Intelligent parking management method and system based on license plate recognition

Abstract

The invention relates to the technical field of intelligent parking management, and discloses an intelligent parking management method and system based on license plate recognition, and the method comprises the steps: carrying out the parking lot parking space recognition of a preprocessed parking lot panoramic image; performing license plate recognition on different parking space areas of the parking lot to obtain unoccupied parking space distribution information; calculating to obtain the optimal parking space of the currently driven-in vehicle; a vehicle driving-in route between the driving-in vehicle and the optimal parking space is planned, and the driving-in vehicle drives to the optimal parking space according to the vehicle driving-in route to be parked. The linear functions of the edge points in the panoramic image of the parking lot at different parking space included angles are accumulated and counted to obtain the positions of parking spaces in the parking lot, license plate recognition is performed on different parking space areas to obtain the distribution of unoccupied parking spaces in the parking lot, quantitative evaluation is performed on different unoccupied parking spaces in combination with multi-dimensional quantitative indexes, and the optimal parking space is selected. And route planning of the optimal parking space and the current position of the vehicle is carried out, and the management efficiency of the parking lot is improved.

Description

Technical Field

[0001] The present invention relates to the field of smart parking management, and in particular to a smart parking management method and system based on license plate recognition. Background Art

[0002] With the acceleration of urbanization, the number of cars has increased year by year, and traffic congestion and parking difficulties have become increasingly prominent. According to relevant statistics, the parking demand in large and medium-sized cities has far exceeded the number of available parking spaces. In some busy commercial and residential areas, the scarcity of parking space resources has led to disorderly parking of vehicles, affecting traffic order and even causing social conflicts. At the same time, the inefficient parking management not only affects the parking experience of car owners, but also brings operational difficulties to parking lot managers. In this context, how to effectively utilize limited parking resources and improve parking management efficiency has become an urgent problem to be solved. License plate recognition, as an intelligent traffic management technology, has been widely used in smart city construction in recent years. By combining computer vision, image processing and artificial intelligence technologies, LPR can automatically recognize the license plate of the vehicle, thereby realizing automatic management of the vehicle. In parking lot management, the intelligent parking management method based on license plate recognition can not only greatly improve the efficiency of parking management, but also simplify the parking process of car owners, improve user experience and reduce manual intervention. However, the current license plate recognition technology can only recognize the license plate, and cannot plan and manage parking, causing chaos in the parking lot. Summary of the invention

[0003] In view of this, the present invention provides a smart parking management method based on license plate recognition, which reasonably arranges parking spaces and improves parking lot management efficiency through intelligent analysis of vehicle destinations and parking space distribution in parking lots.

[0004] To achieve the above object, the present invention provides a smart parking management method based on license plate recognition, comprising the following steps:

[0005] S1: Collecting and preprocessing the panoramic image of the parking lot, and identifying parking spaces on the preprocessed panoramic image of the parking lot;

[0006] S2: Perform license plate recognition on different parking spaces in the parking lot, mark the parking spaces with recognized license plates as non-vacant parking spaces, and mark other parking spaces as vacant parking spaces, to obtain vacant parking space distribution information;

[0007] S3: Calculate the optimal parking space by combining the current position of the incoming vehicle with the distribution information of the available parking spaces, wherein multi-dimensional quantitative evaluation is an implementation method for calculating the optimal parking space;

[0008] S4: planning a vehicle entry route between the entering vehicle and the optimal parking space, and sending the vehicle entry route to the entering vehicle by wireless communication, so that the entering vehicle drives to the optimal parking space according to the vehicle entry route and parks.

[0009] As a further improvement method of the present invention:

[0010] Optionally, in step S1, collecting the panoramic image of the parking lot and preprocessing it includes:

[0011] The parking lot panoramic image I is collected by overhead shooting, and the parking lot panoramic image I is preprocessed, wherein the preprocessing process is as follows:

[0012] Convert the parking lot panoramic image I into pixel matrix form:

[0013] I=(I(p,q)) P×Q

[0014] I(p,q)=max{I R (p,q),I G (p,q),I B (p,q)}

[0015] in:

[0016] (I(p,q)) P×Q represents a matrix form of P rows and Q columns, where P represents the number of pixel rows of the parking lot panoramic image I, Q represents the number of pixel columns of the parking lot panoramic image I, and I(p,q) represents the pixel value of the pixel (p,q) in the p-th row and q-th column of the parking lot panoramic image I, p∈[1,P],q∈[1,Q];

[0017] I R (p,q),I G (p,q),I B (p, q) represents the color value of pixel (p, q) in R, G, and B color channels respectively;

[0018] max{I R (p,q),I G (p,q),I B (p,q)} means to select I R (p,q),I G (p,q),I B The maximum value among (p,q);

[0019] The pixels in the parking lot panoramic image I are smoothed and filtered, where the smoothing filter result of the pixel (p, q) is g(p, q):

[0020]

[0021] in:

[0022] σ represents a scale factor; in the embodiment of the present invention, σ is set to 2;

[0023] exp(·) represents an exponential function with a natural constant as the base;

[0024] Based on the smoothing filtering results of the pixels, a preprocessed parking lot panoramic image g = (g(p,q)) is constructed. P×Q , parking lot parking spaces are recognized on the preprocessed parking lot panoramic image g.

