Indoor parking lot surveying and mapping method, system and equipment based on artificial intelligence and medium

Abstract

The invention discloses an indoor parking lot surveying and mapping method, system and device based on artificial intelligence and a medium, and relates to the technical field of surveying and mapping, and the method comprises the steps: obtaining first image data and second image data which are successively collected in a parking lot, and determining a passable area of the parking lot according to the second image data; performing vehicle state change detection on the passable area according to the first image data and the second image data to obtain at least one changed parking space; for each changed parking space, determining whether a vehicle on the changed parking space is an unrecorded vehicle according to the second image data, and if yes, putting the changed parking space into a target changed parking space set; based on a minimum spanning tree algorithm, generating parking space sequence data according to the target change parking space set; and generating a model based on a surveying and mapping scheme, and determining surveying and mapping parameters according to the parking space sequence data. The method has the effect of improving the intelligent level of parking lot surveying and mapping.

Description

Technical Field

[0001] The present application relates to the technical field of surveying and mapping, and in particular to an indoor parking lot surveying and mapping method, system, equipment and medium based on artificial intelligence. Background Art

[0002] Indoor parking lot mapping is an important technical link in urban planning and traffic management. With the acceleration of urbanization and the growing demand for parking, the scientific management of parking lots is particularly important. Traditional indoor parking lot mapping mainly relies on manual operation and is not closely integrated with modern technology, which limits the efficiency and accuracy of mapping.

[0003] With the rapid development of automation and intelligent technology, indoor parking lot mapping methods are gradually developing in the direction of intelligence. Using advanced technologies such as laser scanning, drone aerial photography, and 3D modeling, it is possible to obtain accurate data of parking lots in a short time and generate highly detailed maps. However, many parking lots still use traditional methods, such as manual measurement and simple plane layout drawing, which is not only time-consuming and labor-intensive, but also prone to errors, and the resulting data updates are relatively lagging.

[0004] Existing intelligent parking systems are mainly based on information management. Although they are combined with certain data analysis and real-time monitoring functions, the overall level of intelligence is still insufficient. On the one hand, many parking lots lack high-precision real-time data collection methods, and information such as parking lot layout and parking space usage often rely on regular manual maintenance; on the other hand, the dynamic changes within the parking lot, such as the frequency of use of parking spaces and vacancy, lack real-time monitoring and feedback mechanisms, resulting in low management efficiency.

[0005] In summary, although the technical conditions for indoor parking lot mapping are constantly improving, the overall level of intelligence is still low, and it is necessary to further introduce advanced technologies and intelligent management concepts to improve the operational efficiency and user experience of parking lots. The improvement of intelligence can not only improve the management efficiency of parking lots, but also provide users with more convenient parking services and promote the smooth and harmonious development of urban traffic. Summary of the invention

[0006] In order to improve the intelligence level of parking lot mapping, the present application provides an indoor parking lot mapping method, system, equipment and medium based on artificial intelligence.

[0007] In the first aspect, the present application provides an indoor parking lot mapping method based on artificial intelligence, which adopts the following technical solutions: Indoor parking lot mapping method based on artificial intelligence, including: Acquire first image data and second image data collected successively in the parking lot, and determine a passable area of ​​the parking lot according to the second image data; Performing vehicle status change detection on the passable area according to the first image data and the second image data to obtain at least one changed parking space, wherein the changed parking space is a changed and occupied parking space; For each changed parking space, determining whether the vehicle in the changed parking space is an unrecorded vehicle according to the second image data, and if so, placing the changed parking space into a target changed parking space set, wherein the unrecorded vehicle is a vehicle that is not parked in the parking lot; Calling the three-dimensional model of the parking lot, based on the minimum spanning tree algorithm, and changing the parking space set according to the target to generate parking space sequence data, wherein the parking space sequence data includes a parking space sequence and a distance sequence corresponding to the parking space sequence; A model is generated based on a surveying and mapping scheme, and surveying and mapping parameters are determined according to the parking space sequence data. A surveying and mapping vehicle is controlled to survey the parking lot according to the surveying and mapping parameters, wherein the surveying and mapping parameters include an optimal number of the surveying and mapping vehicles and a surveying and mapping order of the target changed parking spaces by the surveying and mapping vehicles.

[0008] By adopting the above technical solution, the first image data and the second image data collected in the parking lot are first obtained, and the passable area of ​​the parking lot is determined according to the second image data. Then, the vehicle state change detection is performed on the passable area according to the first image data and the second image data to obtain at least one changed parking space, wherein the changed parking space is a changed and occupied parking space, and then for each changed parking space, it is determined according to the second image data whether the vehicle in the changed parking space is an unrecorded vehicle. If so, the changed parking space is placed in the target changed parking space set, wherein the unrecorded vehicle is a vehicle that is not parked in the parking lot, and then the three-dimensional image data of the parking lot is called. The model is based on the minimum spanning tree algorithm, and generates parking space sequence data according to the target change parking space set, wherein the parking space sequence data includes the parking space sequence and the distance sequence corresponding to the parking space sequence. Finally, the model is generated based on the surveying and mapping scheme, and the surveying and mapping parameters are determined according to the parking space sequence data. The surveying and mapping parameters are used to control the surveying and mapping vehicle to survey the parking lot, wherein the surveying and mapping parameters include the optimal number of surveying and mapping vehicles and the surveying and mapping order of the surveying and mapping vehicle for the target change parking spaces. Through the above method, the surveying and mapping of vehicles in the parking lot is realized during the use of the parking lot, the intelligent level of the parking lot surveying is improved, and the management and monitoring of the vehicles in the parking lot are strengthened.

