Vehicle identification system and method based on big data

Abstract

The invention discloses a vehicle identification system and method based on big data. The system comprises a data acquisition module used for acquiring image data, video data, radar data, laser radar data, infrared sensor data, GPS data, speed sensor data, accelerometer and gyroscope data, environment sensor data, vehicle-mounted sensor data and sonic sensor data; according to the vehicle identification system based on big data, through data fusion of various sensors, the identification precision and robustness are improved, compared with a traditional single sensor system, multi-source data can keep efficient identification in a complex environment, image data can be supplemented under low visibility and severe weather, the influence of factors such as illumination changes is reduced, and the vehicle identification efficiency is improved. The image enhancement and noise suppression technology further improves the data quality, ensures that a stable and accurate recognition result can still be provided in various complex scenes, and is widely applied to traffic management scenes such as urban roads and parking lots.

Description

Technical Field

[0001] The present invention relates to the technical field of information statistics, and in particular to a vehicle identification system and method based on big data. Background Art

[0002] Among the many traffic management and security monitoring technologies, vehicle recognition technology based on image processing and video analysis has made significant progress. Image data collected by cameras can realize functions such as vehicle identification, tracking, and license plate recognition. It is widely used in scenarios such as road monitoring, parking lot management, and traffic flow analysis. However, image data is easily interfered with in complex environments, especially in low light, many obstructions, or bad weather, resulting in reduced recognition accuracy.

[0003] To overcome these limitations, more and more researchers and engineers have begun to explore vehicle recognition systems that integrate multiple sensors. These systems combine data from multiple sensors, such as radar, laser radar (LiDAR), infrared sensors, GPS, accelerometers, etc., and improve the accuracy and robustness of vehicle recognition through data fusion technology. For example, radar and lidar sensors can provide stable distance and speed information in low visibility or poor lighting conditions, while infrared sensors can help identify the outline of vehicles at night or in bad weather.

[0004] In addition, with the help of big data analysis technology, valuable traffic patterns and trends can be extracted from massive data, providing more accurate traffic flow forecasts, traffic accident analysis and road safety warnings. The use of big data not only optimizes the accuracy of vehicle identification, but also enables traffic managers to grasp traffic conditions in real time and make more scientific and effective decisions.

[0005] Therefore, a multi-sensor integrated vehicle identification system based on big data came into being. By combining image data with other sensor data and applying deep learning and big data analysis methods, the accuracy of vehicle identification and the overall performance of the system can be significantly improved, promoting the further development of intelligent transportation systems. Summary of the Invention

[0006] The main purpose of the present invention is to provide a vehicle identification system and method based on big data, which can effectively solve the abnormal problem of being unable to provide stable distance and speed information under conditions of poor visibility or light.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: a vehicle identification system based on big data, the system comprising:

[0008] Data acquisition module: used to collect image data, video data, radar data, lidar data, infrared sensor data, GPS data, speed sensor data, accelerometer and gyroscope data, environmental sensor data, vehicle-mounted sensor data, acoustic wave sensor data and other data;

[0009] Data processing module: used for preprocessing, feature extraction and fusion of the image data, video data and sensor data;

[0010] Vehicle identification module: Identify vehicle type, brand, color and license plate number information based on deep learning algorithms;

[0011] Data storage module: used to store vehicle identification results and data;

[0012] Data analysis module: used to perform big data analysis on vehicle identification results, generate statistical reports and support real-time monitoring.

[0013] Preferably, the data acquisition module includes a surveillance camera, which is located at key locations of traffic arteries and parking lots.

[0014] Preferably, the data processing module adopts image enhancement technology, noise suppression technology and target detection algorithm.

[0015] Preferably, the vehicle identification module performs vehicle feature learning based on a convolutional neural network (CNN) and automatically recognizes license plate numbers in combination with license plate recognition technology.

[0016] Preferably, the data storage module adopts a distributed storage method.

[0017] Preferably, the data analysis module is based on a big data analysis algorithm and is capable of discovering traffic flow patterns, vehicle type distribution, abnormal events and trends from historical data.