[0025] Optionally, the performing parking lot parking space recognition on the pre-processed parking lot panoramic image g includes:

[0026] S11: Calculate the gradient value of any pixel in the parking lot panoramic image g, where the gradient value of pixel (p, q) is grad(p, q):

[0027]

[0028] grad1(p,q)=g(p+1,q)-g(p-1,q)

[0029] grad2(p,q)=g(p,q+1)-g(p,q-1)

[0030] in:

[0031] grad1(p,q) represents the gradient value of pixel (p,q) in the horizontal direction, and grad2(p,q) represents the gradient value of pixel (p,q) in the vertical direction;

[0032] S12: Calculate the neighborhood edge value of any pixel in the parking lot panoramic image g, where the neighborhood edge value of pixel (p, q) is Neighbor (p, q):

[0033] Neighbor(p,q)=(max{Ω(p,q)}>0)Λ(min{Ω(p,q)}<0)

[0034] in:

[0035] Λ represents the logical AND operation. If (max{Ω(p,q)}>0) and (min{Ω(p,q)}<0) are both 1, then (max{Ω(p,q)}>0)Λ(min{Ω(p,q)}<0) is 1;

[0036] Ω(p,q) represents the pixel gradient value set of pixel (p,q) within the 3×3 neighborhood, max{Ω(p,q) represents the maximum value in the pixel gradient value set Ω(p,q), and min{Ω(p,q)} represents the minimum value in the pixel gradient value set Ω(p,q);

[0037] If max{Ω(p,q)}>0, then (max{Ω(p,q)}>0)=1, otherwise (max{Ω(p,q)}>0)=0; if min{Ω(p,q)}<0, then (min{Ω(p,q)}<0)=1, otherwise (min{Ω(p,q)}<0)=0;

[0038] S13: If the neighborhood edge value of the pixel is greater than 0, it means that there is an edge point in the 3×3 neighborhood pixel range of the pixel, and the pixel with the largest gradient value is traversed in the 3×3 neighborhood pixel range of the pixel as the edge pixel;

[0039] S14: Construct a coordinate system with the center of the parking lot panoramic image g as the coordinate origin, transform the pixel coordinates of the edge pixels into the coordinate system, and obtain the coordinate transformation result of the edge pixels, where the coordinate transformation result of the sth edge pixel is s∈[1,S], S represents the total number of edge pixels, Represents the horizontal coordinate of the sth edge pixel in the coordinate system, Represents the vertical coordinate of the sth edge pixel in the coordinate system;

[0040] S15: Set the parking space angles to θ1, θ2, θ3, where θ1, θ2, θ3 are 0 respectively. π; Calculate the cumulative value of all edge pixels at different parking angles and the number of times the cumulative value appears, where the sth edge pixel at the parking angle θ i The cumulative value is i∈[1,3]:

[0041]

[0042] The accumulated values ​​and the corresponding parking space angles whose occurrence times are higher than a preset threshold are retained, and the retained accumulated values ​​and parking space angles are used as a set of parking space lines, wherein the parking space line is a polar coordinate representation, the accumulated value is a vertical distance from the parking space line to the origin of the coordinates, and the parking space angle is an angle between the parking space line and the horizontal axis;

[0043] S16: If the parking space angles between adjacent parking space lines are perpendicular to each other, and the distance between the centers of the parking space lines is less than a preset threshold,

[0044] Then the adjacent parking space lines are regarded as a group of parking spaces, otherwise the parking space lines are regarded as a group of parking spaces;

[0045] S17: Repeat step S16 to fuse the parking space lines to obtain N parking spaces in the parking lot, where N represents the total number of parking spaces after the fusion process, and convert the coordinate positions of the parking spaces to the parking lot area to obtain N parking space areas in the parking lot. In the embodiment of the present invention, the coordinate conversion method of the pixel coordinate system-camera coordinate system-world coordinate system is an implementation method for converting the coordinate positions of the parking spaces to the parking lot area.

[0046] Optionally, in step S2, license plate recognition is performed on different parking spaces in the parking lot, and parking spaces with recognized license plates are marked as non-vacant parking spaces, and other parking spaces are marked as vacant parking spaces, including:

[0047] The parking space images of different parking spaces in the parking lot are collected, where the parking space image of the nth parking space area is I n , for parking space image I n Carry out license plate recognition, where the license plate recognition process is:

[0048] S21: Parking space image I n Performing grayscale processing, and calculating the gradient value of the pixel based on the grayscale processing result;

[0049] S22: Combine the gradient values ​​of pixels and use the Canny edge detection algorithm to detect the edge of the parking space image I n Detect edge pixels in the

[0050] S23: using dilation and erosion methods in sequence to perform morphological processing on edge pixels, connecting edge pixels to form an edge closed area;

[0051] S24: Calculate the size information of the edge closed area, and select the edge closed area whose size information meets the size threshold as the license plate candidate area;

[0052] S25: extracting HOG features and SIFT features of the license plate candidate area, and using the support vector machine model to detect whether there is a license plate in the license plate candidate area. If a license plate is detected, it means that the license plate is recognized in the parking space of the nth parking space area, otherwise it means that the license plate is not recognized;

[0053] The parking spaces with recognized license plates are marked as non-vacant parking spaces, and the other parking spaces are marked as vacant parking spaces, and the vacant parking space distribution information L of the parking lot is obtained:

[0054] L={L m |m∈[1,M]}

[0055] in:

[0056] L m It represents the coordinate position of the parking area where the mth vacant parking space is located, and M represents the total number of vacant parking spaces in the parking lot.