[0009] Optionally, the step of determining the passable area of ​​the parking lot according to the second image data includes: Performing grayscale conversion on the second image data according to the second image data to obtain a corresponding grayscale image; A pixel matrix corresponding to the grayscale image is obtained according to the grayscale image, and elements in the pixel matrix are compared with a preset threshold value to obtain first coordinate data and second coordinate data, wherein the first coordinate data is coordinate data corresponding to a matrix region in the pixel matrix whose element values ​​are within a preset range, and the second coordinate data is coordinate data corresponding to a matrix region in the pixel matrix whose element values ​​are not within the preset range, The area corresponding to the first coordinate data is regarded as a passable area, and the area corresponding to the second coordinate data is regarded as a prohibited passable area.

[0010] By adopting the above technical solution, in order to determine the passable area of ​​the parking lot, the second image data is gray-converted according to the second image data to obtain the corresponding gray-scale image, and then the pixel matrix corresponding to the gray-scale image is obtained according to the gray-scale image, and the elements in the pixel matrix are compared with a preset threshold to obtain the first coordinate data and the second coordinate data, wherein the first coordinate data is the coordinate data corresponding to the matrix area in which the element values ​​in the pixel matrix are within the preset range, and the second coordinate data is the coordinate data corresponding to the matrix area in which the element values ​​in the pixel matrix are not within the preset range, and finally the area corresponding to the first coordinate data is used as the passable area, and the area corresponding to the second coordinate data is used as the prohibited area.

[0011] Optionally, the step of performing vehicle status change detection on the passable area according to the first image data and the second image data to obtain a parking space change includes: Extracting features in the passable area according to the first image data to obtain a first vehicle state feature; Extracting features in the passable area according to the second image data to obtain a second vehicle state feature; A target parking space to be changed in the passable area is determined by comparing the first vehicle state feature with the second vehicle state feature.

[0012] By adopting the above technical solution, in order to obtain the changed parking space, feature extraction is performed on the passable area according to the first image data to obtain the first vehicle state feature, and then feature extraction is performed on the passable area according to the second image data to obtain the second vehicle state feature. Finally, the first vehicle state feature and the second vehicle state feature are compared to determine the target changed parking space in the passable area.

[0013] Optionally, for each changed parking space, the step of determining whether the vehicle in the changed parking space is an unrecorded vehicle according to the second image data includes: For each changed parking space, performing license plate clarity recognition on the changed parking space according to the second image data to obtain the license plate clarity of the changed parking space; Determine whether the license plate clarity is greater than a preset value according to the license plate clarity; if so, obtain the license plate number according to the second image data, and match the license plate number in a license plate number data table; if no match is found, it indicates that the vehicle in the changed parking space is an unrecorded vehicle, wherein the license plate number data table includes license plate number data and vehicle feature data, and the license plate number data table is used to indicate a mapping relationship between the license plate number data and the vehicle feature data; If not, feature recognition is performed on the vehicle in the changed parking space based on the second image data to obtain vehicle features, and the vehicle is matched in the license plate number data table. If the match is unsuccessful, it indicates that the vehicle in the changed parking space is an unrecorded vehicle.

[0014] By adopting the above technical solution, in order to determine whether the vehicle in the changed parking space is an unrecorded vehicle, for each changed parking space, the license plate clarity recognition is performed on the changed parking space according to the second image data to obtain the license plate clarity of the changed parking space, and then it is determined whether the license plate clarity is greater than a preset value according to the license plate clarity. If the license plate clarity is greater than the preset value, the license plate number is obtained according to the second image data, and the license plate number is matched in the license plate number data table. If no match is found, it indicates that the vehicle in the changed parking space is an unrecorded vehicle, wherein the license plate number data table includes license plate number data and vehicle feature data, and the license plate number data table is used to indicate the mapping relationship between the license plate number data and the vehicle feature data. If the license plate clarity is not greater than the preset value, the feature recognition is performed on the vehicle in the changed parking space according to the second image data to obtain the vehicle feature, and the vehicle is matched in the license plate number data table. If no match is successful, it indicates that the vehicle in the changed parking space is an unrecorded vehicle.

[0015] Optionally, the step of calling the three-dimensional model of the parking lot, based on a minimum spanning tree algorithm, and changing the parking space set according to the target to generate parking space sequence data includes: Calling the three-dimensional model of the parking lot, and determining a starting parking space for mapping according to the three-dimensional model, wherein the starting parking space for mapping is the parking space in the target change parking space set that is closest to the storage point of the mapping vehicle; Taking the initial surveyed parking space as the starting point, based on the minimum spanning tree algorithm and according to the target change parking space set, the parking space sequence data is determined.

