A high speed violation detection method and system

Abstract

The application discloses a high-speed violation detection method and system, and relates to the technical field of intelligent traffic management. The method comprises the following steps: constructing a high-speed violation monitoring library; presetting a clustering algorithm, and classifying the high-speed violation monitoring library according to clustering results, wherein the classification information comprises multi-level classification; extracting primary classification features, matching a collection device, and embedding the primary classification features into the matching collection device as a screening condition; starting the matching collection device to collect and monitor high-speed vehicles in a preset area, and using the screening condition to iteratively compare the monitoring information; when the screening condition is matched, sending calling information, calling associated high-level classification features of the primary classification features corresponding to the screening condition; using the associated high-level classification features to match the monitoring information in order from small to large according to the size of the levels, and obtaining a matching result; and based on the matching result, identifying violation information to position and feed back the vehicle.

Description

Technical Field

[0001] This invention relates to the field of intelligent traffic management technology, and in particular to a method and system for detecting violations on highways. Background Technology

[0002] With the continuous development of urban transportation and the increase in traffic flow, traffic violations on highways have become a serious problem, easily leading to traffic accidents and congestion. Highway violation detection is characterized by high speed and high detection difficulty. Existing highway violation detection methods suffer from technical problems such as high false positive and false negative rates for specific traffic violations, large discrimination errors, and ambiguous distinctions in violation levels. Summary of the Invention

[0003] The purpose of this application is to provide a method and system for detecting high-speed traffic violations. This addresses the technical problems in existing technologies, such as high false positive and false negative rates for specific traffic violations, large discrimination errors, and ambiguous violation severity classifications.

[0004] To address the above technical issues, this application provides a high-speed violation detection method and system.

[0005] Firstly, this application provides a high-speed violation detection method, wherein the method includes:

[0006] By analyzing highway violation cases, a highway violation monitoring database is constructed. The database is then clustered using a pre-defined clustering algorithm, and classified according to the clustering results. The classification information includes multi-level classifications, which characterize the different levels of identification of highway violation events. Primary classification features are extracted from the classification information, and data acquisition devices are matched according to the classification information. These primary classification features are embedded into the matching data acquisition devices as filtering conditions. The matching data acquisition devices are activated to collect and monitor highway vehicles within a preset area, and the monitoring information is traversed and compared using the filtering conditions. When a match is found, a call is sent to invoke the associated high-level classification features of the primary classification features corresponding to the filtering conditions. The monitoring information is matched sequentially according to the level of the associated high-level classification features, from smallest to largest, to obtain matching results. Based on the matching results, violation information is identified, and vehicle location feedback is provided.

[0007] Secondly, this application also provides a high-speed violation detection system, wherein the system includes:

[0008] The system comprises the following modules: a monitoring database module, which analyzes highway violation cases to construct a highway violation monitoring database; a clustering and classification module, which clusters the highway violation monitoring database using a preset clustering algorithm and classifies it according to the clustering results, wherein the classification information includes multi-level classification, which characterizes the different feature recognition levels of highway violation events; a device configuration module, whose feature matching module extracts primary classification features from the classification information, matches acquisition devices according to the classification information, and embeds the primary classification features as filtering conditions into the matched acquisition devices; and a monitoring acquisition module. The monitoring and matching module is used to activate the matching and acquisition device to collect and monitor high-speed vehicles within a preset area, and to compare and traverse the monitoring information using screening conditions; the grade feature module is used to send a call message when it matches the screening conditions, and call the associated high-level classification features of the primary classification features corresponding to the screening conditions; the grade matching module is used to match the monitoring information in ascending order of grade using the associated high-level classification features to obtain the matching results; the location feedback module is used to identify violation information and provide location feedback for the vehicle based on the matching results.

[0009] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0010] This application constructs a highway violation monitoring database by analyzing highway violation cases; it uses a pre-set clustering algorithm to cluster the database and classifies it according to the clustering results. The classification information includes multi-level classifications, which characterize the different levels of recognition of highway violation events. Primary classification features are extracted from the classification information and matched with acquisition devices according to the classification information. These primary classification features are embedded into the matching acquisition devices as filtering conditions. The matching acquisition devices are activated to collect and monitor highway vehicles within a preset area, and the monitoring information is traversed and compared using the filtering conditions. When a match is found, a call is sent to invoke the associated high-level classification features of the primary classification features corresponding to the filtering conditions. The monitoring information is matched using the associated high-level classification features in ascending order of level to obtain matching results. Based on the matching results, violation information is identified, and vehicle location feedback is provided. This approach has the technical effect of improving the ability to monitor traffic violations, reducing false positives and false negatives, improving the accuracy of judgment, and accurately determining the level of violation.

