A Local Binarization Method And Its Application

With the development of science and technology,digital image processing is being applicated more and more widely,involved in all walks of life ? Image binarization is an important content of digital image processing,an important means to extract target information from image?It is one of the most commonly used method of binarization threshold method to set the threshold value by a certain method to classify all pixels,threshold value method includes global threshold method and local binarization method ? In general,the existing binarization method can deal with a lot of images well,however,due to the use of a unified threshold as a division standard of all pixels in the area of global or local,in the treatment of low contrast image,the traditional global binarization and local binarization method tend to lose part of the target object details,unable to obtain satisfactory processing effect?To solve the problem of low contrast image target extraction,this paper puts forward an improved local binarization method-point by point method,the method sets up a corresponding to each pixel of the image segmentation threshold by adopting the idea of point binarization,whichcan show the detail of the image more fully and extract target information of low contrast image more accurately?Compared with traditional binarization methods,target extraction on low contrast image advantage makes the point by point method in image defect detection has great potential for application?Point by point method,of course,there are also some disadvantages,this article through the concrete analysis points out the deficiency and improvement direction of point by point method?Finally,through the TFT-LCD Mura defect detection project,the practical application effect of the point by point method is shown in the aspect of low contrast image flaw detection.