[0016] By adopting the above technical solution, in order to generate parking space sequence data, the three-dimensional model of the parking lot is called, and the starting mapping parking space is determined according to the three-dimensional model, wherein the starting mapping parking space is the parking space closest to the mapping car storage point in the target change parking space set, and then the parking space sequence data is determined based on the minimum spanning tree algorithm and the target change parking space set with the starting mapping parking space as the starting point.

[0017] Optionally, the objective function of the surveying and mapping scheme generation model is:

[0018] The constraints of the objective function are:

[0019] Among them, among them, represents the sum of the kth segment, represents the jth element in the distance sequence, n represents the number of parts into which the distance sequence is divided, d represents the distance sequence, X is the maximum value of each part, Indicates the starting position of the first paragraph. represents the end position of the nth segment, m represents the total number of elements in the distance sequence, Indicates the starting position of the kth segment, Indicates the starting position of the k+1th segment.

[0020] Optionally, the step of generating a model based on a surveying and mapping scheme and determining surveying and mapping parameters according to the parking space sequence data includes: Generate a model through a surveying and mapping scheme, and determine the optimal number of surveying and mapping vehicles according to the distance sequence; The order in which the surveying and mapping vehicle surveys the target changed parking spaces is determined by the constraint conditions of the objective function and according to the optimal number, the parking space sequence and the distance sequence.

[0021] By adopting the above technical scheme, in order to determine the surveying and mapping parameters, a model is generated through the surveying and mapping scheme, and the optimal number of surveying and mapping vehicles is determined according to the distance sequence. Then, through the constraints of the objective function and based on the optimal number, parking space sequence and distance sequence, the surveying and mapping order of the target changed parking space by the surveying and mapping vehicle is determined.

[0022] In the second aspect, the present application also provides an indoor parking lot mapping system based on artificial intelligence, which adopts the following technical solutions: Indoor parking lot mapping system based on artificial intelligence, including: A passable area determination module, used to obtain first image data and second image data collected successively in the parking lot, and determine the passable area of ​​the parking lot according to the second image data; a changed parking space generation module, configured to detect a vehicle state change in the passable area according to the first image data and the second image data, and obtain at least one changed parking space, wherein the changed parking space is a changed and occupied parking space; a target changed parking space set generation module, for each changed parking space, determining whether the vehicle in the changed parking space is an unrecorded vehicle according to the second image data, and if so, placing the changed parking space into a target changed parking space set, wherein the unrecorded vehicle is a vehicle that is not parked in the parking lot; A parking space sequence data generating module, configured to call the three-dimensional model of the parking lot, based on a minimum spanning tree algorithm, and generate parking space sequence data according to the target changing parking space set, wherein the parking space sequence data includes a parking space sequence and a distance sequence corresponding to the parking space sequence; The surveying and mapping module generates a model based on the surveying and mapping scheme, determines surveying and mapping parameters according to the parking space sequence data, and controls the surveying and mapping vehicle to survey the parking lot according to the surveying and mapping parameters, wherein the surveying and mapping parameters include the optimal number of the surveying and mapping vehicles and the surveying and mapping order of the target changed parking spaces by the surveying and mapping vehicles.

[0023] In a third aspect, the present application also provides a computer device, which adopts the following technical solution: A computer device comprises a memory and a processor, wherein the memory stores a computer program executable on the processor, and the processor implements the method described in the first aspect when executing the computer program.

[0024] In a fourth aspect, the present application further provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium stores a computer program that can be loaded by a processor and execute the method described in the first aspect.

[0025] In summary, the present application includes at least the following beneficial technical effects: firstly, the first image data and the second image data collected successively in the parking lot are obtained, and the passable area of ​​the parking lot is determined according to the second image data, and then the vehicle status change detection is performed on the passable area according to the first image data and the second image data to obtain at least one changed parking space, wherein the changed parking space is a parking space that has been changed and occupied, and then for each changed parking space, it is determined according to the second image data whether the vehicle in the changed parking space is an unrecorded vehicle, and if so, the changed parking space is placed in the target changed parking space set, wherein the unrecorded vehicle is a vehicle that is not parked in the parking lot, and then the parking space is called. The three-dimensional model of the parking lot is based on the minimum spanning tree algorithm, and according to the target change parking space set, the parking space sequence data is generated, wherein the parking space sequence data includes the parking space sequence and the distance sequence corresponding to the parking space sequence. Finally, the model is generated based on the surveying and mapping scheme, and the surveying and mapping parameters are determined according to the parking space sequence data. The surveying and mapping vehicle is controlled to survey the parking lot according to the surveying and mapping parameters, wherein the surveying and mapping parameters include the optimal number of surveying and mapping vehicles and the surveying and mapping order of the surveying and mapping vehicle for the target change parking spaces. Through the above method, the surveying and mapping of vehicles in the parking lot is realized during the use of the parking lot, the intelligent level of the parking lot surveying is improved, and the management and monitoring of vehicles in the parking lot are strengthened. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a schematic diagram of the overall process of an embodiment of the present application.

[0027] Figure 2 It is a schematic diagram of the structure of the system of this application.