[0011] The above description is merely an overview of the technical solution of this application. In order to more clearly explain the technical means of this application, and to enable its implementation in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are described below. Attached Figure Description

[0012] The embodiments of the present invention and the following brief description are illustrated in conjunction with the figures, which are described below:

[0013] Figure 1 This is a flowchart illustrating a high-speed violation detection method according to this application;

[0014] Figure 2 This is a schematic diagram illustrating the process of constructing a high-speed violation monitoring database in a high-speed violation detection method according to this application;

[0015] Figure 3 This is a schematic diagram of the structure of a high-speed violation detection system according to this application.

[0016] Figure labeling: Monitoring library module 11, Clustering and classification module 12, Equipment configuration module 13, Monitoring and acquisition module 14, Level feature module 15, Level matching module 16, Location feedback module 17. Detailed Implementation

[0017] This application provides a high-speed violation detection method, which solves the technical problems of high false positive and false negative rates, large discrimination errors, and ambiguous violation level discrimination in the prior art for specific traffic violations.

[0018] The overall approach adopted in this technical embodiment to solve the above problems is as follows:

[0019] First, a highway violation monitoring database is constructed by analyzing highway violation cases. A pre-set clustering algorithm is used to cluster the database, and the database is then classified according to the clustering results. The classification information includes multi-level classifications, which characterize the different levels of recognition of highway violation events. Primary classification features are extracted from the classification information and matched with acquisition devices according to the classification information. These primary classification features are then embedded into the matching acquisition devices as filtering conditions. The matching acquisition devices are activated to collect and monitor highway vehicles within a preset area, and the monitoring information is traversed and compared using the filtering conditions. When a match is found, a call is sent to retrieve the associated high-level classification features of the primary classification features corresponding to the filtering conditions. The monitoring information is then matched using the associated high-level classification features in ascending order of level to obtain matching results. Based on the matching results, violation information is identified, and vehicle location feedback is provided. This method has the technical effect of improving the ability to monitor traffic violations, reducing false positives and false negatives, improving the accuracy of judgment, and accurately determining the level of violation.

[0020] To better understand the above technical solutions, the following detailed description will be provided in conjunction with the accompanying drawings and specific embodiments. It should be noted that the described embodiments are only a part of the embodiments of this application, not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention. Furthermore, it should be noted that, for ease of description, only the parts related to this invention are shown in the accompanying drawings, not all of them.

[0021] Example 1

[0022] like Figure 1 As shown, this application provides a high-speed violation detection method, the method comprising:

[0023] S100: By analyzing cases of highway violations, a highway violation monitoring database will be built;

[0024] The Highway Violation Monitoring Database is a specialized database used to monitor, match, and evaluate highway violations. It contains analysis results and discriminant features of multiple cases related to various types of highway violations under diverse conditions. The database is acquired through the collection and analysis of highway violation cases, including images, videos, and related data of various types of highway violations.

[0025] Furthermore, the step S100, which involves analyzing highway violation cases to construct a highway violation monitoring database, includes:

[0026] S110: Semantic analysis of high-speed violation rules is performed through the semantic analysis module to extract violation discrimination features;

[0027] S120: Collect missed cases from the current violation identification system's case database;

[0028] S130: Use the overlooked cases to perform violation discrimination feature matching, determine the overlooked violation discrimination features, and construct the high-speed violation monitoring database based on the overlooked violation discrimination features.

[0029] Semantic analysis, an artificial intelligence analysis technique, is used to understand and extract content expressed in contexts such as text, speech, or images. Semantic analysis of highway violation rules utilizes natural language processing and machine learning technologies to conduct in-depth analysis of violation rules on highways. This efficiently and accurately transforms abstract and lengthy textual highway violation rules into clear rules and discriminative features that machines can understand, enabling more accurate identification and detection of violations.