[0018] In addition, the present invention also provides a method for a vehicle identification system applied to the above-mentioned big data, the method comprising:

[0019] Step 1: Use sensors and cameras to collect vehicle image data, video data, radar data, lidar data, infrared sensor data, GPS data, speed sensor data, accelerometer and gyroscope data, environmental sensor data, vehicle-mounted sensor data, and acoustic sensor data in real time;

[0020] Step 2: Preprocessing the collected image data and video data to improve data quality;

[0021] Step 3: Analyze the image data using a deep learning model and a convolutional neural network (CNN) to identify vehicle type, brand, color, and license plate number information;

[0022] Step 4: Match the recognition result with the vehicle information in the storage database and save the recognition result;

[0023] Step 5: Use big data analysis methods to conduct statistical analysis on the recognition results, generate traffic reports, and support real-time monitoring.

[0024] Preferably, in step 2, when pre-processing the image data, image denoising, contrast enhancement and light interference removal techniques are adopted.

[0025] Preferably, in step three, a convolutional neural network (CNN) is used to extract and classify features of the vehicle image, and the license plate recognition algorithm is combined to automatically recognize the license plate number.

[0026] Preferably, in step five, the big data analysis method includes but is not limited to in-depth analysis of vehicle flow and traffic accident rate information, and providing decision support for traffic management departments.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] 1. In the present invention, the big data-based vehicle recognition system greatly improves recognition accuracy and robustness by fusing data from multiple sensors. Compared with traditional single-sensor systems, multi-source data can maintain efficient recognition in complex environments and supplement image data in low visibility and bad weather, thereby reducing the impact of factors such as lighting changes. Image enhancement and noise suppression technologies further improve data quality, ensuring that stable and accurate recognition results can still be provided in various complex scenarios. It is widely used in traffic management scenarios such as urban roads and parking lots.

[0029] 2. In the present invention, by adopting convolutional neural networks (CNN) for deep learning and vehicle feature extraction, the vehicle type, brand, color and license plate number can be efficiently identified. Combined with license plate recognition technology, the recognition accuracy can be further improved. The deep learning algorithm has strong adaptive capabilities and can continuously optimize the recognition model from historical data to adapt to changes in different scenarios. The system can achieve real-time monitoring and rapid identification, helping traffic management departments to improve the level of intelligent management.

[0030] 3. In the present invention, the data analysis module provides decision support for traffic management through in-depth analysis of a large amount of vehicle identification data. It can discover traffic flow patterns, abnormal events and trends, help traffic management departments optimize signal control, improve traffic planning, and predict traffic demand. The system not only improves the efficiency and accuracy of traffic management, but also provides a scientific basis for traffic accident prevention and urban traffic planning, and helps the construction of smart cities. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 is a system block diagram of the present invention;

[0032] Figure 2 Flowchart of the present invention. DETAILED DESCRIPTION

[0033] In order to make the technical means, creative features, objectives and effects achieved by the present invention easier to understand, the present invention is further described below in conjunction with specific implementation methods.

[0034] A vehicle identification system based on big data, the system comprising:

[0035] Data acquisition module: used to collect image data, video data, radar data, lidar data, infrared sensor data, GPS data, speed sensor data, accelerometer and gyroscope data, environmental sensor data, vehicle-mounted sensor data, acoustic wave sensor data and other data;

[0036] A variety of sensors and devices acquire various types of data in real time, including images, videos, radar, lidar, infrared sensors, GPS, speed sensors, accelerometers, gyroscopes, environmental sensors, vehicle sensors, and acoustic sensors. This data is synchronously collected and transmitted to the processing unit for subsequent data processing, analysis, and decision-making modules. Through precise data synchronization and sensor fusion, the system can comprehensively and real-timely perceive environmental information, providing the basic data support for subsequent identification, decision-making, and control.

