size of parking space b directly affects whether the vehicle can be parked smoothly. Generally speaking, the larger the parking space, the stronger its ability to accommodate different types of vehicles. The parking space size threshold c max It is expressed as the maximum available parking space area; the larger the attribute value h(b,a3) of parking space b in parking space safety a3, the safer the parking space b is. The value range of the safety factor sm of the parking space surrounding environment is 0 to 1, which is expressed as the safety of the surrounding environment; at the same time, the data in different units mentioned above are converted to the same scale to facilitate weighted summation.

[0105] By considering multiple factors, including the availability of parking spaces, their location, size, and safety, the system can provide users with a more comprehensive and high-quality parking space selection service. For parking space location, the system considers the distance from the vehicle owner's current location. For parking space size, the system considers the ratio of the actual size of the parking space to the maximum available area. In actual situations, it may also be necessary to consider that different types of vehicles have different requirements for parking space size. For parking space safety, the system considers the number of cameras, the brightness of the parking lot image, and the safety factor of the surrounding environment. The more cameras there are and the wider the coverage, the safer the parking space. The brightness conditions of the parking lot image help generate clear images and ensure the safety of the parking space. At the same time, the surrounding environment of the parking space can also affect the safety of the parking space. For example, parking spaces near exits, fire lanes, or busy roads may pose safety hazards. Attribute values ​​are obtained based on the three parking space attributes of parking space location, parking space size, and parking space safety, helping the system provide car owners with more accurate and personalized parking space selection services.

[0106] The path planning module is used to obtain the parking path through the path planning strategy based on the parking space selection result and the current location of the owner.

[0107] Path planning strategies include:

[0108] The target parking space is set according to the parking space selection result, the distance of the owner's current position is set to 0, the distance of the vacant parking space in the parking lot is set to infinity, and the vacant parking spaces in the parking lot are set to an array. When the array is not empty, the first vacant parking space is extracted from the array. If the first vacant parking space is the target parking space, the loop process is exited;

[0109] If the first vacant parking space is not the target parking space, find the second vacant parking space adjacent to the first vacant parking space, calculate the distance from the owner's current position to the second vacant parking space, and record it as the difference distance. If the difference distance is less than the distance of the second vacant parking space, update the distance to the second vacant parking space and update the array synchronously until the target parking space is found. By backtracking the vacant parking spaces, a parking path from the owner's current position to the target parking space is generated.

[0110] In a specific application, the first vacant parking space extracted is the one with the shortest distance from the owner's current position, and the distance from the owner's current position to the second vacant parking space needs to be calculated using the Euclidean distance formula. The basic idea of ​​the above path planning strategy is to select an vacant parking space closest to the owner's current position from the vacant parking spaces that have not been extracted each time, and then update the distances to its adjacent vacant parking spaces. Through the above path planning, the best path from the owner's current position to the target parking space can be effectively found, which not only ensures the rationality of the selection of the target parking space, but also optimizes the search for the optimal path by using distance calculation and arrays. Efficiency, where the distance to the second available parking space refers to the actual distance from the driver's current location to the second available space. If the calculated distance difference is less than the currently stored distance of the second available space, the distance of the second available space is updated. Therefore, the new distance will be used to replace the old distance with each update to ensure that the distance information in the array is up to date. The ultimate goal is to find the shortest parking path to the target parking space. At the same time, in actual situations, it is also necessary to consider obstacles, stairs, and elevators in the path. Try to choose a path that is easy to pass through and avoid areas with heavy traffic or obstructed vision to ensure the safety of the driver.

[0111] The navigation control module is used to provide navigation guidance based on the parking path, guide the owner to move to the target parking space, and generate feedback results for the owner.