[0030] An exemplary semantic analysis of highway violation rules includes: Keyword extraction: First, during semantic analysis, keywords, phrases, and terms are extracted from the rule text. These keywords help identify the characteristics and conditions of the violation. Semantic relationship analysis: Next, the semantic analysis module analyzes the semantic relationships in the rules to determine the logical relationships between different conditions. This helps to understand the complexity of the rules and the dependencies between conditions. Feature extraction: Then, based on the analysis of keywords and semantic relationships, features related to the violation are extracted. These features include time, location, vehicle status, driving trajectory, etc. Pattern recognition: The extracted features are then used to train machine learning models, including classifiers, neural network models, etc. These models determine whether a violation has occurred based on the extracted features. Violation discrimination feature generation: After training, features for violation discrimination are generated based on the model's output. These features are used in subsequent violation detection processes.

[0031] The current violation identification system case library refers to the existing case library of high-speed violation identification systems, which includes cases that were missed or not accurately identified, i.e., oversighted identification cases, and has important reference value. Further analysis and improvement of oversighted identification cases will enable the subsequent technical effect of targeted analysis of features with large discrimination errors or that cannot be identified by the current detection system.

[0032] As described in the above embodiments, the omission-identified cases are used for violation feature matching, and the discriminative features are extracted and stored as omission violation discriminative features. This enables targeted matching and detection of violation cases that are difficult to detect at high speed and have a high probability of omission and misidentification in existing technologies.

[0033] Optionally, a high-speed violation monitoring database can be constructed based on omission violation discrimination features. For violation features and violation cases with high recognition rates and mature discrimination, their content can be retained. For violation features and violation cases with problems in the existing high-speed violation detection, omission violation discrimination features can be used to supplement or replace them. While achieving the integrity and coverage of the high-speed violation monitoring database, existing mature discrimination features can be effectively utilized to reduce repetitive violation discrimination feature matching, thereby improving the construction efficiency of the high-speed violation monitoring database.

[0034] Furthermore, step S120 also includes collecting cases of oversight in the current violation identification system's case database:

[0035] S121: Analyze the current violation identification system to obtain the identification error rate;

[0036] S122: Based on a preset error rate threshold, the error rate of the method is screened to identify cases of error violations;

[0037] S123: Collect the violation record database, and compare the violation record database with the identification record data of the current violation identification system to determine the missing cases;

[0038] S124: Integrate the error violation cases and missing cases to obtain the omission identification cases.

[0039] The current violation detection system's detection status includes the number of detected violations, the number of normal detections, and the number of false positives. The method identification error rate is the ratio of the number of false positives for multiple violations to the total number of detected violations. Optionally, a preset error rate threshold is determined based on the target false positive rate. For example, if the target false positive rate in a region is less than 2%, then the preset error rate threshold is greater than or equal to 2%. The method identification error rate is filtered based on the preset error rate threshold. If a method identification error rate meets the preset error rate threshold, the violation case corresponding to this method identification error rate is set as an erroneous violation case.

[0040] The violation record database refers to a database containing multiple confirmed violations. These violation records are compared with the identification record data of the current violation identification system. If some cases recorded as violations in reality are not identified by the current system, these unidentified cases are considered missing cases.

[0041] In one feasible embodiment, optionally, the missed identification cases also include cases where traffic rules have changed but the current violation identification system has not been updated in a timely manner, resulting in missed detections.

[0042] In one feasible embodiment, optionally, the omission identification cases also include cases where the current violation identification system has identified a violation but failed to correctly determine the level or severity of the violation, resulting in a false detection.

[0043] S200: Cluster the high-speed violation monitoring database using a preset clustering algorithm, and classify the high-speed violation monitoring database according to the clustering results. The classification information includes multi-level classification, which is used to characterize the degree of recognition of different features of high-speed violation events.

[0044] Clustering algorithms are machine learning algorithms used to divide data into similar groups or clusters, including K-Means clustering algorithm, hierarchical clustering algorithm, DBSCAN (Density-Based Spatial Clustering of Applications with Noise), Mean Shift clustering algorithm, GMM (Gaussian Mixture Model), etc.

