Research and Implementation of Vehicle Monitoring Methods in Intelligent Bayonet System
|Course||Control Theory and Control Engineering|
|Keywords||vehicle detection three difference method Kalman Filtering license plate location video speed|
The intelligent bayonet monitoring system adopts intelligent technique to monitor specified intersections around the clock, and extracts characteristic information of passing motor vehicles by related image processing technologies, accordingly realizes remote supervision of appointed vehicles. So it can provide with a powerful technical support as far as purifying traffic environment and acting against vehicles crime are concered. The vehicle monitoring system plays an important role in the intelligent bayonet monitoring system. It’s able to detect the passing vehicle, extract the license plate information and determine the vehicle speed. However, this technology also exists some problems, including background updating is easily affected by illumination, license plate location inaccuracy and the speed of the vehicle judgment inaccuracy. To reduce the environmental interference and improve the monitoring accuracy, this paper makes an improvement to the original monitoring method by the utilization of various features and algorithm. The details are as follows:In the vehicle detection algorithm, background updating of this paper is based on the kalman Filter. The kalman Filter is widely used in the background update, which can estimate the current signal while no all the historical data. But in the traditional kalman Filter algorithm, the background updating parameters was fixed. So this algorithm could not adapt to the changing illumination, and background updating needs much more time. To solve above problems, this paper applied a background updating algorithm with adaptive acquired updating parameter by the Otsu threshold method. This parameter can change with the scene illumination. And to save the time of the calculation, the result of the frame difference and background updating are combined. This method was less time-consuming, and the result of the background updating was accurate.In the license plate location algorithm, because the edge of the license plate is very rich and extracted easily, line scan license plate location method based on edge information was widely used. But the horizontal edges of the license plate could interfere the detection result. And the merge process of the suspicious line was very complex. In this paper, to reduce the interference of the horizontal edges, the vertical edges were enhanced and then extracted; to simplify the merger process of the suspicious line, this paper extracted the starting point of the lower boundary suspicious line as the coordinates of the lower left corner. This method improved the positioning accuracy and reduced the time-consuming.In the vehicle speed judgment algorithm, the measurement for the vehicle-speed based on the detection line with the advantages of simple and effective was widely recognized. But the speed detection lines were easily affected by illumination resulting in mission inspection and erroneous inspection. To reduce the impact of weather and illumination, the speed detection lines were widened. This paper solved the problem of the mission inspection by added a trigger detection. The experimental results showed that the vehicle speed measurement accuracy was improved and the missing rate was reduced.Aiming at the application of the improved algorithm in this paper, on the basis of the analysis of the monitoring environment, a video speed measurement system was designed and established. and it realized the monitoring functions, including vehicle detection, license plate location and speed judgment. The experiments show that the monitoring method in this system reduces the environmental interference, improves the monitoring accuracy, and has certain practicability.