[0112] The logic for providing navigation guidance based on the parking path, guiding the driver to the target parking space, and generating driver feedback includes:

[0113] When the car owner generates a parking space selection result, the car owner's current position is determined through wireless technology, and navigation guidance is provided according to the parking path to ensure that the car owner can get accurate navigation information throughout the parking process, avoiding troubles caused by navigation interruption or inaccuracy. The car owner is guided to move to the target parking space through voice broadcast. The position determination through wireless technology improves the accuracy of positioning and the precision of navigation, allowing car owners to find vacant parking spaces more easily. The feedback data of car owners on navigation guidance is collected, and the collected feedback data is analyzed to generate car owner feedback results. The collected feedback data includes information such as satisfaction with the navigation service and improvement suggestions, and identifies problems and deficiencies in the system, and proposes targeted improvement measures, which can help to continuously improve the performance of the system and user experience.

[0114] The feedback adjustment module is used to dynamically adjust the parking space selection results based on the effect of navigation guidance and the feedback results of the car owner.

[0115] The parking space selection results are dynamically adjusted based on the effectiveness of navigation guidance and feedback from car owners. For example, if the attribute values ​​of multiple available parking spaces are the same, these calculation results are cached, and the multiplication operation of the weight coefficient of the parking space attribute and the attribute value of the parking space attribute is parallelized to improve calculation efficiency.

[0116] To optimize the calculation formula for parking space selection values, it is necessary to express the weight coefficients and attribute values ​​of parking space attributes in matrix form, accelerate the calculation process through matrix operations, and dynamically adjust the weight coefficients of parking space attributes based on the frequency of parking space use and user preferences.

[0117] Example 2

[0118] In this embodiment, a method flow chart of an intelligent parking space recognition and navigation method based on artificial intelligence is provided, such as Figure 4 As shown, an intelligent parking space recognition and navigation method based on artificial intelligence includes:

[0119] S1, obtain parking lot images through sensor network;

[0120] A sensor network consisting of cameras, lidars, and infrared sensors is deployed in the parking lot. The camera installation angle must be selected at a suitable height and angle based on the parking lot layout to maximize field of view coverage. High-resolution cameras are used in key areas (such as entrances and exits) for clear vehicle identification. The lidar's scanning frequency must be adjusted based on the volume of traffic in the parking lot. During high traffic, the scanning frequency must be increased to capture more dynamic information and ensure that the lidar covers the entire parking lot to avoid blind spots. At night or in low-light environments, the sensitivity of the infrared sensor must be increased to ensure effective vehicle monitoring. The infrared sensor's monitoring range must be adjusted based on actual needs to ensure coverage of important areas of the parking lot. Adaptive adjustment of the sensor network using the two formulas above enables efficient parking lot image acquisition and environmental perception, facilitating subsequent identification of parking space vacancy.

[0121] S2, identifying the vacant status of parking spaces based on parking lot images;

[0122] In the specific application, the parking lot image is preprocessed, including denoising, scaling and normalization. The regions are divided into the left, center and right sides of the parking lot image to generate regional images. The regional images are input into the convolutional neural network for feature extraction. The convolutional neural network automatically learns the key features of the regional images through multi-layer convolution and pooling operations. The extracted regional features are input into the fully connected layer for classification to determine the vacancy status of parking spaces in the parking lot. In the convolutional neural network, the basic calculation of regional feature extraction is expressed as:

[0123] F m =σ(w m *F m-1 +ε m );

[0124] Where, F m represents the feature map of the mth layer, σ represents the activation function, w m represents the convolution kernel of the mth layer, * represents the convolution operation, F m-1 represents the feature map of the m-1th layer, ε m represents the bias term.

[0125] A state threshold is set and the vacant status of parking spaces in the parking lot is identified based on the output probability. When a misidentification or missed identification is detected during re-inspection, relevant re-inspection samples are collected and the learning rate of the convolutional neural network is adjusted based on the relevant re-inspection samples. By adjusting the learning rate of the convolutional neural network, the extraction of regional features and the classification of the vacant status of parking spaces are optimized, and new regional images are used for retraining. The above process can effectively improve the recognition accuracy of vacant parking spaces in the parking lot and adapt to the ever-changing environment.

