Abnormal parking identification method based on artificial intelligence large model

Abstract

The invention discloses an abnormal parking identification method based on an artificial intelligence large model, and belongs to the technical field of vehicle identification, and the method comprises the steps: 1, obtaining the related data information, related to vehicle parking identification, of a monitoring region from a plurality of data terminals; step 2, according to the planning in the monitoring area, dividing a parking available area and a parking prohibited area; 3, judging whether an abnormal parking behavior exists in the parking forbidding area or not; and 4, judging whether an abnormal parking behavior exists in the parking area or not. The parking position can be monitored in real time, manual inspection is not needed, the efficiency can be effectively improved, the parking available area and the parking prohibited area are divided according to planning in the monitoring area, analysis is carried out according to related data information in the parking available area and the parking prohibited area, and the parking position can be monitored in real time. Different scene requirements can be covered in a targeted manner, the accuracy of abnormal recognition is improved, and the recognition efficiency can be remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle identification, and in particular relates to an abnormal parking identification method based on an artificial intelligence large model. Background Art

[0002] With the continued acceleration of urbanization, the number of vehicles has exploded, and traffic congestion has become increasingly severe. Parking management has become a major challenge in urban transportation. Illegal and irregular parking behaviors are common, seriously disrupting normal traffic order and posing a significant challenge to urban traffic management. For example, vehicles occupying emergency lanes, fire lanes, or sidewalks for extended periods of time can hinder the normal flow of vehicles and pedestrians. In emergency situations, they can even severely hinder the timely arrival of rescue vehicles, delaying the optimal time for assistance.

[0003] Traditional methods for detecting parking violations rely primarily on manual patrols and surveillance camera playback. Manual patrols suffer from limited coverage and low efficiency, making it difficult to conduct comprehensive, real-time monitoring of large areas. Surveillance camera playback, on the other hand, requires significant manpower and time to review and cannot provide real-time processing. Consequently, violations are often discovered long after they occur, delaying their resolution. Summary of the Invention

[0004] The purpose of the present invention is to provide an abnormal parking recognition method based on an artificial intelligence large model to solve the problems faced in the above-mentioned background technology.

[0005] The purpose of the present invention can be achieved through the following technical solutions:

[0006] An abnormal parking identification method based on an artificial intelligence large model, the method comprising:

[0007] Step 1: Obtain relevant data information related to the monitoring area and vehicle parking identification from multiple data terminals, and perform data cleaning and normalization on the relevant data information;

[0008] Step 2: Divide the parking area into permitted areas and prohibited areas according to the planning within the monitoring area;

[0009] Step 3: Analyze the collected related data information of the prohibited parking area to determine whether there is any abnormal parking behavior in the prohibited parking area;

[0010] Step 4: Based on the historical data information of the parking area, a recognition matrix model is constructed, and the real-time collected related data information is input into the corresponding recognition matrix model to determine whether there is abnormal parking behavior in the parking area.

[0011] Furthermore, the associated data information includes vehicle location information, monitoring information of the monitoring area, and vehicle status information.

[0012] Furthermore, the working method of step three is:

[0013] In the prohibited parking area, the monitoring equipment in the area obtains the vehicle's parking time t, the deviation position L between the vehicle center and the lane center, and the angle θ between the vehicle direction and the lane direction. At the same time, the vehicle status information, including the engine start and stop status C, is obtained from the vehicle networking system. F , Double flash lights on state C D and vehicle speed state C V ;

[0014] By formula The first outlier R A ;

[0015] when When , it is determined that there is abnormal parking behavior in the prohibited parking area;

[0016] Among them, τ is the regional influence coefficient, L0 is the set deviation threshold, is the first abnormality judgment threshold.

[0017] Furthermore, the engine start-stop state C F , Double flash lights on state C D and vehicle speed state C V The acquisition method is:

[0018] When the engine is turned off C F =1, otherwise C F =0.5, when the hazard lights are on, C D =1, otherwise C D =0.1;

[0019] When the vehicle speed is higher than 10km / s, C V =0; when the speed is lower than 10km / s, C V The value increases by 0.1. When the vehicle speed is 0, C V =1.