[0028] Figure 3 It is a structural block diagram of the computer device of the present application. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0030] The embodiment of the present application discloses an indoor parking lot mapping method based on artificial intelligence.

[0031] Reference Figure 1 , an indoor parking lot mapping method based on artificial intelligence, including: Step S11, acquiring first image data and second image data collected successively in the parking lot, and determining a passable area of ​​the parking lot according to the second image data.

[0032] It is understandable that in the present application, cameras are installed at various locations in the parking lot, and the cameras will capture images of the parking lot at a preset frequency.

[0033] Step S12: Detect vehicle status changes in the passable area based on the first image data and the second image data to obtain at least one changed parking space.

[0034] Among them, the changed parking space refers to the parking space that has been changed and occupied.

[0035] It is understandable that in actual applications, the change of vehicle status in a parking space can be divided into three situations: the first is the change from the state of a car in the parking space to the state of no car, the second is the change from the state of no car to the state of a car, and the third involves the replacement of one vehicle with another. This application specifically refers to the change of the latter two states, namely, the change from the state of no car to the state of a car, and the replacement of vehicles in the same parking space.

[0036] Step S13: for each changed parking space, determine whether the vehicle in the changed parking space is an unrecorded vehicle according to the second image data; if so, put the changed parking space into the target changed parking space set.

[0037] Among them, unrecorded vehicles are vehicles that are not parked in the parking lot.

[0038] It should be noted that in this application, all vehicles parked in the parking lot will be recorded, such as the vehicle's license plate, brand, color, shape, etc. Vehicles that have never been in the parking lot will not be recorded, and such vehicles are unrecorded vehicles.

[0039] Step S14, calling the three-dimensional model of the parking lot, based on the minimum spanning tree algorithm, and changing the parking space set according to the target to generate parking space sequence data.

[0040] The parking space sequence data includes a parking space sequence and a distance sequence corresponding to the parking space sequence.

[0041] It should be noted that the minimum spanning tree is designed to connect all vertices (i.e. parking spaces) without forming a loop and to minimize the total weight of the edges. In the scenario of parking space changes, the process of generating parking space sequence data based on the minimum spanning tree (MST) algorithm can be divided into the following steps: First, define the target change parking space set and the parking spaces in the target change parking space set: regard each parking space as a vertex in the graph, use numbers (such as P1, P2, P3, etc.) to represent different parking spaces, and use C={P1, P2, P3, ..., Pn} to represent the target change parking space set; then define the connection relationship between the vertices in the graph, that is, define edges between each pair of parking spaces, and assign corresponding weights to each edge according to the distance between the parking spaces. In this application, the weight of each edge is the distance between the two parking spaces corresponding to the edge; then apply the minimum spanning tree algorithm, and you can choose Prim algorithm or Kr algorithm The Kruskal algorithm is used to find the minimum spanning tree. For the Prim algorithm, it is applicable to dense graphs and gradually expands the minimum spanning tree. For the Kruskal algorithm, it is applicable to sparse graphs, and the edge weights are sequentially connected to form acyclic trees from small to large. Finally, the parking space sequence data is generated, that is, the path connecting the parking spaces is extracted from the minimum spanning tree, that is, the order from one parking space to another is formed, and the data is recorded. The generated parking space sequence data is formatted for subsequent analysis or application. For example, the output format can be: the parking space sequence is P1->P3->P2->P4, and the distance sequence is 10->13->9.

[0042] Step S15, generating a model based on the surveying and mapping scheme, determining surveying and mapping parameters according to the parking space sequence data, and controlling the surveying and mapping vehicle to survey the parking lot according to the surveying and mapping parameters.

[0043] Among them, the surveying parameters include the optimal number of surveying vehicles and the surveying order of the surveying vehicles for the target changed parking spaces.

[0044] In the above implementation, the first image data and the second image data collected in the parking lot are first obtained, and the passable area of ​​the parking lot is determined according to the second image data. Then, the vehicle state change detection is performed on the passable area according to the first image data and the second image data to obtain at least one changed parking space, wherein the changed parking space is a changed and occupied parking space. Then, for each changed parking space, it is determined according to the second image data whether the vehicle in the changed parking space is an unrecorded vehicle. If so, the changed parking space is placed in the target changed parking space set, wherein the unrecorded vehicle is a vehicle that is not parked in the parking lot. Then, the three-dimensional model of the parking lot is called. A model is proposed, based on the minimum spanning tree algorithm, and according to the target changed parking space set, a parking space sequence data is generated, wherein the parking space sequence data includes a parking space sequence and a distance sequence corresponding to the parking space sequence. Finally, a model is generated based on a surveying and mapping scheme, and according to the parking space sequence data, surveying and mapping parameters are determined, and according to the surveying and mapping parameters, a surveying and mapping vehicle is controlled to survey the parking lot, wherein the surveying and mapping parameters include the optimal number of surveying and mapping vehicles and the surveying and mapping order of the surveying and mapping vehicle for the target changed parking spaces. Through the above method, the surveying and mapping of vehicles in the parking lot is realized during the use of the parking lot, the intelligent level of the parking lot surveying is improved, and the management and monitoring of vehicles in the parking lot are strengthened.