[0126] S3. Providing parking space selection based on the availability of parking spaces and generating a parking space selection result;

[0127] By considering multiple factors, including the availability of parking spaces, their location, size, and safety, the system can provide users with a more comprehensive and high-quality parking space selection service. For parking space location, the system considers the distance from the vehicle owner's current location. For parking space size, the system considers the ratio of the actual size of the parking space to the maximum available area. In actual situations, it may also be necessary to consider that different types of vehicles have different requirements for parking space size. For parking space safety, the system considers the number of cameras, the brightness of the parking lot image, and the safety factor of the surrounding environment. The more cameras there are and the wider the coverage, the safer the parking space. The brightness conditions of the parking lot image help generate clear images and ensure the safety of the parking space. At the same time, the surrounding environment of the parking space can also affect the safety of the parking space. For example, parking spaces near exits, fire lanes, or busy roads may pose safety hazards. Attribute values ​​are obtained based on the three parking space attributes of parking space location, parking space size, and parking space safety, helping the system provide car owners with more accurate and personalized parking space selection services.

[0128] S4. Obtain a parking path through a path planning strategy based on the parking space selection result and the current location of the vehicle owner;

[0129] In a specific application, the first vacant parking space extracted is the one with the shortest distance from the owner's current position, and the distance from the owner's current position to the second vacant parking space needs to be calculated using the Euclidean distance formula. The basic idea of ​​the above path planning strategy is to select an vacant parking space closest to the owner's current position from the vacant parking spaces that have not been extracted each time, and then update the distances to its adjacent vacant parking spaces. Through the above path planning, the optimal path from the owner's current position to the target parking space can be effectively found, which not only ensures the rationality of the selection of the target parking space, but also optimizes the search efficiency of the optimal path by using distance calculation and arrays. At the same time, in actual situations, it is also necessary to consider obstacles, stairs, elevators, etc. in the path, try to choose a path that is easy to pass through, avoid passing through areas with large traffic or obstructed vision, and ensure the safety of the owner.

[0130] S5. Provide navigation guidance based on the parking path, guide the car owner to move to the target parking space, and generate feedback results for the car owner;

[0131] The logic for providing navigation guidance based on the parking path, guiding the driver to the target parking space, and generating driver feedback includes:

[0132] When the car owner generates a parking space selection result, the car owner's current position is determined through wireless technology, and navigation guidance is provided according to the parking path to ensure that the car owner can get accurate navigation information throughout the parking process, avoiding troubles caused by navigation interruption or inaccuracy. The car owner is guided to move to the target parking space through voice broadcast. The position determination through wireless technology improves the accuracy of positioning and the precision of navigation, allowing car owners to find vacant parking spaces more easily. The feedback data of car owners on navigation guidance is collected, and the collected feedback data is analyzed to generate car owner feedback results. The collected feedback data includes information such as satisfaction with the navigation service and improvement suggestions, and identifies problems and deficiencies in the system, and proposes targeted improvement measures, which can help to continuously improve the performance of the system and user experience.

[0133] S6. Dynamically adjust the parking space selection result based on the effect of the navigation guidance and the feedback from the car owner.

[0134] The parking space selection results are dynamically adjusted based on the effectiveness of navigation guidance and feedback from car owners. For example, if the attribute values ​​of multiple available parking spaces are the same, these calculation results are cached, and the multiplication operation of the weight coefficient of the parking space attribute and the attribute value of the parking space attribute is parallelized to improve calculation efficiency.

[0135] To optimize the calculation formula for parking space selection values, it is necessary to express the weight coefficients and attribute values ​​of parking space attributes in matrix form, accelerate the calculation process through matrix operations, and dynamically adjust the weight coefficients of parking space attributes based on the frequency of parking space use and user preferences.

[0136] The specific content of an intelligent parking space recognition and navigation method based on artificial intelligence can be found in an intelligent parking space recognition and navigation system based on artificial intelligence, which will not be described in detail here.