[0037] Data processing module: used for preprocessing, feature extraction and fusion of image data, video data and sensor data;

[0038] After receiving images, videos, and sensor data from the data acquisition module, the system first preprocesses this data, including denoising, enhancement, correction, and standardization, to improve data quality and usability. Next, it extracts important visual features from images and videos using feature extraction algorithms (such as edge detection and key point recognition), while also extracting physical features such as motion, speed, and direction from the sensor data. Finally, it fuses this data from different sensors, combining multimodal information to generate more accurate and comprehensive environmental perception results, providing reliable basic data for subsequent recognition, decision-making, and control.

[0039] Vehicle identification module: Identify vehicle type, brand, color and license plate number information based on deep learning algorithms;

[0040] First, a clear vehicle image is obtained from a camera or video stream through an image processing algorithm. Then, a pre-trained deep learning model (such as a convolutional neural network (CNN)) is used to classify and analyze the image to identify the type, brand, and color of the vehicle. For license plate number recognition, the module will accurately find the license plate area through a license plate positioning algorithm and use OCR (optical character recognition) technology to extract the license plate number information. Through a multi-level deep learning model, the image and features are gradually analyzed and optimized. The module can efficiently and accurately output various types of vehicle information.

[0041] Data storage module: used to store vehicle identification results and data;

[0042] Responsible for storing the results generated by the vehicle identification module (such as vehicle type, brand, color, license plate number, etc.) and related sensor data (such as images, videos, sensor readings, etc.) in a database or data warehouse. First, the module will format the recognition results to ensure data standardization and consistency. Then, the data is marked by timestamp or event and stored in a suitable data structure (such as a relational database, NoSQL database or cloud storage). In addition, the module will regularly back up data to ensure data security and recoverability. Through efficient data storage and management, the system can support subsequent query, analysis and report generation.

[0043] Data analysis module: used to perform big data analysis on vehicle identification results, generate statistical reports and support real-time monitoring.

[0044] By performing big data processing and analysis on the stored vehicle identification results and other related data, valuable trends, patterns and statistical information can be extracted. First, the module aggregates, filters and classifies a large number of identification results, performs data cleaning and preprocessing, and then uses data mining and machine learning algorithms to conduct in-depth analysis to generate statistical reports such as vehicle flow, vehicle model distribution, and license plate recognition accuracy. At the same time, the module supports real-time monitoring and provides immediate feedback and alerts through dynamic analysis of real-time data streams to help real-time decision-making. The analysis results and reports can be displayed through a visual interface to facilitate users to view data and support decision-making.

[0045] The data acquisition module includes surveillance cameras, which are located at key locations of traffic arteries and parking lots.

[0046] By placing surveillance cameras at key locations on major traffic arteries and parking lots, video image data is collected in real time. These cameras continuously shoot and transmit images or video streams to the data processing system according to preset time intervals or trigger conditions. The cameras can be equipped with high-definition, infrared or night vision functions as needed to ensure stable operation under different lighting and weather conditions. The collected image data will be transmitted to the data processing module for further analysis and processing. At the same time, sensor data (such as temperature, humidity, vehicle speed, etc.) may also be collected and transmitted synchronously through the module to form a comprehensive traffic and environmental data source.

[0047] The data processing module uses image enhancement technology, noise suppression technology and target detection algorithm.

[0048] The data processing module first receives the original image or video data from the data acquisition module, then applies image enhancement techniques (such as contrast enhancement and brightness adjustment) to improve the image quality, making the image clearer under different environmental conditions. Then, it uses noise suppression techniques (such as filtering algorithms) to remove noise and interference in the image, further improving the signal-to-noise ratio of the image. Finally, the module uses target detection algorithms (such as YOLO and Faster R-CNN) to analyze the processed image, identify targets such as vehicles and license plates, and extract valuable information to provide accurate input for subsequent vehicle recognition and data analysis.

[0049] The vehicle identification module learns vehicle features based on convolutional neural networks (CNN) and automatically recognizes license plate numbers in combination with license plate recognition technology.