[0057] Optionally, the step S3 combines the current position of the incoming vehicle with the vacant parking space distribution information to calculate the optimal parking space, including:

[0058] The optimal parking space is calculated by combining the current position Loc of the incoming vehicle and the vacant parking space distribution information L, where the current position Loc is the entrance position of the incoming vehicle into the parking lot, and the multi-dimensional quantitative evaluation is the implementation method for calculating the optimal parking space. The calculation process of the optimal parking space is as follows:

[0059] S31: constructing a multi-dimensional quantitative evaluation index, wherein the multi-dimensional quantitative evaluation index includes a parking space distance index, a parking space redundancy index, a parking space safety index, and an access convenience index;

[0060] S32: Obtain the index information of different vacant parking spaces according to the multi-dimensional quantitative evaluation index, where the index information of the mth vacant parking space is represents the index values ​​of the parking distance index, parking redundancy index, parking safety index and access convenience index of the mth vacant parking space in sequence;

[0061] S33: Standardize the index information of the vacant parking spaces, where the index information F m The normalized result is In order The normalized processing result is:

[0062]

[0063]

[0064]

[0065]

[0066] S34: Calculate the weights of different indicators to form an indicator weight vector w=[w1,w2,w3,w4], where w1,w2,w3,w4 are the weights of the parking space distance indicator, the parking space redundancy indicator, the parking space safety indicator, and the access convenience indicator respectively;

[0067]

[0068] in:

[0069] w j is the weight of the jth indicator, j∈[1,4], the first to fourth indicators are parking distance indicator, parking redundancy indicator, parking safety indicator and access convenience indicator;

[0070] μ jrepresents the mean of the M groups of normalized index values ​​associated with the jth index, σ j It represents the standard deviation of the M groups of normalized index values ​​associated with the jth index;

[0071] S35: Use the index weight vector w to weight the normalized index information to obtain weighted index information of vacant parking spaces, where the index information f m The corresponding weighted index information is

[0072] S36: Based on the weighted index information of all vacant parking spaces, construct the ideal weighted index information f + And the negative ideal weighted index information f - :

[0073]

[0074] S37: Combine ideal weighted index information f + And the negative ideal weighted index information f - The support of each vacant parking space is calculated, and the vacant parking space with the highest support is selected as the optimal parking space.

[0075] Optionally, in step S32, obtaining index information of different vacant parking spaces according to multi-dimensional quantitative evaluation indexes includes:

[0076] S321: Calculate the Euclidean distance between the parking area coordinate position of the vacant parking space and the current position Loc as the index value of the parking distance index of the vacant parking space;

[0077] S322: Calculate the number of vacant parking spaces on both sides of the vacant parking space and the ratio of vacant parking spaces in the horizontal direction to form the index value of the vacant parking space in the parking space redundancy index, where the index value of the mth vacant parking space in the parking space redundancy index is

[0078]

[0079] in:

[0080] Num m Indicates the number of vacant parking spaces on both sides of the mth vacant parking space;

[0081] P m Indicates the ratio of vacant parking spaces in the horizontal direction of the mth vacant parking space;

[0082] S323: Obtain the lighting brightness of the coordinate position of the parking area where the vacant parking space is located as the index value of the parking safety index of the vacant parking space;

[0083] S324: Calculate the Euclidean distance between the coordinate position of the parking space area where the vacant parking space is located and the elevator entrance of the parking lot, and use it as the index value of the vacant parking space in the access convenience index.

[0084] Optionally, in the step S4, planning the vehicle driving route between the incoming vehicle and the optimal parking space includes:

[0085] Plan the vehicle driving route between the incoming vehicle and the optimal parking space, where the planning process of the vehicle driving route is as follows:

[0086] S41: Initialize and generate a traversal node list list, and add the position of the optimal parking space to list;

[0087] S42: Represent the parking lot in terms of position nodes, where each position node is the inflection point of the road in the parking lot;

[0088] S43: Extract the latest added position from list as the parent node, and traverse the position nodes node that are adjacent to the parent node and not in list, calculate the path distance between the position node node and the parent node, and use it as the path cost Cost1 of the position node node;

[0089] S44: Starting from the position node node, traverse the jump point that is not in list and is the closest to the current position Loc among the position nodes adjacent to the starting point, calculate the path distance Cost2 between the jump point and the optimal parking space and the path distance Cost3 between the jump point and the parent node. If Cost2 < min{Cost1, Cost3}, then use the position node node as the parent node and turn to step S45, otherwise return to step S43 to re-traverse the position nodes; where the jump point is a position node adjacent to the position node node or a position node with an obstacle between it and the position node node;

[0090] S45: Add the parent node to list, return to step S43, and continue until the current position Loc is added to list;

[0091] S46: Extract the position node sequence from the traversal node list list from the current position Loc to the optimal parking space as the vehicle driving route;

[0092] Send the vehicle driving route to the incoming vehicle by wireless communication, and the incoming vehicle drives to the optimal parking space according to the vehicle driving route for parking.

[0093] Optionally, in the step S37, combining the ideal weighted index information f + and the negative ideal weighted index information f - Calculate the support degree of each vacant parking space, including:

[0094] The support of the mth vacant parking space is:

[0095]

[0096] in:

[0097] Support m represents the support of the mth vacant parking space.

[0098] In order to solve the above problems, the present invention provides a smart parking management system based on license plate recognition, the system comprising:

[0099] The parking space recognition module is used to collect and pre-process the panoramic image of the parking lot, and recognize the parking spaces in the parking lot based on the pre-processed panoramic image of the parking lot;

[0100] The parking space management and allocation module is used to perform license plate recognition on different parking space areas of the parking lot, mark the parking spaces with recognized license plates as non-vacant parking spaces, and mark other parking spaces as vacant parking spaces, obtain the vacant parking space distribution information, and calculate the optimal parking space by combining the current position of the incoming vehicle with the vacant parking space distribution information;

[0101] The route planning device is used to plan a vehicle entry route between the entering vehicle and the optimal parking space, and transmit the vehicle entry route to the entering vehicle by wireless communication, so that the entering vehicle drives to the optimal parking space according to the vehicle entry route and parks.