[0045] Furthermore, the high-speed violation monitoring database is classified according to the clustering results, wherein the classification information includes multi-level classification. Step S200 includes:

[0046] S210: Classify the high-speed violation monitoring database according to the clustering results and determine the classification information, wherein the classification information includes violation type and violation characteristics;

[0047] S220: Based on the aforementioned violation characteristics, conduct an impact analysis to determine the core violation characteristics and auxiliary violation characteristics;

[0048] S230: Divide the threshold influence range of the core violation features and auxiliary violation features respectively, and determine the multi-level identification threshold;

[0049] S240: Sort the core violation features, auxiliary violation features and multi-level identification thresholds in order from core violation features to auxiliary violation features, and from the largest identification threshold range to the smallest identification threshold range;

[0050] S250: Use the sorting results to classify the violation features in the classification information.

[0051] Optionally, the clustering results include multiple groups or clusters, representing multiple violation categories or violation features, and the violation type corresponds to the violation feature. One violation type can correspond to multiple violation features.

[0052] Among them, the impact analysis of violation features refers to analyzing whether the violation features play a qualitative or quantitative role in the corresponding violation type. Qualitative violation features, namely core violation features, are used to determine whether the behavior to be detected has a violation of the current category, while quantitative violation features, namely auxiliary violation features, are used to determine the degree of violation of the corresponding type of violation features.

[0053] For example, for the violation of crossing the lane line, the core violation feature is the overlap of the vehicle's wheels with the lane line, and the auxiliary violation feature is the duration of the overlap. Among them, the core violation feature can be divided into crossing dashed lines, crossing solid lines, and crossing guide lines according to the scope of impact; the auxiliary violation feature can be divided into crossing the line for less than 3 seconds and crossing the line for more than 3 seconds.

[0054] S300: Extract primary classification features from classification information, match the acquisition device according to the classification information, and embed the primary classification features as filtering conditions into the matching acquisition device;

[0055] This process involves extracting primary classification features related to each violation type and characteristic from the classification information. These features are common to highway violation cases, including vehicle speed, lane position, and vehicle type. These features will be used to select appropriate data collection equipment, enabling the system to more accurately identify different violation types.

[0056] Based on the extracted primary classification features, the system matches these features with different data acquisition devices. Different acquisition devices may focus on different parameters when collecting data, including vehicle speed, lane position, etc. Optionally, vehicle speed features and vehicle type features are matched with vehicle speed detection devices and embedded as filtering conditions into the vehicle speed detection devices. For example, for small passenger cars, the maximum speed is no more than 120 km / h and the minimum speed is no less than 80 km / h; for large passenger cars, the maximum speed is no more than 100 km / h and the minimum speed is no less than 60 km / h; for trucks, the maximum speed is no more than 80 km / h and the minimum speed is no less than 60 km / h.

[0057] By embedding the primary classification features as filtering conditions into the matching acquisition device, the system can select appropriate acquisition devices for data acquisition based on different violation types and features. This improves acquisition efficiency and provides more targeted data support for subsequent violation detection and analysis.

[0058] S400: Start the matching and acquisition device to collect and monitor high-speed vehicles in the preset area, and use the filtering conditions to traverse and compare the monitoring information.

[0059] The data collection equipment refers to devices used to acquire vehicle information for violation detection, including cameras, radar, laser rangefinders, geomagnetic sensors, and vehicle weighing systems. The monitored information is then traversed and compared to determine whether any violations exist on highways within a preset area; if so, the type of violation is also determined.

[0060] Optionally, before using the filtering conditions to traverse and compare the monitoring information, it also includes feature recognition of the monitoring information collected from high-speed vehicles within the preset area to obtain the vehicle monitoring features.

[0061] Furthermore, the matching and acquisition device is activated to collect and monitor high-speed vehicles within a preset area. When the matching and acquisition device is a camera, step S400 further includes:

[0062] S410: The video camera captures images of vehicles passing through a preset area to obtain monitoring video information;

[0063] S420: The monitoring video information is preprocessed by the video processing extraction module, and video is extracted according to a preset step size to obtain a first step length extracted video and a second step length extracted video, wherein the first step length is greater than the second step length;

[0064] S430: Use the video extracted in the first step as primary monitoring information and the video extracted in the second step as secondary monitoring information;

[0065] S440: Perform feature extraction on the primary monitoring information, match the feature extraction results with the screening conditions, and then perform feature extraction on the secondary monitoring information after successful matching.

[0066] Optionally, when the acquisition device is a camera, a digital sensor is used to capture the optical information of the scene, including a CCD sensor (Charge-Coupled Device) or a CMOS sensor (Complementary Metal-Oxide-Semiconductor). Preferably, a CMOS sensor is used, which has the advantages of low power consumption, small size, and high integration. For example, the acquisition resolution is 3200*2400, the aspect ratio is 4:3, and the acquisition frame rate is 60FPS.