[0050] The vehicle identification module uses convolutional neural networks (CNN) to perform deep learning on the collected image data and automatically extract key features of the vehicle, such as appearance, shape, color, etc. CNN learns high-level feature representations of the vehicle through multi-layer convolution and pooling operations. Subsequently, combined with license plate recognition technology (such as OCR technology or license plate recognition dedicated network), the module can accurately locate and identify the license plate from the image and extract the license plate number. The entire process includes image preprocessing, feature extraction, license plate area positioning and character recognition, and ultimately generates complete vehicle identification information (including vehicle features and license plate number), providing basic data for subsequent data storage and analysis.

[0051] The data storage module adopts distributed storage.

[0052] The data storage module adopts a distributed storage method to distribute and store the collected vehicle identification data, images, videos and related analysis results on multiple servers or storage nodes to improve the data storage capacity, access speed and fault tolerance of the system. The data is first processed and divided into multiple data blocks, and then these data blocks are distributed to different storage nodes through a distributed file system (such as HDFS, Ceph, etc.). Each node is responsible for storing a part of the data, and the reliability and high availability of the data are ensured through data redundancy and backup mechanisms. When accessing data, the system automatically locates the corresponding node according to the request and quickly returns the required data, ensuring the efficiency of real-time query and large-scale data storage.

[0053] The data analysis module is based on big data analysis algorithms and can discover traffic flow patterns, vehicle type distribution, abnormal events and trends from historical data.

[0054] The data analysis module is based on big data analysis algorithms. It first obtains historical traffic data, vehicle identification information, event logs and other data sources from the distributed storage system. After data cleaning and preprocessing to remove invalid information and abnormal data, it uses machine learning, cluster analysis, time series analysis and other methods to deeply explore the traffic flow patterns, vehicle type distribution and change trends in the data. The module can identify peak traffic hours and common traffic bottleneck areas, and use anomaly detection algorithms to promptly detect abnormal events such as traffic accidents and illegal parking. The analysis results will generate visual reports and prediction models to provide a basis for traffic management and decision-making, and support real-time adjustment and optimization of traffic scheduling.

[0055] In addition, the present invention also provides a method for applying to the above-mentioned vehicle identification system based on big data, the method comprising:

[0056] Step 1: Use sensors and cameras to collect vehicle image data, video data, radar data, lidar data, infrared sensor data, GPS data, speed sensor data, accelerometer and gyroscope data, environmental sensor data, vehicle-mounted sensor data, and acoustic sensor data in real time;

[0057] Step 2: Preprocess the collected image data and video data to improve data quality;

[0058] Step 3: Analyze the image data using a deep learning model and a convolutional neural network (CNN) to identify vehicle type, brand, color, and license plate number information;

[0059] Step 4: Match the recognition result with the vehicle information in the storage database and save the recognition result;

[0060] Step 5: Use big data analysis methods to conduct statistical analysis on the recognition results, generate traffic reports, and support real-time monitoring.

[0061] In step 2, when preprocessing the image data, image denoising, contrast enhancement and light interference removal techniques are used.

[0062] During the image data preprocessing stage, image denoising techniques, such as median filtering or Gaussian filtering, are first applied to remove noise and interference from the image, improving image quality. Next, contrast enhancement (such as histogram equalization or adaptive contrast enhancement) is used to improve the clarity of image details, making the target object more visible, especially in low-light or blurred environments. Finally, illumination interference removal techniques, such as illumination normalization or brightness adjustment, are used to eliminate image deviations caused by illumination changes, ensuring that the image remains consistent and stable under different lighting conditions. This series of preprocessing steps ensures high-quality and stable input images, providing more accurate image data for subsequent target detection and analysis.

[0063] In step three, a convolutional neural network (CNN) is used to extract and classify vehicle images, and the license plate recognition algorithm is combined to automatically recognize the license plate number.

[0064] In this workflow, a convolutional neural network (CNN) is first used to extract features from the input vehicle image. Through multiple layers of convolution, pooling, and fully connected layers, the CNN automatically learns and extracts key vehicle features, such as appearance, color, and brand, and then classifies the vehicle. The extracted features are then passed through a classification network to determine the vehicle type or status. The license plate area is then located in the vehicle image, and a license plate recognition algorithm (such as optical character recognition (OCR) technology or a specific license plate recognition network) is applied to further process the license plate and identify the license plate number. Through image preprocessing, feature extraction, region localization, and character recognition, the system automatically identifies vehicle information and associates the license plate number with vehicle features, completing the entire vehicle identification and license plate number extraction task.

