High definition intelligent checkpoint system (video mode)

Project Overview
The new generation of high-definition intelligent checkpoint system integrates multiple patented technologies and adopts various IT high-tech technologies, such as video encoding and decoding technology, embedded system technology, storage technology, network technology, intelligent technology, and pattern recognition technology. It can be widely applied in networked intelligent road monitoring systems, security checkpoint systems, security reporting station systems, vehicle violation capture and management systems, highway pre inspection and weighing systems, speeding warning and reminder systems, traffic information release systems, traffic information release guidance, congestion charging systems, and other fields.
Background and Challenges
With the improvement of China's comprehensive strength and national income level, motor vehicles are growing rapidly at a rate of 10% to 20% per year. The pace of road construction is accelerating, and the level of urbanization in the country is also constantly improving. The contradiction between traffic management status and demand is further intensifying, and the number of criminal and public security cases related to traffic is also increasing year by year, especially cases such as causing accidents or driving along highways after committing crimes, stealing motor vehicles, and driving illegally.
In this situation, how to use advanced technological means to improve the level of urban traffic management, suppress traffic accidents, crack down on vehicle related cases, deter criminals, and enhance the comprehensive management level of social security has become an urgent problem that needs to be solved by the public security and transportation departments.
In response to the above challenges, based on rich experience in the field of video surveillance, a high-definition intelligent checkpoint system solution has been launched. The main purpose is to collect vehicle characteristic data such as motor vehicle image capture, vehicle license plate recognition, control and compare alarms, and intercept alarms at inspection stations. The system automatically records the composition, flow distribution, and violations of vehicles operating on the road year-round; Provide important technical means and evidence for the rapid correction of traffic violations, the rapid investigation of traffic accident evasion, motor vehicle theft, and license plate fraud cases; At the same time, it provides important foundation and operational data for traffic management, traffic planning, and road maintenance, playing an important role in urban security and traffic management processes. It is of great significance to solve the shortage of public security forces and improve the level of traffic law enforcement.
Solution
system architecture
The system adopts video detection method, which is a vehicle detection method based on full screen analysis. Its core principle is to use background modeling and motion foreground extraction technology to differentiate the current frame image from the background model to obtain motion foreground pixels, and then process these motion foreground pixels to obtain vehicle information. The system uses 2 million, 3 million, or 5 million CCD high-definition integrated cameras as the collection main body. A single camera of 2 million covers one lane in one direction, a single camera of 3 million covers two lanes in one direction, and a single camera of 5 million covers three lanes in one direction; Synchronize support for flash and LED strobe lights for supplementary lighting. The system structure is simple and easy to install and maintain.
The system mainly consists of a front-end data collection subsystem, a network transmission subsystem, and a back-end management subsystem. The overall structure of the system is as follows:

System Features
1. Full screen analysis technology
The full frame analysis technology combines advanced background modeling and motion foreground extraction techniques internationally, enabling us to effectively extract moving vehicles, non motorized vehicles, and pedestrian targets from videos, avoiding the false detection effects caused by shadows and camera shake.
1) The image processing technology adopted by the system effectively improves the detection rate of targets in low contrast situations and enhances the management capability of highways in adverse weather conditions such as foggy and rainy days.
2) The target tracking technology adopted by the system combines prediction and detection to accurately track the separated moving targets, creating extremely favorable conditions for accurately judging the target's driving trajectory and greatly improving the reliability of target behavior detection.
2. Application of high-density integration technology:
The camera adopts a high-definition CCD+high-definition ISP+high-performance DSP architecture design, integrating core functions such as high-definition video capture, high-definition video processing, and license plate recognition.
Camera front intelligent recognition algorithm, firstly, can share the intelligent calculation pressure of the system; Secondly, the image source required for analysis is closest to the real environment, resulting in more accurate analysis results.
3. Intelligent recognition technology for license plate front-end:
The vehicle license plate automatic recognition algorithm is integrated into the checkpoint capture unit, and the recognition results are directly output by the checkpoint capture unit, which improves the recognition accuracy and response time, reduces the workload of the front-end control host, and has a higher cost-effectiveness.
Under the conditions of no fog, standardized license plate placement, no pollution, and excluding five small vehicles, the system's all-weather license plate recognition accuracy is not less than 95%.

Example of identifying license plate types

4. Intelligent analysis applications and deep data mining:
On the premise of ensuring the accuracy of license plate recognition, a large number of vehicle association records are constructed into a data warehouse and analyzed through data mining techniques on software platforms. More accurate execution of fake license plate vehicle analysis, vehicle correlation analysis, and other technical means that help public security solve cases.
Support the sharing of data and information resources at the district, county, city, and even provincial levels, complete the unified automatic control and alarm of the entire network, vehicle trajectory tracking, data mining of criminal investigation and case solving principles, and central emergency command and mobilization of comprehensive police force for case handling and other business functions.

Vehicle trajectory tracking