[0137] Example 3

[0138] In this embodiment, a computer device is provided, including a memory and a processor, wherein the memory is used to store instructions and the processor is used to execute the instructions, so that the computer device performs the steps of implementing the above-mentioned artificial intelligence-based intelligent parking space recognition and navigation method.

[0139] Example 4

[0140] In this embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed, the steps of the above-mentioned intelligent parking space recognition and navigation method based on artificial intelligence are implemented.

[0141] The computer-readable storage medium includes various media for storing program codes, such as a USB flash drive, a mobile hard disk, a read-only memory, a random access memory, a magnetic disk, or an optical disk.

[0142] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, and all of these should be included in the scope of the claims of the present invention.

Claims

1. An intelligent parking space recognition and navigation system based on artificial intelligence, characterized in that: include: Image acquisition module, parking space recognition module, state interaction module, path planning module, navigation control module and feedback adjustment module; The image acquisition module is used to acquire parking lot images through a sensor network; The parking space recognition module is used to recognize the vacant state of a parking space according to the parking lot image; The state interaction module is used to provide parking space selection according to the vacancy state of the parking space and generate a parking space selection result. The strategy for generating the parking space selection result includes obtaining a parking space selection value according to the attribute value of the parking space attribute of the vacancy state of the parking space, and generating the parking space selection result according to the parking space selection value. The parking space attributes include the parking space location, the parking space size and the parking space safety, wherein the parking space attribute is the attribute value of the parking space safety, which is calculated and processed according to the number of cameras in the parking lot, the current brightness of the parking lot image and the safety factor of the parking space surrounding environment; The path planning module is used to obtain a parking path through a path planning strategy according to the parking space selection result and the current location of the car owner; The navigation control module is used to provide navigation guidance according to the parking path, guide the owner to move to the target parking space, and generate feedback results from the owner; The feedback adjustment module is used to dynamically adjust the parking space selection result according to the effect of the navigation guidance and the feedback result of the car owner.

2. The intelligent parking space recognition and navigation system based on artificial intelligence as claimed in claim 1, characterized in that: The strategy for identifying the idle state of a parking space includes: The captured parking lot image is preprocessed, and the parking lot image is divided into regions to generate a regional image, feature extraction is performed on the regional image to obtain regional features, and classification is performed based on the regional features to obtain the vacancy status of the parking space; The function expression of the idle state of the parking space is as follows: In the formula, q represents the idle state of parking spaces in the parking lot, 1 represents an idle parking space, and 0 represents an occupied parking space. j )>T represents the region feature p of the jth region image j The probability of being classified as a free parking space is greater than the state threshold T, else means except P (q = 1 | p j )>Situations other than T.

3. The intelligent parking space recognition and navigation system based on artificial intelligence as claimed in claim 2, characterized in that: The strategy of providing parking space selection according to the free status of the parking space and generating the parking space selection result includes: Record the vacant parking spaces and the attributes of the vacant parking spaces, the parking space attributes including the parking space location, parking space size and parking space safety, assign a weight coefficient to each parking space attribute, calculate the attribute value of each parking space attribute, obtain the parking space selection value according to the weight coefficient and the attribute value of each parking space attribute, and generate the parking space selection result according to the parking space selection value; The function expression of the parking space selection value is as follows: In the formula, C represents the parking space selection value, B represents the set of free parking spaces, A represents the set of parking space attributes for each free parking space, and w a represents the weight coefficient of parking space attribute a, and h(b,a) represents the attribute value of parking space b in parking space attribute a.