[0102] In order to solve the above problem, the present invention further provides an electronic device, the electronic device comprising:

[0103] A memory storing at least one instruction;

[0104] Communication interface, enabling electronic equipment to communicate; and

[0105] The processor executes the instructions stored in the memory to implement the above-mentioned smart parking management method based on license plate recognition.

[0106] In order to solve the above problems, the present invention also provides a computer-readable storage medium, in which at least one instruction is stored. The at least one instruction is executed by a processor in an electronic device to implement the above-mentioned smart parking management method based on license plate recognition.

[0107] Compared with the prior art, the present invention proposes a smart parking management method based on license plate recognition, which has the following advantages:

[0108] First, this scheme proposes a parking space recognition method for a parking lot. It obtains a panoramic image of the parking lot and performs pixel gradient calculation. Combined with the sign change of the pixel gradient, the neighborhood pixel area where the edge point exists is determined to improve the retrieval efficiency of the edge point. The straight line function of the edge point in the image at different parking space angles is generated and accumulated and counted. The straight line function with a higher number of occurrences is selected as the parking space line in the panoramic image of the parking lot, and the parking space line is aggregated to obtain the parking space in the parking lot. The real-time calculation of the distribution information of the vacant parking spaces in the parking lot is realized, and the positioning efficiency of the vacant parking spaces is improved.

[0109] At the same time, this plan proposes a vacant parking space management method, which combines multi-dimensional quantitative indicators to conduct quantitative evaluation of different vacant parking spaces, selects a comprehensive evaluation of the distance from the current position of the vehicle, the degree of redundancy next to the vacant parking spaces, the safety of the vacant parking spaces, and the convenience of entry and exit, and selects the vacant position that is most suitable for the incoming vehicle as the optimal parking space, and plans the route between the optimal parking space and the current position of the vehicle to improve the parking efficiency of the parking lot. BRIEF DESCRIPTION OF THE DRAWINGS

[0110] Figure 1 A flowchart of a smart parking management method based on license plate recognition provided by one embodiment of the present invention;

[0111] Figure 2 A functional module diagram of a smart parking management system based on license plate recognition provided by an embodiment of the present invention;

[0112] Figure 2 In: 100 smart parking management system based on license plate recognition, 101 parking space recognition module, 102 parking space management allocation module, 103 route planning device;

[0113] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0114] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0115] The embodiment of the present application provides a smart parking management method based on license plate recognition. The execution subject of the smart parking management method based on license plate recognition includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided by the embodiment of the present application. In other words, the smart parking management method based on license plate recognition can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc.

[0116] Embodiment 1:

[0117] S1: Collect and pre-process a panoramic image of the parking lot, and identify parking spaces on the pre-processed panoramic image of the parking lot.

[0118] In step S1, the panoramic image of the parking lot is collected and preprocessed, including:

[0119] The parking lot panoramic image I is collected by overhead shooting, and the parking lot panoramic image I is preprocessed, wherein the preprocessing process is as follows:

[0120] Convert the parking lot panoramic image I into pixel matrix form:

[0121] I=(I(p,q)) P×Q

[0122] I(p,q)=max{I R (p,q),I G (p,q),I B (p,q)}

[0123] in:

[0124] (I(p,q)) P×Q represents a matrix form of P rows and Q columns, where P represents the number of pixel rows of the parking lot panoramic image I, Q represents the number of pixel columns of the parking lot panoramic image I, and I(p,q) represents the pixel value of the pixel (p,q) in the p-th row and q-th column of the parking lot panoramic image I, p∈[1,P],q∈[1,Q];

[0125] I R (p,q),I G (p,q),I B (p, q) represents the color value of pixel (p, q) in R, G, and B color channels respectively;

[0126] max{I R (p,q),I G (p,q),I B (p,q)} means to select I R (p,q),I G (p,q),I B The maximum value among (p,q);

[0127] The pixels in the parking lot panoramic image I are smoothed and filtered, where the smoothing filter result of the pixel (p, q) is g(p, q):

[0128]

[0129] in:

[0130] σ represents a scale factor; in the embodiment of the present invention, σ is set to 2;

[0131] exp(·) represents an exponential function with a natural constant as the base;

[0132] Based on the smoothing filter results of the pixels, a preprocessed parking lot panoramic image g = (g(p,q)) is constructed. P×Q , parking lot parking spaces are recognized on the preprocessed parking lot panoramic image g.

[0133] The performing parking lot parking space recognition on the pre-processed parking lot panoramic image g comprises:

[0134] S11: Calculate the gradient value of any pixel in the parking lot panoramic image g, where the gradient value of pixel (p, q) is grad(p, q):

[0135]

[0136] grad1(p,q)=g(p+1,q)-g(p-1,q)

[0137] grad2(p,q)=g(p,q+1)-g(p,q-1)

[0138] in:

[0139] grad1(p,q) represents the gradient value of pixel (p,q) in the horizontal direction, and grad2(p,q) represents the gradient value of pixel (p,q) in the vertical direction;

[0140] S12: Calculate the neighborhood edge value of any pixel in the parking lot panoramic image g, where the neighborhood edge value of pixel (p, q) is Neighbor (p, q):

[0141] Neighbor(p,q)=(max{Ω(p,q)}>0)∧(min{Ω(p,q)}<0)

[0142] in:

[0143] ∧ represents the logical AND operation. If (max{Ω(p,q)}>0) and (min{Ω(p,q)}<0) are both 1, then (max{Ω(p,q)}>0)∧(min{Ω(p,q)}<0) is 1;