[0045] As a further embodiment of the surveying and mapping system, the step of determining the passable area of ​​the parking lot according to the second image data includes: Step S21, performing grayscale conversion on the second image data according to the second image data to obtain a corresponding grayscale image.

[0046] It can be understood that grayscale conversion, in the field of computer vision and image processing, is a process of converting a color image into a grayscale image. A grayscale image is an image that contains only brightness information but not color information, and its pixel values ​​usually represent grayscale levels from black (pixel value 0) to white (pixel value 255) (in 8-bit images). The basic idea of ​​grayscale conversion is to calculate the corresponding grayscale value based on the color value of each pixel of the color image. For RGB color images, each pixel consists of three color channels: red (R), green (G), and blue (B). The grayscale conversion algorithm comprehensively considers the values ​​of these three color channels and obtains a grayscale value through weighted summation or other methods. A common grayscale conversion method is the weighted average method, which assigns different weights to the three channels of R, G, and B according to the different sensitivities of the human eye to different colors. A commonly used weight assignment is: Gray=0.299×R+0.587×G+0.114×B; This formula is based on psychological experiments and simulates the sensitivity of the human eye to the three primary colors of RGB. The grayscale value calculated by this method can better preserve the brightness information and contrast of the original color image. In addition to the weighted average method, there are other grayscale conversion methods, such as the maximum value method (taking the maximum value of R, G, and B as the grayscale value), the minimum value method (taking the minimum value of R, G, and B as the grayscale value), and the average value method (taking the average value of R, G, and B as the grayscale value).

[0047] Step S22, obtaining a pixel matrix corresponding to the grayscale image according to the grayscale image, and comparing elements in the pixel matrix with a preset threshold to obtain first coordinate data and second coordinate data.

[0048] The first coordinate data is coordinate data corresponding to a matrix region whose element values ​​in the pixel matrix are within a preset range, and the second coordinate data is coordinate data corresponding to a matrix region whose element values ​​in the pixel matrix are not within the preset range.

[0049] Step S23: The area corresponding to the first coordinate data is regarded as a passable area, and the area corresponding to the second coordinate data is regarded as a prohibited passable area.

[0050] It is understandable that the prohibited area may be a construction area or an area where obstacles exist.

[0051] In the above embodiment, in order to determine the passable area of ​​the parking lot, the second image data is converted into grayscale according to the second image data to obtain a corresponding grayscale image, and then a pixel matrix corresponding to the grayscale image is obtained according to the grayscale image, and the elements in the pixel matrix are compared with a preset threshold to obtain first coordinate data and second coordinate data, wherein the first coordinate data is the coordinate data corresponding to the matrix area in which the element values ​​in the pixel matrix are within a preset range, and the second coordinate data is the coordinate data corresponding to the matrix area in which the element values ​​in the pixel matrix are not within the preset range, and finally the area corresponding to the first coordinate data is used as a passable area, and the area corresponding to the second coordinate data is used as a prohibited area.

[0052] As a further implementation of the surveying and mapping system, the vehicle status change detection is performed on the passable area according to the first image data and the second image data to obtain the step of changing the parking space, including: Step S31, extracting features in the passable area according to the first image data to obtain a first vehicle state feature.

[0053] Step S32: extracting features in the passable area according to the second image data to obtain second vehicle state features.

[0054] It is understandable that in image processing and computer vision, feature extraction within the passable area usually includes the following steps: preprocessing, that is, converting the image into a format suitable for analysis, such as grayscale, normalization and noise removal, and edge detection of the image (such as using the Canny algorithm) to highlight important structures; region segmentation, that is, using thresholding, region growing or deep learning methods (such as semantic segmentation) to segment the image into different regions to identify passable areas; feature extraction, such as color features, shape features, and texture features; feature selection and fusion, usually through algorithms (such as PCA or LDA) to select the most representative features to reduce dimensions and redundant information, and fuse features from different sources (such as color, shape, texture, etc.) to build a more comprehensive feature set; generate vehicle state features, that is, after completing feature extraction, combine different features and extract features reflecting the vehicle state, such as position, speed, direction, etc., through machine learning models or rules.

[0055] Step S33, comparing the first vehicle state feature with the second vehicle state feature to determine a target parking space to be changed within the passable area.

[0056] In the above embodiment, in order to obtain the changed parking space, feature extraction is performed on the passable area according to the first image data to obtain the first vehicle state feature, and then feature extraction is performed on the passable area according to the second image data to obtain the second vehicle state feature. Finally, the first vehicle state feature and the second vehicle state feature are compared to determine the target changed parking space in the passable area.

[0057] As a further implementation of the surveying and mapping system, for each changed parking space, the step of determining whether the vehicle in the changed parking space is an unrecorded vehicle according to the second image data includes: Step S41: for each changed parking space, the license plate clarity of the changed parking space is recognized according to the second image data to obtain the license plate clarity of the changed parking space.

[0058] Step S42, judging whether the license plate clarity is greater than a preset value according to the license plate clarity, if so, obtaining the license plate number according to the second image data, and matching the license plate number in the license plate number data table, if no match is found, it indicates that the vehicle in the changed parking space is an unrecorded vehicle.

