3D Histogram Based Image Processing Algorithm And Its Application In Image Segmentation And Image Enhancement

The image enhancement technology is a large class of the basic image processingtechnologies; its goal is to get more good and useful images by processing the images.There are a lot of image enhancement algorithms. The histogram equalization is one ofthe most commonly used algorithms and one of the most effective methods. Forgray-scale images, there has been a very mature1-D histogram equalization methodbased on the cumulative distribution function (cdf). However, a typical color image hasthree color components: red(R), green (G) and blue (B). The histogram equalization of acolor image is more complex than1-D equalization due to the multidimensional nature ofcolor signal. With the popularity of color image applications, color image enhancementalgorithms are also emerging. Currently, the most common color histogram equalizationmethods are the equalization method based on each RGB color channel, the equalizationmethod based on the three-dimensional joint probability and the equalization methodbased on I component or SI components in HSI color space. However, in the1-Dhistogram equalization, the image gray levels may be excessively merged, leading tolosing detail and causing false contours.This paper studied the related theories and technologies of the color histogramequalization algorithm, and it analyzed the color image enhancement algorithm based on3-D color histogram equalization particularly. The algorithm would present someproblems for dim color images. So, we do the following improvements to solve theproblems.For the false contour in the1-D histogram equalization, the Logarithm histogramequalization algorithm is used instead of1-D histogram equalization method. Theimproved method can reduce gray-scale merger so as to remove false contours greatly.Using the logarithm histogram equalization method for many times can obtain bettereffect of image processing.For images turning pale by the decrease of color-saturations，the paper make someimprovements. Try not to be space processing and do the corresponding processing according to the algorithm in RGB color space. The improvement can make more pixelsof the processed color images keep their color-saturations.Through experiment and analysis, test results with dim color images present a bettereffect of image enhancement.