[0144] Ω(p,q) represents the pixel gradient value set of pixel (p,q) within the 3×3 neighborhood, max{Ω(p,q) represents the maximum value in the pixel gradient value set Ω(p,q), and min{Ω(p,q)} represents the minimum value in the pixel gradient value set Ω(p,q);

[0145] If max{Ω(p,q)}>0, then (max{Ω(p,q)}>0)=1, otherwise (max{Ω(p,q)}>0)=0; if min{Ω(p,q)}<0, then (min{Ω(p,q)}<0)=1, otherwise (min{Ω(p,q)}<0)=0;

[0146] S13: If the neighborhood edge value of the pixel is greater than 0, it means that there is an edge point in the 3×3 neighborhood pixel range of the pixel, and the pixel with the largest gradient value is traversed in the 3×3 neighborhood pixel range of the pixel as the edge pixel;

[0147] S14: Construct a coordinate system with the center of the parking lot panoramic image g as the coordinate origin, transform the pixel coordinates of the edge pixels into the coordinate system, and obtain the coordinate transformation result of the edge pixels, where the coordinate transformation result of the sth edge pixel is s∈[1,S], S represents the total number of edge pixels, Represents the horizontal coordinate of the sth edge pixel in the coordinate system, Represents the vertical coordinate of the sth edge pixel in the coordinate system;

[0148] S15: Set the parking space angles to θ1, θ2, θ3, where θ1, θ2, θ3 are 0 respectively. π; Calculate the cumulative value of all edge pixels at different parking angles and the number of times the cumulative value appears, where the sth edge pixel at the parking angle θ i The cumulative value is i∈[1,3]:

[0149]

[0150] The accumulated values ​​and the corresponding parking space angles whose occurrence times are higher than a preset threshold are retained, and the retained accumulated values ​​and parking space angles are used as a set of parking space lines, wherein the parking space line is a polar coordinate representation, the accumulated value is a vertical distance from the parking space line to the origin of the coordinates, and the parking space angle is an angle between the parking space line and the horizontal axis;

[0151] S16: If the parking space angles between adjacent parking space lines are perpendicular to each other, and the distance between the centers of the parking space lines is less than a preset threshold,

[0152] Then the adjacent parking space lines are regarded as a group of parking spaces, otherwise the parking space lines are regarded as a group of parking spaces;

[0153] S17: Repeat step S16 to fuse the parking space lines to obtain N parking spaces in the parking lot, where N represents the total number of parking spaces after the fusion process, and convert the coordinate positions of the parking spaces to the parking lot area to obtain N parking space areas in the parking lot. In the embodiment of the present invention, the coordinate conversion method of the pixel coordinate system-camera coordinate system-world coordinate system is an implementation method for converting the coordinate positions of the parking spaces to the parking lot area.

[0154] S2: Perform license plate recognition on different parking spaces in the parking lot, mark the parking spaces with recognized license plates as non-vacant parking spaces, and mark other parking spaces as vacant parking spaces, to obtain vacant parking space distribution information.

[0155] In step S2, the license plate recognition is performed on different parking spaces in the parking lot, and the parking spaces with recognized license plates are marked as non-vacant parking spaces, and other parking spaces are marked as vacant parking spaces, including:

[0156] The parking space images of different parking spaces in the parking lot are collected, where the parking space image of the nth parking space area is I n , for parking space image I n Carry out license plate recognition, where the license plate recognition process is:

[0157] S21: Parking space image I n Performing grayscale processing, and calculating the gradient value of the pixel based on the grayscale processing result;

[0158] S22: Combine the gradient values ​​of pixels and use the Canny edge detection algorithm to detect the edge of the parking space image I n The edge image is detected

[0159] S23: using dilation and erosion methods in sequence to perform morphological processing on edge pixels, connecting edge pixels to form an edge closed area;

[0160] S24: Calculate the size information of the edge closed area, and select the edge closed area whose size information meets the size threshold as the license plate candidate area;

[0161] S25: extracting HOG features and SIFT features of the license plate candidate area, and using the support vector machine model to detect whether there is a license plate in the license plate candidate area. If a license plate is detected, it means that the license plate is recognized in the parking space of the nth parking space area, otherwise it means that the license plate is not recognized;

[0162] The parking spaces with recognized license plates are marked as non-vacant parking spaces, and the other parking spaces are marked as vacant parking spaces, and the vacant parking space distribution information L of the parking lot is obtained:

[0163] L={L m |m∈[1,M]}

[0164] in:

[0165] L m It represents the coordinate position of the parking area where the mth vacant parking space is located, and M represents the total number of vacant parking spaces in the parking lot.

[0166] S3: Calculate the optimal parking space by combining the current position of the incoming vehicle with the distribution information of available parking spaces.

[0167] The step S3 calculates the optimal parking space by combining the current position of the incoming vehicle with the distribution information of available parking spaces, including:

[0168] The optimal parking space is calculated by combining the current position Loc of the incoming vehicle and the vacant parking space distribution information L, where the current position Loc is the entrance position of the incoming vehicle into the parking lot, and the multi-dimensional quantitative evaluation is the implementation method for calculating the optimal parking space. The calculation process of the optimal parking space is as follows:

[0169] S31: constructing a multi-dimensional quantitative evaluation index, wherein the multi-dimensional quantitative evaluation index includes a parking space distance index, a parking space redundancy index, a parking space safety index, and an access convenience index;

[0170] S32: Obtain the index information of different vacant parking spaces according to the multi-dimensional quantitative evaluation index, where the index information of the mth vacant parking space is represents the index values ​​of the parking distance index, parking redundancy index, parking safety index and access convenience index of the mth vacant parking space in sequence;

[0171] S33: Standardize the index information of the vacant parking spaces, where the index information F m The normalized result is In order The normalized processing result is:

[0172]

[0173]

[0174]

[0175]

[0176] S34: Calculate the weights of different indicators to form an indicator weight vector w=[w1,w2,w3,w4], where w1,w2,w3,w4 are the weights of the parking space distance indicator, the parking space redundancy indicator, the parking space safety indicator, and the access convenience indicator respectively;

[0177]

[0178] in:

[0179] wj is the weight of the jth indicator, j∈[1,4], the first to fourth indicators are parking distance indicator, parking redundancy indicator, parking safety indicator and access convenience indicator;

[0180] μ j represents the mean of the M groups of normalized index values ​​associated with the jth index, σ j It represents the standard deviation of the M groups of normalized index values ​​associated with the jth index;

[0181] S35: Use the index weight vector w to weight the normalized index information to obtain weighted index information of vacant parking spaces, where the index information f m The corresponding weighted index information is

[0182] S36: Based on the weighted index information of all vacant parking spaces, construct the ideal weighted index information f + And the negative ideal weighted index information f - :

[0183]

[0184]

[0185] S37: Combine ideal weighted index information f + And the negative ideal weighted index information f - The support of each vacant parking space is calculated, and the vacant parking space with the highest support is selected as the optimal parking space.

[0186] In step S32, the index information of different vacant parking spaces is obtained according to the multi-dimensional quantitative evaluation index, including:

[0187] S321: Calculate the Euclidean distance between the parking area coordinate position of the vacant parking space and the current position Loc as the index value of the parking distance index of the vacant parking space;

[0188] S322: Calculate the number of vacant parking spaces on both sides of the vacant parking space and the ratio of vacant parking spaces in the horizontal direction to form the index value of the vacant parking space in the parking space redundancy index, where the index value of the mth vacant parking space in the parking space redundancy index is

[0189]

[0190] in:

[0191] Num m Indicates the number of vacant parking spaces on both sides of the mth vacant parking space;

[0192] Pm represents the proportion of available parking spaces in the horizontal direction of the m-th available parking space;

[0193] S323: Obtain the illumination brightness at the coordinate position of the parking space area where the available parking space is located, and use it as the index value of the available parking space in the parking space safety index;

[0194] S324: Calculate the Euclidean distance between the coordinate position of the parking space area where the available parking space is located and the elevator entrance of the parking lot, and use it as the index value of the available parking space in the access convenience index.

[0195] S4: Plan the vehicle driving route between the incoming vehicle and the optimal parking space, and send the vehicle driving route to the incoming vehicle by wireless communication. The incoming vehicle drives to the optimal parking space according to the vehicle driving route for parking.

[0196] In the S4 step, planning the vehicle driving route between the incoming vehicle and the optimal parking space includes:

[0197] Plan the vehicle driving route between the incoming vehicle and the optimal parking space, where the planning process of the vehicle driving route is as follows:

[0198] S41: Initialize and generate a traversal node list list, and add the position of the optimal parking space to list;

[0199] S42: Represent the parking lot in a position node form, where each position node is the inflection point of the road in the parking lot;

[0200] S43: Extract the latest added position from list as the parent node, and traverse the position nodes node adjacent to the parent node that are not in list. Calculate the path distance between the position node node and the parent node, and use it as the path cost Cost1 of the position node node;

[0201] S44: Starting from the position node node, traverse the hop points that are not in list and are the closest to the current position Loc among the position nodes adjacent to the starting point. Calculate the path distance Cost2 between the hop point and the optimal parking space and the path distance Cost3 between the hop point and the parent node. If Cost2 < min{Cost1, Cost3}, then use the position node node as the parent node and go to step S45; otherwise, return to step S43 to re-traverse the position nodes; where the hop point is a position node adjacent to the position node node or a position node with an obstacle between it and the position node node;

[0202] S45: Add the parent node to list, return to step S43, and continue until the current position Loc is added to list;

[0203] S46: extracting a position node sequence from the current position Loc to the optimal parking space from the traversed node list list as the vehicle entry route;

[0204] The vehicle's entry route is sent to the entering vehicle by wireless communication, and the entering vehicle drives to the optimal parking space according to the vehicle's entry route and parks.

[0205] Embodiment 2:

[0206] like Figure 2 As shown, it is a functional module diagram of a smart parking management system based on license plate recognition provided by an embodiment of the present invention, which can implement the smart parking management method based on license plate recognition in Example 1.

[0207] The smart parking management system 100 based on license plate recognition of the present invention can be installed in an electronic device. According to the functions implemented, the smart parking management system based on license plate recognition can include a parking space recognition module 101, a parking space management allocation module 102 and a route planning device 103. The module of the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0208] The parking space recognition module 101 is used to collect and pre-process the panoramic image of the parking lot, and recognize the parking spaces in the parking lot on the pre-processed panoramic image of the parking lot;

[0209] The parking space management and allocation module 102 is used to perform license plate recognition on different parking space areas of the parking lot, mark the parking space with the recognized license plate as a non-vacant parking space, mark the other parking spaces as vacant parking spaces, obtain vacant parking space distribution information, and calculate the optimal parking space by combining the current position of the incoming vehicle with the vacant parking space distribution information;

[0210] The route planning device 103 is used to plan a vehicle entry route between the vehicle and the optimal parking space, and send the vehicle entry route to the vehicle using wireless communication. The vehicle drives to the optimal parking space according to the vehicle entry route and parks.

[0211] In detail, each module in the smart parking management system 100 based on license plate recognition in the embodiment of the present invention is used in the same manner as above. Figure 1 The smart parking management method based on license plate recognition described in the invention has the same technical means and can produce the same technical effects, so I will not go into details here.

