Research Of Image Enhancement And Image Denoising

Images are an important source of information that human beings can gain fromthe outside world, and image processing techniques are closely bound up with theproduction of human life. But an image could be noisy or low contrasted, because of theinfluence of different shot environment, cameras or transmission conditions, which willreduce the quality of the processed images. While there are variety of preprocessingmethods can improve the quality of those images effectively, such as imageenhancement, image denoising and so on. Nowadays, researchers have presented manymethods in image enhancement for different factors which can affect the quality ofimages. More specifically, we can use the traditional filtering methods or methods basedon fractional partial differential equation (PDE) to deal with noise, or we can use themethods based on histogram equalization to deal with contrast enhancement. In thisstudy we have presented two methods to deal with different scenes. The main content ofthis research study is as follow:1) In this study we have analyzed some traditional methods for image enhancementand have compared their advantages and disadvantages in different applications. Themethods, such as spatial filtering, spectral filtering and morphological operations, arethe main focus of this study. Due to their simplicity and efficiency, these methods areoften applied in real world applications.2) In order to cope with complex and difficult applications, this study has proposeda method to deal with color images denoising, which combined PDE and quaternion. Inthis method, the integer order is replaced by fractional order in the energy function,because models PDE-based can comprehensive the advantages of integer order-based;and a quaternion matrix is applied on a color image to deal with the defects ofanamorphose when a color image is processed in RGB three channels independently.The experimental results show that the proposed method produces better results intexture enhancement and noise reduction than some other traditional color imagedenoising methods.3) This study also discusses the research status of histogram equalization methodsfor brightness preservation and has explained their principle and contribution in thedevelopment history of this algorithm. In order to control the intensity of brightnesspreservation and contrast enhancement, we have presented a method of range limited peak-separate fuzzy histogram equalization for image contrast enhancement(RLPSFHE), which is proposed by us. In the RLSFHE, an intensity factor is applied tocontrol the amount of brightness preservation. The experiment results have showed thatthe proposed method can gain a better trade-off between brightness preservation andcontrast enhancement than some other available methods.