[0059] The license plate number data table includes license plate number data and vehicle characteristic data, and the license plate number data table is used to represent the mapping relationship between the license plate number data and the vehicle characteristic data.

[0060] Specifically, it is determined whether the license plate clarity is greater than a preset value based on the license plate clarity. If the license plate clarity is greater than the preset value, the license plate number is obtained based on the second image data, and the license plate number is matched in the license plate number data table. If no match is found, it indicates that the vehicle in the changed parking space is an unrecorded vehicle. If a match is found, it indicates that the vehicle in the changed parking space is not an unrecorded vehicle.

[0061] Step S43, if not, then feature recognition is performed on the vehicle in the changed parking space according to the second image data to obtain vehicle features, and the vehicle is matched in the license plate number data table. If the match is unsuccessful, it indicates that the vehicle in the changed parking space is an unrecorded vehicle.

[0062] Specifically, if the clarity of the license plate is not greater than a preset value, the feature recognition of the vehicle in the changed parking space is performed based on the second image data to obtain the vehicle features, and the vehicle is matched in the license plate number data table. If the match is unsuccessful, it means that the vehicle in the changed parking space is an unrecorded vehicle. If the match is successful, it means that the vehicle in the changed parking space is not an unrecorded vehicle.

[0063] In the above embodiment, in order to determine whether the vehicle in the changed parking space is an unrecorded vehicle, for each changed parking space, the license plate clarity recognition is performed on the changed parking space according to the second image data to obtain the license plate clarity of the changed parking space, and then it is determined whether the license plate clarity is greater than a preset value according to the license plate clarity. If the license plate clarity is greater than the preset value, the license plate number is obtained according to the second image data, and the license plate number is matched in the license plate number data table. If no match is found, it indicates that the vehicle in the changed parking space is an unrecorded vehicle, wherein the license plate number data table includes license plate number data and vehicle feature data, and the license plate number data table is used to indicate the mapping relationship between the license plate number data and the vehicle feature data. If the license plate clarity is not greater than the preset value, the feature recognition is performed on the vehicle in the changed parking space according to the second image data to obtain the vehicle feature, and the vehicle is matched in the license plate number data table. If the match is not successful, it indicates that the vehicle in the changed parking space is an unrecorded vehicle.

[0064] As a further implementation method of the surveying and mapping system, the steps of calling the three-dimensional model of the parking lot, based on the minimum spanning tree algorithm, and changing the parking space set according to the target to generate the parking space sequence data include: Step S51, calling the three-dimensional model of the parking lot, and determining the starting parking space for mapping according to the three-dimensional model, wherein the starting parking space for mapping is the parking space in the target change parking space set that is closest to the storage point of the mapping vehicle; Step S52, starting from the initial surveyed parking space, based on the minimum spanning tree algorithm and according to the target change parking space set, determines the parking space sequence data.

[0065] In the above embodiment, in order to generate parking space sequence data, the three-dimensional model of the parking lot is called, and the starting surveying parking space is determined based on the three-dimensional model, wherein the starting surveying parking space is the parking space closest to the storage point of the surveying vehicle in the target change parking space set, and then the parking space sequence data is determined based on the minimum spanning tree algorithm and the target change parking space set with the starting surveying parking space as the starting point.

[0066] As a further implementation method of the surveying and mapping system, the objective function of the surveying and mapping scheme generation model is:

[0067] The constraints of the objective function are:

[0068] in, represents the sum of the kth segment, represents the jth element in the distance sequence, n represents the number of parts into which the distance sequence is divided, d represents the distance sequence, and X is the maximum value of each part. Indicates the starting position of the first paragraph. Indicates the end position of the nth segment, m indicates the total number of elements in the distance sequence, Indicates the starting position of the kth segment, Indicates the starting position of the k+1th segment.

[0069] It should be noted that for the objective function ( ), which means that this value is made as small as possible through the optimization process. In other words, it is hoped that the distance sequence can be divided into as few parts as possible; represents the sum of the kth segment, specifically the kth segment in the distance sequence Positions to The sum of the positions; represents the jth element in the distance sequence, Indicates that in the distance sequence, the index is from arrive Elements Sum, that is, calculate the total distance of the kth segment; X is a given maximum value, indicating that the sum of each segment cannot exceed this value; ∀k=1,2,…,n means that these constraints apply to all segments, that is, from the first segment to the nth segment; for the start and end positions of the segment (that is, ), Indicates the starting position of the first segment, which is fixed to 1, meaning it starts from the beginning of the sequence. Indicates the end position of the nth segment, which is equal to m+1, indicating that after the sequence is divided, there is one more position, which symbolically ends to facilitate the processing of the last segment; for position constraints (i.e. ), Indicates the starting position of the kth segment, which needs to be within the distance sequence. Indicates that each segment starts at Cannot be less than 1 and less than the starting position of the next segment ,and It cannot exceed m (m represents the total length of the distance sequence or the total number of elements). This position constraint on k ensures that the division of each segment in the distance sequence is valid and orderly. This constraint applies to every k from 1 to n−1, i.e., the starting position of the first n−1 segments.