[0212] It should be understood that the embodiment is for illustration only and the scope of the patent application is not limited to this structure.

[0213] It should be noted that the serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments. And the terms "including", "comprising" or any other variants thereof in this article are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "including a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0214] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0215] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A smart parking management method based on license plate recognition, characterized in that: The method comprises: S1: Collecting and preprocessing the panoramic image of the parking lot, and identifying parking spaces on the preprocessed panoramic image of the parking lot; S2: Perform license plate recognition on different parking spaces in the parking lot, mark the parking spaces with recognized license plates as non-vacant parking spaces, and mark other parking spaces as vacant parking spaces, to obtain vacant parking space distribution information; S3: Calculate the optimal parking space by combining the current position of the incoming vehicle and the distribution information of available parking spaces; S4: planning a vehicle entry route between the entering vehicle and the optimal parking space, and sending the vehicle entry route to the entering vehicle by wireless communication, so that the entering vehicle drives to the optimal parking space according to the vehicle entry route and parks.

2. A smart parking management method based on license plate recognition as claimed in claim 1, characterized in that: In step S1, the panoramic image of the parking lot is collected and preprocessed, including: The parking lot panoramic image I is collected by overhead shooting, and the parking lot panoramic image I is preprocessed, wherein the preprocessing process is as follows: Convert the parking lot panoramic image I into pixel matrix form: I=(I(p,q)) P×Q I(p,q)=max{I R (p,q),I G (p,q),I B (p,q)} in: (I(p,q)) P×Q represents a matrix form of P rows and Q columns, where P represents the number of pixel rows of the parking lot panoramic image I, Q represents the number of pixel columns of the parking lot panoramic image I, and I(p,q) represents the pixel value of the pixel (p,q) in the p-th row and q-th column of the parking lot panoramic image I, p∈[1,P],q∈[1,Q]; I R (p,q),I G (p,q),I B (p, q) represents the color value of pixel (p, q) in R, G, and B color channels respectively; max{I R (p,q),I G (p,q),I B (p,q)} means to select I R (p,q),I G (p,q),I B The maximum value among (p,q); Perform smoothing filtering on the pixels in the parking lot panoramic image I, where the smoothing filtering result of the pixel (p, q) is g(p, q); Based on the smoothing filtering results of the pixels, a preprocessed parking lot panoramic image g = (g(p,q)) is constructed. P×Q , parking lot parking spaces are recognized on the preprocessed parking lot panoramic image g.

3. A smart parking management method based on license plate recognition as claimed in claim 2, characterized in that: The performing parking lot parking space recognition on the pre-processed parking lot panoramic image g comprises: S11: Calculate the gradient value of any pixel in the parking lot panoramic image g, where the gradient value of pixel (p, q) is grad(p, q); S12: Calculate the neighborhood edge value of any pixel in the parking lot panoramic image g, where the neighborhood edge value of pixel (p, q) is Neighbor (p, q); S13: If the neighborhood edge value of the pixel is greater than 0, it means that there is an edge point in the 3×3 neighborhood pixel range of the pixel, and the pixel with the largest gradient value is traversed in the 3×3 neighborhood pixel range of the pixel as the edge pixel; S14: Construct a coordinate system with the center of the parking lot panoramic image g as the coordinate origin, transform the pixel coordinates of the edge pixels into the coordinate system, and obtain the coordinate transformation result of the edge pixels, where the coordinate transformation result of the sth edge pixel is s∈[1,S], S represents the total number of edge pixels, Represents the horizontal coordinate of the sth edge pixel in the coordinate system, Represents the vertical coordinate of the sth edge pixel in the coordinate system; S15: Set the parking space angles to θ1, θ2, θ3, where θ1, θ2, θ3 are 0 respectively. π; Calculate the cumulative values ​​of all edge pixels at different parking space angles and the number of times the cumulative values ​​appear; The accumulated values ​​and the corresponding parking space angles with a number of occurrences higher than a preset threshold are retained, and the retained accumulated values ​​and parking space angles are used as a set of parking space lines; S16: If the parking space angles between adjacent parking space lines are perpendicular to each other and the distance between the centers of the parking space lines is less than a preset threshold, the adjacent parking space lines are regarded as a group of parking spaces; otherwise, the parking space lines are regarded as a group of parking spaces; S17: Repeat step S16 to fuse the parking space lines to obtain N parking spaces in the parking lot, where N represents the total number of parking spaces after the fusion process, and convert the coordinate positions of the parking spaces to the parking lot area to obtain N parking space areas in the parking lot.

4. A smart parking management method based on license plate recognition as claimed in claim 3, characterized in that: In step S2, the license plate recognition is performed on different parking spaces in the parking lot, and the parking spaces with recognized license plates are marked as non-vacant parking spaces, and other parking spaces are marked as vacant parking spaces, including: The parking space images of different parking spaces in the parking lot are collected, where the parking space image of the nth parking space area is I n , for parking space image I n Carry out license plate recognition, where the license plate recognition process is: S21: Parking space image I n Performing grayscale processing, and calculating the gradient value of the pixel based on the grayscale processing result; S22: Combine the gradient values ​​of pixels and use the Canny edge detection algorithm to detect the edge of the parking space image I n Detect edge pixels in the S23: using dilation and erosion methods in sequence to perform morphological processing on edge pixels, connecting edge pixels to form an edge closed area; S24: Calculate the size information of the edge closed area, and select the edge closed area whose size information meets the size threshold as the license plate candidate area; S25: extracting HOG features and SIFT features of the license plate candidate area, and using the support vector machine model to detect whether there is a license plate in the license plate candidate area. If a license plate is detected, it means that the license plate is recognized in the parking space of the nth parking space area, otherwise it means that the license plate is not recognized; Mark the parking space with the recognized license plate as a non - vacant parking space, and mark other parking spaces as vacant parking spaces to obtain the distribution information L of vacant parking spaces in the parking lot.