[0067] Preprocessing refers to image processing techniques performed to improve image quality, reduce noise, and provide better input for subsequent processing. These include noise removal, contrast enhancement, image smoothing, color space conversion, and motion compensation. Preprocessing effectively provides higher-quality image information for subsequent feature extraction.

[0068] Video extraction from monitoring video information refers to extracting key frames from the monitoring video information at certain time intervals, which has the technical effect of reducing data volume and improving processing efficiency. The preset step size refers to the duration of the key frames extracted from the detected video information. The longer the preset step size, the fewer images are extracted from the same duration of detected video information, resulting in smaller file sizes and less subsequent processing. Conversely, the shorter the preset step size, the more images are extracted from the same duration of detected video information, preserving more complete information and facilitating more detailed further analysis and processing.

[0069] Optionally, the preset step size includes a first step size and a second step size. The first step size is larger than the second step size, and the second step size is one-eighth of the first step size. The primary monitoring information extracted in the first step size is used to extract core violation features and match them with screening conditions to determine the type of violation behavior, and then determine the feature type for extracting secondary monitoring information. The secondary monitoring information removed in the second step size is used to further extract auxiliary violation features.

[0070] Alternatively, a SlowFast network can be used to construct a model for feature extraction from primary and secondary monitoring information. The SlowFast network consists of a Slow Path and a Fast Path, which improves the efficiency and accuracy of video feature extraction by simultaneously processing both low and high frame rate portions of the video.

[0071] S500: When a match is found with the filtering criteria, a call message is sent to call the associated high-level classification feature of the primary classification feature corresponding to the filtering criteria.

[0072] Specifically, when the vehicle monitoring features corresponding to the high-speed vehicle monitoring information collected within a preset area match the screening conditions, it indicates that a certain type of high-speed violation exists on the high-speed vehicle within the preset area. Once the category of the high-speed violation is determined, a request is sent to retrieve the associated high-level classification features of the primary classification features corresponding to the screening conditions to determine the degree of violation of the high-speed vehicle within the preset area.

[0073] S600: The monitoring information is matched according to the level of the association high-level classification features in ascending order to obtain the matching results;

[0074] For example, for speeding violations, detection information is matched based on high-level classification features, i.e., speed is matched. The high-level classification features, in ascending order of severity, are as follows: Violation Level 1: Exceeding the speed limit by less than 10%, no penalty; Violation Level 2: Exceeding the speed limit by 10% to less than 20%, a fine of 50 yuan and 3 demerit points; Violation Level 3: Exceeding the speed limit by 20% to less than 30%, a fine of 50 yuan and 6 demerit points; Violation Level 4: Exceeding the speed limit by 30% to less than 50%, a fine of 200 yuan and 6 demerit points; Violation Level 5: Exceeding the speed limit by 50% to less than 70%, a fine of 1000 yuan and 12 demerit points, and possible license revocation; Violation Level 6: Exceeding the speed limit by 70%, a fine of 2000 yuan and 12 demerit points, and possible license revocation.

[0075] S700: Based on the matching results, identify violation information and provide location feedback for the vehicle.

[0076] The vehicle location feedback process involves using monitoring and data collection equipment to locate violating vehicles and then feeding the location results and violation information back to the traffic management platform. Optionally, vehicle location information may include current location, speed, and expected departure location. This location feedback facilitates timely intervention and penalties for violations by management personnel, strengthening the supervision and management of violations and ultimately improving highway safety and traffic efficiency.

[0077] Furthermore, in one feasible embodiment, the steps further include:

[0078] S710: Based on the aforementioned high-speed violation monitoring database, conduct a risk level assessment of violation events to determine the violation risk level;

[0079] S720: Set a mandatory violation flag for violation identification features whose violation risk level exceeds the preset mandatory threshold;

[0080] S730: When monitoring information triggers a mandatory violation flag, a warning message is sent.

[0081] Optionally, the risk level assessment is performed in two steps. First, a baseline risk index is determined based on the type of violation. For example, the baseline risk index for speeding is higher than that for crossing the line, and the baseline risk index for illegal parking is higher than that for speeding. Next, an additional baseline risk index is determined. For example, for speeding, a lower speeding level results in a higher additional risk index. Finally, based on a preset mandatory threshold and the violation risk level, the risk level of the violation is assessed, where the violation risk level is equal to the sum of the baseline risk index and the additional baseline risk index.

