An Adaptive Linear CCD Image Correction Algorithm Based On Regional Characteristics

Recently, linear CCD (Charge Coupled Device) and its application research havemade remarkable progresses, especially in the fields of line grading and processing, surfacequality assessment, high-speed data acquisition and so on. Domestic and foreign scholarshave undertaken extensive researches on linear CCD imaging technology, making theimaging technology even more perfect as well as the linear CCD measurement technology,high-speed image acquisition and processing system and the image data redundancycompression technology. Based on the analysis of distortion image caused by the mismatchof the athletes speed and the color linear CCD camera in athletics, this paper proposes onelinear CCD imaging correction algorithm based on the regional characteristics.First, this paper discusses the extraction methods and its key technologies of manydifferent features about edge, angle and area, as well as the analysis of advantages anddisadvantages of various methods. Then, this paper introduces the principles based on theregional characteristics, basic methods and the processing procedure, focusing on severalcommon algorithms. Finally, in order to solve the distortion image taken by the color linearCCD camera in the athletics, this paper proposes the linear CCD imaging correctionalgorithm based on the regional characteristics. This method first does the filtering forde-noising to the acquisition image. Then taking use of the frame difference method toremove the redundant information and trim the image, obtaining the threshold by using theiterative method from the trimmed image and even the binary image according to thethreshold. Then segmenting the image using morphological processing methods, we canget the obliquity, the aspect ratio of the athletes head, the corresponding angle conversionfactor and together with the two-dimensional transformation matrix. At last, the corrected color image is obtained after interpolation projection transformation to the trimmed image.During the research process of this paper, large amounts of simulation experiments throughthe MATLAB platform on the PC.This project improves the problem of image distortion caused by the mismatchbetween the camera scan synchronization signal and the athletes’ moving speed. It also canadaptively stretch or compress the distortion image, making it in consistent with humanvisual requirements, thereby improving the accuracy of the end timing.At the last part of this paper, the geometric transformation is taken to the distortionimage. Then the error range of the algorithm is indirectly obtained through the algorithmsimulation analysis. And by adding noise to the image, this paper tests the anti-noise abilityof the algorithm. What’s more, chapter5introduces the evaluation method of imagealgorithm, taking use of the evaluation method for analyzing the correcting effect. Last,summary and outlook is concluded on the subject.