5. The smart parking management method based on license plate recognition as claimed in claim 4, characterized in that: In step S3, calculating the optimal parking space by combining the current position of the incoming vehicle with the distribution information of vacant parking spaces includes: Calculating the optimal parking space by combining the current position Loc of the incoming vehicle with the distribution information L of vacant parking spaces, where the current position Loc is the entrance position of the incoming vehicle entering the parking lot. The calculation process of the optimal parking space is as follows: S31: Construct a multi - dimensional quantitative evaluation index, where the multi - dimensional quantitative evaluation index includes a parking space distance index, a parking space redundancy index, a parking space safety index, and an access convenience index; S32: Obtain the index information of different vacant parking spaces according to the multi-dimensional quantitative evaluation index, where the index information of the mth vacant parking space is represents the index values ​​of the parking distance index, parking redundancy index, parking safety index and access convenience index of the mth vacant parking space in sequence; S33: Standardize the index information of the vacant parking spaces, where the index information F m The normalized result is In order The normalized processing results; S34: Calculate the weights of different indexes to form an index weight vector w = [w1, w2, w3, w4], where w1, w2, w3, w4 are the weights of the parking space distance index, the parking space redundancy index, the parking space safety index, and the access convenience index in sequence; S35: Use the index weight vector w to weight the normalized index information to obtain weighted index information of vacant parking spaces, where the index information f m The corresponding weighted index information is S36: Construct ideal weighted index information f based on the weighted index information of all available parking spaces + And the negative ideal weighted index information f - ; S37: Combine ideal weighted index information f + And the negative ideal weighted index information f - The support of each vacant parking space is calculated, and the vacant parking space with the highest support is selected as the optimal parking space.

6. A smart parking management method based on license plate recognition as claimed in claim 5, characterized in that: In step S32, obtaining the index information of different vacant parking spaces according to the multi - dimensional quantitative evaluation index includes: S321: Calculate the Euclidean distance between the coordinate position of the parking space area where the vacant parking space is located and the current position Loc as the index value of the vacant parking space in the parking space distance index; S322: Calculate the number of vacant parking spaces on both sides of the vacant parking space and the ratio of vacant parking spaces in the horizontal direction to form the index value of the vacant parking space in the parking space redundancy index, where the index value of the mth vacant parking space in the parking space redundancy index is S323: Obtain the illumination brightness of the coordinate position of the parking space area where the vacant parking space is located as the index value of the vacant parking space in the parking space safety index; S324: Calculate the Euclidean distance between the coordinate position of the parking space area where the vacant parking space is located and the elevator entrance of the parking lot as the index value of the vacant parking space in the access convenience index.

7. The smart parking management method based on license plate recognition as claimed in claim 1, characterized in that: In step S4, planning the vehicle driving route between the incoming vehicle and the optimal parking space includes: Planning the vehicle driving route between the incoming vehicle and the optimal parking space. The planning process of the vehicle driving route is as follows: S41: Initialize and generate a traversal node list list, and add the position of the optimal parking space to list; S42: Represent the parking lot in a position - node form, where each position node is the inflection point of the road in the parking lot; S43: Extract the latest added position from list as the parent node, and traverse the position nodes adjacent to the parent node that are not in list. Calculate the path distance between the position node node and the parent node as the path cost Cost1 of this position node node; S44: Starting from this position node node, traverse the hop points that are not in list and are the closest to the current position Loc among the position nodes adjacent to the starting point. Calculate the path distance Cost2 between the hop point and the optimal parking space and the path distance Cost3 between the hop point and the parent node. If Cost2 < min{Cost1, Cost3}, then take this position node node as the parent node and turn to step S45; otherwise, return to step S43 to re - traverse the position nodes. Here, the hop point is a position node adjacent to the position node node or a position node with an obstacle between it and the position node node; S45: Add the parent node to list, and return to step S43 until the current position Loc is added to list; S46: extracting a position node sequence from the current position Loc to the optimal parking space from the traversed node list list as the vehicle entry route; The vehicle's entry route is sent to the entering vehicle by wireless communication, and the entering vehicle drives to the optimal parking space according to the vehicle's entry route and parks.

8. The smart parking management method based on license plate recognition as claimed in claim 5, characterized in that: In the step S37, the ideal weighted index information f + And the negative ideal weighted index information f - Calculate the support of each vacant parking space, including: The support of the mth vacant parking space is: in: represents traversal m∈[1,M], The minimum value of represents traversal m∈[1,M], The maximum value of Support m represents the support of the mth vacant parking space.

9. A smart parking management system based on license plate recognition, characterized in that: The system comprises: The parking space recognition module is used to collect and pre-process the panoramic image of the parking lot, and recognize the parking spaces in the parking lot based on the pre-processed panoramic image of the parking lot; The parking space management and allocation module is used to perform license plate recognition on different parking space areas of the parking lot, mark the parking spaces with recognized license plates as non-vacant parking spaces, and mark other parking spaces as vacant parking spaces, obtain the vacant parking space distribution information, and calculate the optimal parking space by combining the current position of the incoming vehicle with the vacant parking space distribution information; A route planning device is used to plan a vehicle entry route between an entering vehicle and an optimal parking space, and to send the vehicle entry route to the entering vehicle by wireless communication. The entering vehicle drives to the optimal parking space according to the vehicle entry route and parks, thereby realizing a smart parking management method based on license plate recognition as described in any one of claims 1 to 8.

Images