[0082] For example, if the baseline risk index for illegal parking is 10 and the preset mandatory threshold is 8, then a mandatory violation flag will be set for illegal parking and a warning message will be triggered. If the baseline risk index for speeding violations is 5, the additional risk index is 0 when the speeding violation level is 1, and 3 when the speeding violation level is 4. Therefore, when a speeding violation reaches violation level 4, a mandatory violation flag will be set and a warning message will be triggered.

[0083] By setting preset mandatory thresholds and flexibly setting alarm lines for different violations, the system improves its adaptability to different traffic rules under complex road conditions and enhances the accuracy of violation detection and early warning.

[0084] In summary, the high-speed violation detection method provided by this invention has the following technical effects:

[0085] This application constructs a highway violation monitoring database by analyzing highway violation cases; it uses a pre-set clustering algorithm to cluster the database and classifies it according to the clustering results. The classification information includes multi-level classifications, which characterize the different levels of recognition of highway violation events. Primary classification features are extracted from the classification information and matched with acquisition devices according to the classification information. These primary classification features are embedded into the matching acquisition devices as filtering conditions. The matching acquisition devices are activated to collect and monitor highway vehicles within a preset area, and the monitoring information is traversed and compared using the filtering conditions. When a match is found, a call is sent to invoke the associated high-level classification features of the primary classification features corresponding to the filtering conditions. The monitoring information is matched using the associated high-level classification features in ascending order of level to obtain matching results. Based on the matching results, violation information is identified, and vehicle location feedback is provided. This approach has the technical effect of improving the ability to monitor traffic violations, reducing false positives and false negatives, improving the accuracy of judgment, and accurately determining the level of violation.

[0086] Example 2

[0087] Based on the same concept as the high-speed violation detection method in the above embodiment, such as Figure 3 As shown, this application also provides a high-speed violation detection system, the system comprising:

[0088] Monitoring library module 11 is used to build a highway violation monitoring library by analyzing highway violation cases;

[0089] The clustering and classification module 12 is used to cluster the high-speed violation monitoring database using a preset clustering algorithm and classify the high-speed violation monitoring database according to the clustering results. The classification information includes multi-level classification, which is used to characterize the degree of recognition of different features of high-speed violation events.

[0090] The device configuration module 13 is used to extract primary classification features from classification information, match collection devices according to classification information, and embed the primary classification features as filtering conditions into the matching collection devices;

[0091] The monitoring and acquisition module 14 is used to start the matching and acquisition device to collect and monitor high-speed vehicles in the preset area, and to traverse and compare the monitoring information using the filtering conditions.

[0092] The grade feature module 15 is used to send a call message when it matches the filtering conditions, and call the associated high-grade classification features of the primary classification features corresponding to the filtering conditions.

[0093] The grade matching module 16 is used to match the monitoring information according to the grade size from small to large using the associated high-grade classification features to obtain the matching results;

[0094] The location feedback module 17 is used to identify violation information and provide location feedback for the vehicle based on the matching results.

[0095] Furthermore, the monitoring database module 11 also includes a feature matching unit, which is used to perform violation identification feature matching using the omission identification cases to determine the omission violation identification features;

[0096] Furthermore, the clustering and classification module 12 also includes a threshold grading unit, which is used to divide the threshold influence range of the core violation features and auxiliary violation features respectively, and determine the multi-level identification threshold;

[0097] It should be understood that the focus of the embodiments mentioned in this specification is their difference from other embodiments. The specific embodiments in the aforementioned Embodiment 1 are also applicable to the high-speed violation detection system described in Embodiment 2. For the sake of brevity, they will not be elaborated further here.

[0098] It should be understood that the embodiments disclosed in this application and the above description can enable those skilled in the art to implement this application. At the same time, this application is not limited to the embodiments mentioned above; obvious modifications and variations to the embodiments mentioned in this application also fall within the scope of this application's principles.