[0070] It should be further explained that the core of this part of the mathematical model is to define an optimization problem by integrating the above content. Its goal is to divide a distance sequence into several segments so that the sum of each segment is less than or equal to a specific value X, and the number of segments n is expected to be as small as possible. Through the definition of constraints, it is ensured that the position and sum of each segment meet the requirements.

[0071] As a further implementation of the mapping system, the steps of generating a model based on the mapping scheme and determining mapping parameters according to the parking space sequence data include: Step S61, generating a model through a surveying and mapping scheme, and determining the optimal number of surveying and mapping vehicles according to the distance sequence.

[0072] Step S62, through the constraints of the objective function and according to the optimal number, parking space sequence and distance sequence, determine the surveying order of the surveying vehicle for the target changed parking space.

[0073] In the above implementation, in order to determine the surveying parameters, a model is generated through a surveying plan, and the optimal number of surveying vehicles is determined based on the distance sequence. Then, through the constraints of the objective function and based on the optimal number, parking space sequence and distance sequence, the surveying order of the surveying vehicles for the target changed parking spaces is determined.

[0074] The embodiment of the present application also discloses an indoor parking lot mapping system based on artificial intelligence.

[0075] refer to Figure 2 , an indoor parking lot mapping system based on artificial intelligence, including: A passable area determination module, used to obtain the first image data and the second image data collected successively in the parking lot, and determine the passable area of ​​the parking lot according to the second image data; a changed parking space generation module, configured to detect a vehicle state change in the passable area according to the first image data and the second image data, and obtain at least one changed parking space, wherein the changed parking space is a parking space that has been changed and is occupied; a target changed parking space set generation module, for each changed parking space, determining whether the vehicle in the changed parking space is an unrecorded vehicle according to the second image data, and if so, putting the changed parking space into the target changed parking space set, wherein the unrecorded vehicle is a vehicle that is not parked in the parking lot; A parking space sequence data generation module is used to call the three-dimensional model of the parking lot, based on the minimum spanning tree algorithm, and generate parking space sequence data according to the target change parking space set, wherein the parking space sequence data includes a parking space sequence and a distance sequence corresponding to the parking space sequence; The surveying and mapping module generates a model based on the surveying and mapping plan, determines the surveying and mapping parameters according to the parking space sequence data, and controls the surveying and mapping vehicle to survey the parking lot according to the surveying and mapping parameters. The surveying and mapping parameters include the optimal number of surveying and mapping vehicles and the surveying and mapping order of the target changed parking spaces by the surveying and mapping vehicles.

[0076] The artificial intelligence-based indoor parking lot mapping system of the present invention can implement any one of the artificial intelligence-based indoor parking lot mapping methods, and the specific working process of the artificial intelligence-based indoor parking lot mapping system of the present invention can refer to the corresponding process in the above-mentioned artificial intelligence-based indoor parking lot mapping method.

[0077] The embodiment of the present application also discloses a computer device.

[0078] refer to Figure 3 A computer device includes a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the computer program, any one of the above-mentioned artificial intelligence-based indoor parking lot mapping methods is implemented.

[0079] The embodiment of the present application also discloses a computer-readable storage medium.

[0080] A computer-readable storage medium stores a computer program that can be loaded by a processor and execute any of the above-mentioned artificial intelligence-based indoor parking lot mapping methods.

[0081] Among them, computer-readable storage media can be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, apparatus or device; the program code contained on the computer-readable medium can be transmitted using any appropriate medium, including but not limited to wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0082] The above are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Any feature disclosed in this specification (including the abstract and drawings), unless otherwise stated, can be replaced by other equivalent or alternative features with similar purposes. That is, unless otherwise stated, each feature is only an example of a series of equivalent or similar features.

Claims

1. An indoor parking lot mapping method based on artificial intelligence, characterized in that: include: Acquire first image data and second image data collected successively in the parking lot, and determine a passable area of ​​the parking lot according to the second image data; Performing vehicle status change detection on the passable area according to the first image data and the second image data to obtain at least one changed parking space, wherein the changed parking space is a changed and occupied parking space; For each changed parking space, determining whether the vehicle in the changed parking space is an unrecorded vehicle according to the second image data, and if so, placing the changed parking space into a target changed parking space set, wherein the unrecorded vehicle is a vehicle that is not parked in the parking lot; Calling the three-dimensional model of the parking lot, based on the minimum spanning tree algorithm, and changing the parking space set according to the target to generate parking space sequence data, wherein the parking space sequence data includes a parking space sequence and a distance sequence corresponding to the parking space sequence; A model is generated based on a surveying and mapping scheme, and surveying and mapping parameters are determined according to the parking space sequence data. A surveying and mapping vehicle is controlled to survey the parking lot according to the surveying and mapping parameters, wherein the surveying and mapping parameters include an optimal number of the surveying and mapping vehicles and a surveying and mapping order of the target changed parking spaces by the surveying and mapping vehicles.