[0020] Furthermore, the working method of step 4 is:

[0021] Obtain a large amount of normal parking monitoring information and corresponding location information in the parking area, and construct a standard identification matrix G = |G1(x1, y1)...G according to the coordinate parameters of each key position corresponding to the parking position in the parking area. i (x i ,y i )...G n (xn ,y n )|, where n is the number of key positions, G1(x1, y1) is the coordinate point of the first key position, G i (x i ,y i ) is the coordinate point of the jth key position, and i∈[1,n];

[0022] At the same time, the coordinate parameters of each key position corresponding to the parking area are obtained according to the monitoring information, and a parking recognition matrix P = |P1(x1, y1)...P i (x i ,y i )...P n (x n ,y n )|, P1(x1, y1) is the parking coordinate point of the first key position, P i (x i ,y i ) is the stop i coordinate point of the jth key position;

[0023] Based on the standard recognition matrix G and the parking recognition matrix P, the abnormal matrix K = |K1...K i ...K n |, where

[0024] Based on the abnormal matrix K, through the formula The second outlier R B , δ i is the weight ratio of the i-th key position;

[0025] The second outlier R B The second abnormality judgment threshold set by the system To compare, when When , it is determined that there is abnormal parking behavior in the parking area.

[0026] Furthermore, the working method of step 4 also includes:

[0027] When it is determined that there is no abnormal parking behavior in the parking area, the vehicle type of the corresponding parking position is identified through the monitoring equipment, and it is determined whether the vehicle type is a vehicle type that is allowed to park. If the vehicle is identified as an unauthorized vehicle type, it is determined that there is abnormal parking behavior in the parking area.

[0028] Beneficial effects of the present invention:

[0029] The present invention can monitor parking positions in real time, does not require manual inspection, can effectively improve efficiency, and divides parking areas and prohibited parking areas according to the planning within the monitoring area, and analyzes relevant data information in the parking areas and prohibited parking areas, which can cover different scene requirements in a targeted manner, improve the accuracy of abnormal recognition, and significantly improve recognition efficiency.

[0030] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0032] Figure 1 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0033] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0034] In one embodiment, a method for identifying abnormal parking based on an artificial intelligence large model is disclosed, such as Figure 1 As shown, the method includes:

[0035] Step 1: Acquire related data information related to the monitoring area and vehicle parking identification from multiple data terminals, the related data information including vehicle location information, monitoring information of the monitoring area, and vehicle status information, and perform data cleaning and normalization on the related data information;

[0036] Step 2: Divide the parking area into permitted areas and prohibited areas according to the planning within the monitoring area;

[0037] Step 3: Analyze the collected related data information of the prohibited parking area to determine whether there is any abnormal parking behavior in the prohibited parking area;

[0038] Step 4: Based on the historical data information of the parking area, a recognition matrix model is constructed, and the real-time collected related data information is input into the corresponding recognition matrix model to determine whether there is abnormal parking behavior in the parking area.

[0039] Through the above technical solution, the present application first obtains associated data information related to vehicle parking identification in the monitoring area from multiple data terminals. The associated data information includes vehicle location information, monitoring information of the monitoring area, and vehicle status information, and performs data cleaning and normalization processing on the associated data information. By obtaining data information related to vehicle parking identification, it is convenient to comprehensively analyze whether parking is abnormal based on multiple influencing factors, which can increase the accuracy of judgment. Cleaning and normalizing the data information can improve the accuracy of the acquired data. At the same time, data from different sources are normalized to unify different sources of data, which facilitates subsequent calculation and analysis. Then, according to the planning within the monitoring area, parking areas and prohibited parking areas are divided. The associated data information collected in the prohibited parking areas is analyzed to obtain a first outlier value. The size of the first outlier value is used to determine whether there is abnormal parking behavior in the prohibited parking areas. At the same time, based on the historical data information of the parking area, an identification matrix model is constructed. The associated data information collected in real time is input into the corresponding identification matrix model to obtain a second outlier value. The size of the second outlier value is used to determine whether there is abnormal parking behavior in the parking area. In this way, parking locations can be monitored in real time, which does not require manual inspections and can effectively improve efficiency. According to the planning within the monitoring area, parking areas and prohibited parking areas can be divided, and the relevant data information in the parking areas and prohibited parking areas can be analyzed. It can cover the needs of different scenarios in a targeted manner, improve the accuracy of abnormal recognition, and significantly improve recognition efficiency.