4. The intelligent parking space recognition and navigation system based on artificial intelligence as claimed in claim 3, characterized in that: The function expression of the attribute value of parking space b in parking space attribute a is as follows: h(b,a2)=c(b) / c max ; In the formula, h(b,a1) represents the attribute value of parking space b at parking space position a1, d max represents the distance threshold, d(b) represents the actual distance between parking space b and the current position of the owner, h(b,a2) represents the attribute value of parking space b in parking space size a2, c(b) represents the actual parking space size of parking space b, c max represents the parking space size threshold, h(b,a3) represents the attribute value of parking space b in parking space safety a3, β1 represents the quantity weight coefficient, u represents the number of cameras in the parking lot, and u max represents the number threshold of cameras in the parking lot, β2 represents the brightness weight coefficient, l represents the current brightness of the parking lot image, and l max represents the brightness threshold of the parking lot image, β3 represents the safety weight coefficient, and sm represents the safety factor of the parking space surrounding environment.

5. The intelligent parking space recognition and navigation system based on artificial intelligence as claimed in claim 4, characterized in that: The path planning strategy includes: The target parking space is set according to the parking space selection result, the distance of the owner's current position is set to 0, the distance of the free parking space in the parking lot is set to infinity, and the free parking spaces in the parking lot are set to an array. When the array is not empty, the first free parking space is extracted from the array. If the first free parking space is the target parking space, the loop process is exited; If the first vacant parking space is not the target parking space, find the second vacant parking space adjacent to the first vacant parking space, calculate the distance from the owner's current position to the second vacant parking space, and record it as the difference distance. If the difference distance is less than the distance of the second vacant parking space, update the distance to the second vacant parking space and update the array synchronously until the target parking space is found. Generate a parking path from the owner's current position to the target parking space by backtracking the vacant parking spaces.

6. The intelligent parking space recognition and navigation system based on artificial intelligence as claimed in claim 5, characterized in that: The logic of providing navigation guidance according to the parking path, guiding the car owner to move to the target parking space, and generating the car owner's feedback result includes: When the car owner generates a parking space selection result, the current location of the car owner is determined, and navigation guidance is provided according to the parking path. The car owner is guided to move to the target parking space through voice broadcast, and the feedback data of the car owner on the navigation guidance is collected. The collected feedback data is analyzed to generate the car owner feedback result.

7. The intelligent parking space recognition and navigation system based on artificial intelligence as claimed in claim 6, characterized in that: The strategy for acquiring parking lot images through the sensor network includes: The height and angle of the camera are configured according to the layout of the parking lot. The scanning frequency of the LiDAR is adjusted according to the traffic volume in the parking lot. Infrared sensors are configured in the parking lot. The parking lot image is captured through a sensor network combining cameras, LiDAR and infrared sensors. The exposure time of the camera is adjusted according to the brightness of the parking lot image. The function expression for adjusting the scanning frequency of the laser radar according to the traffic flow in the parking lot is as follows: f = f0 × (1 + λ × (r - r0)); Where f represents the scanning frequency of the laser radar after adjustment, f0 represents the standard scanning frequency of the laser radar, λ represents the adjustment coefficient, r represents the current traffic flow in the parking lot, and r0 represents the traffic flow threshold.

8. An intelligent parking space recognition and navigation method based on artificial intelligence, implemented based on an intelligent parking space recognition and navigation system based on artificial intelligence according to any one of claims 1 to 7, characterized in that: include: S1, obtain parking lot images through sensor network; S2, identifying the vacancy status of parking spaces according to the parking lot image; S3. Providing parking space selection according to the vacancy status of the parking space and generating a parking space selection result; S4, obtaining a parking path through a path planning strategy according to the parking space selection result and the current location of the car owner; S5. Provide navigation guidance according to the parking path, guide the car owner to move to the target parking space, and generate feedback results for the car owner; S6. Dynamically adjust the parking space selection result according to the effect of the navigation guidance and the feedback from the car owner.

9. A computer device, characterized in that: include: A memory for storing instructions; The processor is used to execute the instructions so that the computer device implements the intelligent parking space recognition and navigation method based on artificial intelligence as described in claim 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed, the intelligent parking space recognition and navigation method based on artificial intelligence as described in claim 8 is implemented.

Images