2. The indoor parking lot mapping method based on artificial intelligence according to claim 1 is characterized in that: The step of determining the passable area of ​​the parking lot according to the second image data comprises: Performing grayscale conversion on the second image data according to the second image data to obtain a corresponding grayscale image; A pixel matrix corresponding to the grayscale image is obtained according to the grayscale image, and elements in the pixel matrix are compared with a preset threshold value to obtain first coordinate data and second coordinate data, wherein the first coordinate data is coordinate data corresponding to a matrix region in the pixel matrix whose element values ​​are within a preset range, and the second coordinate data is coordinate data corresponding to a matrix region in the pixel matrix whose element values ​​are not within the preset range, The area corresponding to the first coordinate data is regarded as a passable area, and the area corresponding to the second coordinate data is regarded as a prohibited passable area.

3. The indoor parking lot mapping method based on artificial intelligence according to claim 1 is characterized in that: The step of detecting a vehicle status change in the passable area according to the first image data and the second image data to obtain a changed parking space comprises: Extracting features in the passable area according to the first image data to obtain a first vehicle state feature; Extracting features in the passable area according to the second image data to obtain a second vehicle state feature; A target parking space to be changed in the passable area is determined by comparing the first vehicle state feature with the second vehicle state feature.

4. The indoor parking lot mapping method based on artificial intelligence according to claim 1 is characterized in that: The step of determining, for each changed parking space, whether the vehicle in the changed parking space is an unrecorded vehicle according to the second image data comprises: For each changed parking space, performing license plate clarity recognition on the changed parking space according to the second image data to obtain the license plate clarity of the changed parking space; Determine whether the license plate clarity is greater than a preset value according to the license plate clarity; if so, obtain the license plate number according to the second image data, and match the license plate number in a license plate number data table; if no match is found, it indicates that the vehicle in the changed parking space is an unrecorded vehicle, wherein the license plate number data table includes license plate number data and vehicle feature data, and the license plate number data table is used to indicate a mapping relationship between the license plate number data and the vehicle feature data; If not, feature recognition is performed on the vehicle in the changed parking space based on the second image data to obtain vehicle features, and the vehicle is matched in the license plate number data table. If the match is unsuccessful, it indicates that the vehicle in the changed parking space is an unrecorded vehicle.

5. The indoor parking lot mapping method based on artificial intelligence according to claim 1 is characterized in that: The step of calling the three-dimensional model of the parking lot, based on the minimum spanning tree algorithm, and changing the parking space set according to the target to generate parking space sequence data includes: Calling the three-dimensional model of the parking lot, and determining a starting parking space for mapping according to the three-dimensional model, wherein the starting parking space for mapping is the parking space in the target change parking space set that is closest to the storage point of the mapping vehicle; Taking the initial surveyed parking space as the starting point, based on the minimum spanning tree algorithm and according to the target change parking space set, the parking space sequence data is determined.

6. The indoor parking lot mapping method based on artificial intelligence according to claim 1 is characterized in that: The objective function of the surveying and mapping scheme generation model is: The constraints of the objective function are: in, represents the sum of the kth segment, represents the jth element in the distance sequence, n represents the number of parts into which the distance sequence is divided, d represents the distance sequence, X is the maximum value of each part, Indicates the starting position of the first paragraph. represents the end position of the nth segment, m represents the total number of elements in the distance sequence, Indicates the starting position of the kth segment, Indicates the starting position of the k+1th segment.

7. The indoor parking lot mapping method based on artificial intelligence according to claim 6 is characterized in that: The step of generating a model based on the surveying and mapping scheme and determining surveying and mapping parameters according to the parking space sequence data includes: Generate a model through a surveying and mapping scheme, and determine the optimal number of surveying and mapping vehicles according to the distance sequence; The order in which the surveying and mapping vehicle surveys the target changed parking spaces is determined by the constraint conditions of the objective function and according to the optimal number, the parking space sequence and the distance sequence.

8. The indoor parking lot mapping system based on artificial intelligence is characterized by: include: A passable area determination module, used to obtain first image data and second image data collected successively in the parking lot, and determine the passable area of ​​the parking lot according to the second image data; a changed parking space generation module, configured to detect a vehicle state change in the passable area according to the first image data and the second image data, and obtain at least one changed parking space, wherein the changed parking space is a changed and occupied parking space; a target changed parking space set generation module, for each changed parking space, determining whether the vehicle in the changed parking space is an unrecorded vehicle according to the second image data, and if so, placing the changed parking space into a target changed parking space set, wherein the unrecorded vehicle is a vehicle that is not parked in the parking lot; A parking space sequence data generating module, configured to call the three-dimensional model of the parking lot, based on a minimum spanning tree algorithm, and generate parking space sequence data according to the target changing parking space set, wherein the parking space sequence data includes a parking space sequence and a distance sequence corresponding to the parking space sequence; The surveying and mapping module generates a model based on the surveying and mapping scheme, determines surveying and mapping parameters according to the parking space sequence data, and controls the surveying and mapping vehicle to survey the parking lot according to the surveying and mapping parameters, wherein the surveying and mapping parameters include the optimal number of the surveying and mapping vehicles and the surveying and mapping order of the target changed parking spaces by the surveying and mapping vehicles.

9. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program executable on the processor, and the processor implements the method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that: A computer program is stored which can be loaded by a processor and execute the method according to any one of claims 1 to 7.

Images