Claims

1. A high-speed violation detection method, characterized in that, include: By analyzing highway violation cases, a highway violation monitoring database was constructed, including: The semantic analysis module performs semantic analysis on the high-speed violation rules and extracts violation discrimination features. Collect cases that were missed in the current violation identification system's case database; The aforementioned cases of oversight are used to perform violation identification feature matching to determine the oversight violation identification features, and the high-speed violation monitoring database is constructed based on the oversight violation identification features. The collection of overlooked cases from the current violation identification system's case database includes: Analyze the current violation detection system to obtain the identification error rate. Based on a preset error rate threshold, the error rate of the method is screened to identify cases of error violation. Collect a violation record database, and compare the violation record database with the identification record data of the current violation identification system to determine missing cases; The error violation cases and missing cases are integrated to obtain the omission identification cases; The high-speed violation monitoring database is clustered using a preset clustering algorithm, and then classified according to the clustering results. The classification information includes multi-level classification, which is used to characterize the degree of recognition of different features of high-speed violation events. Primary classification features are extracted from the classification information, and the primary classification features are used as filtering conditions to be embedded into the matching and acquisition devices according to the classification information. The matching and acquisition device is activated to collect and monitor high-speed vehicles within a preset area, and the monitoring information is traversed and compared using filtering conditions. When a match is found with the filtering criteria, a call message is sent to call the associated high-level classification feature of the primary classification feature corresponding to the filtering criteria. The monitoring information is matched by using the high-level classification features of the association, in ascending order of level, to obtain the matching results; Based on the matching results, the location of the vehicle is identified and feedback is provided based on the violation information.

2. The method as described in claim 1, characterized in that, The high-speed violation monitoring database is classified according to the clustering results, wherein the classification information includes multi-level classification, including: The high-speed violation monitoring database is classified according to the clustering results to determine the classification information, which includes violation type and violation characteristics. Based on the aforementioned violation characteristics, an impact analysis is conducted to determine the core violation characteristics and auxiliary violation characteristics. The threshold influence ranges of the core violation features and auxiliary violation features are divided respectively to determine multi-level identification thresholds; The core violation features, auxiliary violation features, and multi-level identification thresholds are sorted in order from core violation features to auxiliary violation features, and from the largest identification threshold range to the smallest identification threshold range. The violation features in the classification information are graded using the sorting results.

3. The method as described in claim 1, characterized in that, The matching and acquisition device is activated to collect and monitor high-speed vehicles within a preset area. When the matching and acquisition device is a camera, the process includes: The video is captured by a camera to collect video of vehicles passing through a preset area, and monitoring video information is obtained. The monitoring video information is preprocessed by the video processing extraction module, and video extraction is performed according to a preset step size to obtain the first step length extracted video and the second step length extracted video, wherein the first step length is greater than the second step length. The video extracted in the first step is used as primary monitoring information, and the video extracted in the second step is used as secondary monitoring information. Feature extraction is performed on the primary monitoring information, and the feature extraction results are matched with the screening conditions. When the match is successful, feature extraction is then performed on the secondary monitoring information.

4. The method as described in claim 1, characterized in that, Also includes: Based on the aforementioned high-speed violation monitoring database, a risk level assessment of violation events is conducted to determine the violation risk level. For violation identification characteristics whose violation risk level exceeds the preset mandatory threshold, a mandatory violation label is set; When monitoring information triggers a mandatory violation flag, an early warning message is sent.

5. A high-speed violation detection system, used to perform the method of claim 1, characterized in that, The system includes: The monitoring database module is used to construct a high-speed violation monitoring database by analyzing high-speed violation cases. The clustering and classification module is used to cluster the high-speed violation monitoring database using a preset clustering algorithm and classify the high-speed violation monitoring database according to the clustering results. The classification information includes multi-level classification, which is used to characterize the degree of recognition of different features of high-speed violation events. The device configuration module is used to extract primary classification features from classification information, match acquisition devices according to classification information, and embed the primary classification features as filtering conditions into the matching acquisition devices. The monitoring and acquisition module is used to start the matching acquisition device to collect and monitor high-speed vehicles in a preset area, and to traverse and compare the monitoring information using filtering conditions. The grade feature module is used to send a call message when it matches the filtering conditions, and call the associated high-level classification features of the primary classification features corresponding to the filtering conditions. The grade matching module is used to match the monitoring information according to the grade size from small to large using the associated high-level classification features to obtain the matching result; A location feedback module is used to identify violation information and provide location feedback for the vehicle based on the matching results.

Images