[0040] The working method of step 3 is as follows: in the prohibited parking area, the vehicle parking time t, the deviation position L between the vehicle center and the lane center, and the angle θ between the vehicle direction and the lane direction are obtained through the monitoring equipment in the area. At the same time, the vehicle status information, including the engine start and stop status C, is obtained from the vehicle network system. F , Double flash lights on state C D and vehicle speed state C V ;

[0041] By formula The first outlier R A ;

[0042] when When , it is determined that there is abnormal parking behavior in the prohibited parking area;

[0043] Among them, τ is the regional influence coefficient, L0 is the set deviation threshold, is the first abnormality judgment threshold;

[0044] Engine start-stop state C F , Double flash lights on state C Dand vehicle speed state C V The acquisition method is:

[0045] When the engine is turned off C F =1, otherwise C F =0.5, when the hazard lights are on, C D =1, otherwise C D =0.1;

[0046] When the vehicle speed is higher than 10km / s, C V =0; when the speed is lower than 10km / s, C V The value increases by 0.1. When the vehicle speed is 0, C V =1.

[0047] The above scheme provides a specific method for determining whether there is abnormal parking behavior in a prohibited parking area. In a prohibited parking area, if the vehicle stays for too long, the deviation between the center of the vehicle and the center of the lane is large, or the vehicle is severely tilted in the middle of the road, then it is likely that the vehicle has parked abnormally. Similarly, if the vehicle is turned off, the hazard lights are on, or the speed is very slow in a prohibited parking area, then it is likely that it has parked abnormally in the area. Therefore, the monitoring equipment in the area obtains the vehicle's parking time t, the deviation between the center of the vehicle and the center of the lane L, and the angle θ between the vehicle direction and the lane direction. At the same time, it obtains the vehicle's status information from the vehicle network system, including the engine start and stop status C F , Double flash lights on state C D and vehicle speed state C V , and then through the formula The first outlier R A , where τ is the regional influence coefficient, L0 is the set deviation threshold, is the first abnormality judgment threshold, the deviation threshold and the judgment threshold can be formulated according to historical experience data, and the regional influence coefficient can be determined according to the parking position in the prohibited parking area. For example, it can be determined according to the influence ratio of fire passages, blind paths, emergency lanes, flammable and explosive dangerous areas, and other area types. It is not necessary to describe it here. It can be seen that the larger the first abnormal value is, the greater the possibility of abnormal parking of the vehicle is. Therefore, when hour, is the first abnormality judgment threshold, which is also determined based on historical parking data and experience. If there is abnormal parking behavior in the prohibited parking area, an alarm will be issued. This method can be used to conduct a comprehensive analysis based on the operating status of the vehicle itself in the prohibited parking area and the vehicle image information obtained by the monitoring equipment, so as to more accurately identify abnormal parking of the vehicle, thereby ensuring recognition efficiency and accuracy.

