The Digital Image Identification System Of Highway Pavement Disease Application Research

With the rapid development of our country’s economy, the highway industry in China has been into the period of maintenance and reconstruction from the new period. In pavement maintenance work, the original pavement disease information maintenance management department to make maintenance plan scheme is a very key and important basis. If caught early in the disease of problems, and take corresponding measures, it can greatly cut down the cost of pavement maintenance. Traditional detection method based on artificial vision have been not able to meet the needs of road maintenance, and many deficiencies, such as testing personnel safety, the detection efficiency is low, the result is not accurate, affect normal operation of traffic and so on. Pavement disease detection method based on digital image analysis research and related system development has been scrambling to various countries research topic.This paper introduces the highway types, main disease research disease investigation methods are summarized and the hierarchy of evaluation index and crop diseases; In this paper, the road test is Canada developed ARAN multi-functional test car, this paper introduces the automatic test system for highway pavement diseases structure, and the functions and characteristics of each subsystem, and introduces the ARAN9000type multifunctional test car working principle and working process.This paper mainly studied the basic knowledge of the digital image processing, image enhancement of the commonly used methods include enhancing image transform, gray level transformation, histogram transformation. This paper introduces the principle, mathematical model and its steps, from enhancing actual effect was introduced.the grayscale transform and histogram transform to enhance more effective; Analysis compares the image noise cancellation method, and the introduction of spatial domain image sharpening is actually one of the blurred image denoising processing, make it easier for our target signals are identified; Introduces the implementation process of high-pass filter and gradient sharpening, is concluded high-pass filter processing of the image than the gradient sharpening clearer and more effective. Image segmentation from image processing to image analysis is an important link, this paper introduced a concept of image segmentation, this paper introduces the commonly used image segmentation method of theoretical knowledge and practical implementation process; This article analyse the reason of image degradation and the importance of image restoration, defines the basic theory of image degradation system, introduces the continuous function of the degenerate model and the discrete function degradation model two degradation model. Study caused by uniform linear motion blurred images by specific recovery process, and the programs are the VisualC++implementation of image restoration.This paper studied the basic model of BP neural network, using the extraction of road features as the input value of the BP neural network, designed a three layers BP neural network to classify pavement plant disease image recognition, instructions for pavement disease recognition based on BP neural network technology is feasible.Finally this paper Yuqian road engineering example proves that the digital image automatic pavement distress detection than the manual detectioneffect is better, higher recognition rate, faster, more secure.