The Application Of Digital Image Processing In Intelligent Transportation System

With the rapid development of social industrialization, urbanization, traffic flowincreases dramatically. The impact of traffic congestion, traffic accidents, environmentalpollution and other traffic disasters on the human environment and socio-economicdevelopment are increasing day by day and becoming one of the social nuisances that isdifficult to eradicate. How to manage road operations effectively and ensure the security andefficiency of vehicles has become an important topic in traffic management. Recently,Intelligent Transportation Systems is the frontier research in the field of electronicinformation technology applied in transportation. The application of image processingtechnology in intelligent transportation system is an important cutting-edge research in thefield of intelligent transportation systems, and has an extremely important theoretical andpractical value.Crack detection is the basis of road safety and security. The first part of this articlemainly aimed at the study of algorithm on pavement crack image, and proposed afeature-based crack recognition and extraction algorithm which presented a crack detectionalgorithm based on gray correlation from the perspective of gray scale and continuity of thecrack. For these disturbed factors of pavement holes, defacements and other noise interference,cracks extraction were performed by the following three steps:(1) analyze the cracks’characteristics of the acquired pavement images, build a similarity function of the gray valueand extract suspected information of the cracks according to the similarity comparison;(2)remove the false cracks by the computed measurement of connected domain and transformthe crack detection problem into graph theory one;(3) extract the true cracks through spectralclustering algorithm. After testing many images of concrete pavement with complicatedbackground, it shows that the proposed method can extract the pavement cracks’characteristics actually and have good practicability.Road vehicle detection is essential in terms of safety and high-speed operation ofvehicles. Image registration plays an important role in vehicle detection system and it is thebasic and key steps in digital image processing. Registration accuracy directly affects the subsequent image processing. The second part mainly focused on the image of two frames ofadjacent pavements through registration algorithm. As the accuracy of the image registrationalgorithm based on the gray information and the selected local matching window, then thepoint which embodied more information aroused our interest to do research in this paper.andthe set of the equal distance from the concentric circles was the representative of our interest.The similarity between the two point judged by MSE was to determine match point, withwhich presented the registration algorithm based on two adjacent frames image. Thisalgorithm not only enhanced high precision registration of gray but also improved theaccuracy of location registration and reduced the amount of calculation as well.