[0048] The working method of step 4 is: obtain a large amount of normal parking monitoring information and corresponding location information of the parking area, and construct a standard identification matrix G = |G1(x1, y1)...G according to the coordinate parameters of each key position corresponding to the parking position in the parking area. i (x i ,y i )...G n (x n ,y n )|, where n is the number of key positions, G1(x1, y1) is the coordinate point of the first key position, G i (x i ,y i ) is the coordinate point of the jth key position, and i∈[1,n];

[0049] At the same time, the coordinate parameters of each key position corresponding to the parking area are obtained according to the monitoring information, and a parking recognition matrix P = |P1(x1, y1)...P i (x i ,y i )...P n (x n ,y n )|, P1(x1, y1) is the parking coordinate point of the first key position, P i (x i ,y i ) is the stop i coordinate point of the jth key position;

[0050] Based on the standard recognition matrix G and the parking recognition matrix P, the abnormal matrix K = |K1...K i ...K n |, where

[0051] Based on the abnormal matrix K, through the formula The second outlier R B , δ i is the weight ratio of the i-th key position;

[0052] The second outlier R B The second abnormality judgment threshold set by the system To compare, when When , it is determined that there is abnormal parking behavior in the parking area;

[0053] The working method of step four also includes: when it is determined that there is no abnormal parking behavior in the parking area, the vehicle type corresponding to the parking position is identified through the monitoring equipment, and whether the vehicle type is a vehicle type that is allowed to park. If the vehicle is identified as an unauthorized vehicle type, it is determined that there is abnormal parking behavior in the parking area.

[0054] The above scheme provides a specific method for determining whether there is abnormal parking behavior in the parking area. First, a large amount of normal parking monitoring information and corresponding location information of the parking area are obtained. According to the coordinate parameters of each key position corresponding to the parking position in the parking area, a standard identification matrix G = |G1(x1, y1)...G i (x i ,y i )...G n (x n ,y n )|, where n is the number of key positions, G1(x1, y1) is the coordinate point of the first key position, G i (x i ,y i ) is the coordinate point of the jth key position. In this way, there is a standard position left for multiple key position points of each fixed parking position. Then, the coordinate parameters of each key position of the corresponding position in the parking area are obtained according to the monitoring information, and the parking recognition matrix P = | P1(x1, y1)...P i (x i ,y i )...P n (x n ,y n )|, P1(x1, y1) is the parking coordinate point of the first key position, P i (x i ,y i ) is the parking i coordinate point of the jth key position. In this way, the real-time coordinates of each key position of the parking position can be obtained. According to the coordinates of the two, based on the standard recognition matrix G and the parking recognition matrix P, the abnormal matrix K = |K1...K i ...K n |, where It can be seen that when K i The smaller the value of , the smaller the gap between the real-time parking coordinates and the standard coordinates, which means the possibility of parking anomalies is smaller. Therefore, based on the anomaly matrix K, the formula The second outlier R B , δ iis the weight ratio of the i-th key position, which is independently formulated according to the influence of each key position. For example, the influence of the two side positions on parking is generally slightly higher than that of the front and rear of the vehicle. The weight of the key positions on both sides can be appropriately increased. The specific weight ratio is determined according to the actual situation of the position. From the formula, it can be seen that when the second abnormal value R B The larger the value of , the greater the difference between the coordinates of each key position and the standard coordinates, and the greater the possibility of parking anomaly. Therefore, the second abnormal value R B The second abnormality judgment threshold set by the system To compare, when When the vehicle is parked in a reasonable area, it is judged that there is abnormal parking behavior in the parking area. In this way, the difference between the coordinates of multiple key position points in the parking area and each standard coordinate can be combined to analyze the parking situation of the vehicle. In this way, it is possible to quickly identify whether the vehicle parking is abnormal based on the parking status to ensure recognition efficiency. In addition, since the above method can only determine whether the vehicle is parked in a reasonable location area, if the parked vehicle is of a type that is not allowed to park, such as a motor vehicle parked in a non-motor vehicle, then there is also abnormal parking behavior. Therefore, when it is judged that there is no abnormal parking behavior in the parking area, the monitoring equipment is used to identify the vehicle type corresponding to the parking position and determine whether the vehicle type is a vehicle type that is allowed to park. If the vehicle is identified as an unauthorized vehicle type, it is judged that there is abnormal parking behavior in the parking area, which can improve the accuracy of detection.

