Grayscale And Color Image Contrast Enhancement Of Pde Research

Contrast enhancement is an important part in image processing technology. The purpose is to obviously improve the visual effects and offer resulted images that are more suitable to analyze by altering the distribution of the image gray levels to spread the gray-value dynamic range and enrich the image gray levels. Histogram equalization is an important branch of image contrast enhancement.This paper introduces the foundational principles and current research status of image contrast enhancement. On the basis of studying the algorithm of shape-preserving local contrast enhancement presented by V. Caselles, an improved algorithm is proposed in order to overcome the difficult selection of evolution time step in original PDE and avoid the "spots effects" induced by the layer-separating local contrast enhancement. The improved algorithm firstly modifies the original PDE model of shape-preserving local contrast enhancement of gray image to simplify the selection of evolution time step. Next, the modified model is extended to a more general form to adapt to any specified stretch function for global and local contrast enhancement. A piecewise linear stretch function designing method is suggested to achieve better histogram equalization than accumulated histogram as the gray transform function.This paper proposes a relaxed thresholding scheme followed by a detailed analysis of the cause of the "spots effects" arised in the local contrast enhancement by means of layer-separating sub-images. The new scheme adjusts the connected components with small area of which values are close to the threshold value to the more relevant layer, therefore the undesired spots are effectively avoided.V. Caselles et al presented a shape-preserving color image enhancement by applying the shape-preserving local enhancement to the intensity component of the color image. It is a common method to directly extend the contrast enhancement techniques of gray image to color image. This kind of method applies the enhancement algorithm of gray image to the intensity component of the color image, which is easy to implement: However,it neglects the important fact that intensity component can not preserve the shape of the objects of the original color image in the strict sense, therefore, shape distortion and color distortion may appear in the enhanced result. Based on the research of V. Caselles et al, this paper analyses the cause of the distortion of color image enhancement, and propose an improved shape-preserving color enhancement method. The improved method uses a scalar image which can keep the level sets information of input color image as the intensity components; and then enhance its contrast by the shape-preserving local contrast enhancement method; finally, the enhanced color image is obtained with the original chromaticity component. Experiments of both artificial and natural color images show that the proposed method is hot only efficient for color enhancement but also better than existed methods in the fidelity of color and shape.