[0065] In step five, big data analysis methods include but are not limited to in-depth analysis of vehicle flow and traffic accident rate information, and providing decision support for traffic management departments.

[0066] The big data analysis method first collects information such as vehicle flow and traffic accident rates from multiple data sources (such as traffic sensors, cameras, GPS data, accident reports, etc.), and then removes outliers and invalid data through data cleaning and preprocessing. Then, big data analysis techniques such as cluster analysis, regression analysis, time series prediction and machine learning algorithms are used to deeply explore the fluctuation trends of vehicle flow, high-incidence periods and areas of traffic accidents, identify traffic patterns and potential risk factors, and present the analysis results through data visualization to help traffic management departments monitor traffic conditions in real time, predict traffic peak hours and accident risks, and provide data support and decision-making basis for formulating traffic control strategies, optimizing transportation infrastructure, and improving safety.

[0067] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A vehicle identification system based on big data, characterized in that: The system includes: Data acquisition module: used to collect image data, video data, radar data, lidar data, infrared sensor data, GPS data, speed sensor data, accelerometer and gyroscope data, environmental sensor data, vehicle-mounted sensor data, acoustic wave sensor data and other data; Data processing module: used for preprocessing, feature extraction and fusion of the image data, video data and sensor data; Vehicle identification module: Identify vehicle type, brand, color and license plate number information based on deep learning algorithms; Data storage module: used to store vehicle identification results and data; Data analysis module: used to perform big data analysis on vehicle identification results, generate statistical reports and support real-time monitoring.

2. The vehicle identification system based on big data according to claim 1, characterized in that: The data acquisition module includes monitoring cameras, which are located at key locations of traffic arteries and parking lots.

3. The vehicle identification system based on big data according to claim 1, characterized in that: The data processing module adopts image enhancement technology, noise suppression technology and target detection algorithm.

4. The vehicle identification system based on big data according to claim 1, characterized in that: The vehicle identification module learns vehicle features based on a convolutional neural network (CNN) and automatically recognizes license plate numbers in combination with license plate recognition technology.

5. The vehicle identification system based on big data according to claim 1, characterized in that: The data storage module adopts a distributed storage method.

6. The vehicle identification system based on big data according to claim 1, characterized in that: The data analysis module is based on big data analysis algorithms and can discover traffic flow patterns, vehicle type distribution, abnormal events and trends from historical data.

7. A method for a vehicle identification system based on big data according to any one of claims 1 to 6, characterized in that: The method includes: Step 1: Use sensors and cameras to collect vehicle image data, video data, radar data, lidar data, infrared sensor data, GPS data, speed sensor data, accelerometer and gyroscope data, environmental sensor data, vehicle-mounted sensor data, and acoustic sensor data in real time; Step 2: Preprocessing the collected image data and video data to improve data quality; Step 3: Analyze the image data using a deep learning model and a convolutional neural network (CNN) to identify vehicle type, brand, color, and license plate number information; Step 4: Match the recognition result with the vehicle information in the storage database and save the recognition result; Step 5: Use big data analysis methods to conduct statistical analysis on the recognition results, generate traffic reports, and support real-time monitoring.

8. The vehicle identification method based on big data according to claim 7, characterized in that: In step 2, when pre-processing the image data, image denoising, contrast enhancement and light interference removal techniques are adopted.

9. The vehicle identification method based on big data according to claim 7, characterized in that: In step three, a convolutional neural network (CNN) is used to extract and classify vehicle images, and the license plate recognition algorithm is combined to automatically recognize the license plate number.

10. The vehicle identification method based on big data according to claim 7, characterized in that: In step five, the big data analysis method includes but is not limited to in-depth analysis of vehicle flow and traffic accident rate information, and provides decision support for traffic management departments.

Images