[0055] The above content is merely an example and explanation of the concept of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the claims, they should all fall within the scope of protection of the present invention.

Claims

1. The abnormal parking identification method based on artificial intelligence large model is characterized by: The method comprises: Step 1: Obtain relevant data information related to the monitoring area and vehicle parking identification from multiple data terminals, and perform data cleaning and normalization on the relevant data information; Step 2: Divide the parking area into permitted areas and prohibited areas according to the planning within the monitoring area; Step 3: Analyze the associated data information collected from the prohibited parking areas to determine whether there is any abnormal parking behavior in the prohibited parking areas; Step 4: Based on the historical data information of the parking area, a recognition matrix model is constructed, and the real-time collected related data information is input into the corresponding recognition matrix model to determine whether there is abnormal parking behavior in the parking area.

2. The abnormal parking identification method based on artificial intelligence large model according to claim 1 is characterized in that: The associated data information includes vehicle location information, monitoring information of the monitoring area, and vehicle status information.

3. The abnormal parking identification method based on artificial intelligence large model according to claim 2 is characterized in that: The working method of step three is: In the prohibited parking area, the monitoring equipment in the area obtains the vehicle's parking time t, the deviation position L between the vehicle center and the lane center, and the angle θ between the vehicle direction and the lane direction. At the same time, the vehicle status information, including the engine start and stop status C, is obtained from the vehicle networking system. F , Double flash lights on state C D and vehicle speed state C V ; By formula The first outlier value R A ; when When , it is determined that there is abnormal parking behavior in the prohibited parking area; Among them, τ is the regional influence coefficient, L0 is the set deviation threshold, is the first abnormality judgment threshold.

4. The abnormal parking identification method based on artificial intelligence large model according to claim 3 is characterized in that: The engine start-stop state C F , Double flash lights on state C D and vehicle speed state C V The acquisition method is: When the engine is turned off C F =1, otherwise C F =0.5, when the hazard lights are on, C D =1, otherwise C D =0.1; When the vehicle speed is higher than 10km / s, C V =0; when the speed is lower than 10km / s, C V The value increases by 0.

1. When the vehicle speed is 0, C V =1.

5. The abnormal parking identification method based on artificial intelligence large model according to claim 1 is characterized in that: The working method of step 4 is: Obtain a large amount of normal parking monitoring information and corresponding location information in the parking area, and construct a standard identification matrix G = |G1(x1, y1)...G according to the coordinate parameters of each key position corresponding to the parking position in the parking area. i (x i ,y i )...G n (x n ,y n )|, where n is the number of key positions, G1(x1, y1) is the coordinate point of the first key position, G i (x i ,y i ) is the coordinate point of the jth key position, and i∈[1,n]; At the same time, the coordinate parameters of each key position corresponding to the parking area are obtained according to the monitoring information, and a parking recognition matrix P = |P1(x1, y1)...P i (x i ,y i )...P n (x n ,y n )|, P1(x1, y1) is the parking coordinate point of the first key position, P i (x i ,y i ) is the stop i coordinate point of the jth key position; Based on the standard recognition matrix G and the parking recognition matrix P, the abnormal matrix K = |K1...K i ...K n |, where Based on the abnormal matrix K, through the formula The second outlier R B , δ i is the weight ratio of the i-th key position; The second outlier R B The second abnormality judgment threshold set by the system To compare, when When , it is determined that there is abnormal parking behavior in the parking area.

6. The abnormal parking identification method based on artificial intelligence large model according to claim 5 is characterized in that: The working method of step 4 also includes: When it is determined that there is no abnormal parking behavior in the parking area, the vehicle type of the corresponding parking position is identified through the monitoring equipment, and it is determined whether the vehicle type is a vehicle type that is allowed to park. If the vehicle is identified as an unauthorized vehicle type, it is determined that there is abnormal parking behavior in the parking area.

Images