A Highway Vehicle Detection System Of Green Channel Based On Radar Echo

With the national "Green Channel"policy sustainable and deep development Evading road tolls has grown from a single act into a group phenomenon in recent years.However, a traditional test method is not only time-consuming, but also inefficient; At the same time ,it is easy to miss fake "green channel"vehicles a lot .Therefore, how to design a fast , efficient, non-destructive testing system in the highway is of importace in the aspect of theoretical and practical significance.Through manual existing drawbacks and compare various of non-destructive detect techniqus in the advantage and disadvantage, It roposed a vehicle detection scheme based on radar echo in the highway of"green channel".The article study focuse on the radar echo in the car and analysis its echo in the prohibited items or not,reasearching some algorithms on the radar echo feature and recognition in detail.for example based on secondary statistics GLCM and image recognition BP algorithm.It achieve radar echo in cars automatic analysis and judged.This paper first describes the feature extraction of radar echo,Then, using a Matlab program analysis the gray-scale of radar echo image to find out some different gray level of concentration and the probability distribution of each gray level. Finally, feature extraction. Its purpose is to image classification.There are a variety of methods to describe image characteristics.among them texture features is an important feature. This feature extraction algorithm is based on secondary statistics thought GLCM,On this basis, extracting the radar echo texture parameters in the different situations as a recognition feature.Second,It is radar echo identification,Image recognition is a difficult way in pattern recognition.There are various of algorithms on image recognition.Through analysis in depth multiple image recognition algorithm, radar echo of their own characteristics and site of the actual situation, It introduced neural network in the recognition system of radar echo.BP alogrithm is a key,improving some drawbak,like slow slow convergence, local minimum and determine hidden layers and hidden nodes difficult.Follow,creating a structure model of BP neural network to identify radar echo.Finally, Many experiments and Pratical applicatios show that the system have good good discernment and classification performance. Recognition rate of 75% or more. The system can realize vehicles detection intelligent and recognition, improve test accuracy, reduce inspector labor intensity.