The Simulation of Adaptive Optical Image Even and Pulse Noise and Research of Image Quality Evaluation
Wen Changli,Xu Yuannan,Xu Rong,Liu Changhai,Men Tao,Niu Wei
DOI: https://doi.org/10.1117/12.2031727
2013-01-01
Abstract:As optical image becomes more and more important in adaptive optics area, and adaptive optical telescopes play a more and more important role in the detection system on the ground, and the images we get are so many that we need find a suitable method to choose good quality images automatically in order to save human power, people pay more and more attention in image’s evaluation methods and their characteristics. According to different image degradation model, the applicability of different image’s quality evaluation method will be different. Researchers have paid most attention in how to improve or build new method to evaluate degraded images. Now we should change our way to take some research in the models of degradation of images, the reasons of image degradation, and the relations among different degraded images and different image quality evaluation methods. In this paper, we build models of even noise and pulse noise based on their definition and get degraded images using these models, and we take research in six kinds of usual image quality evaluation methods such as square error method, sum of multi-power of grey scale method, entropy method, Fisher function method, Sobel method, and sum of grads method, and we make computer software for these methods to use easily to evaluate all kinds of images input. Then we evaluate the images’ qualities with different evaluation methods and analyze the results of six kinds of methods, and finally we get many important results. Such as the characteristics of every method for evaluating qualities of degraded images of even noise, the characteristics of every method for evaluating qualities of degraded images of pulse noise, and the best method to evaluate images which affected by tow kinds of noise both and the characteristics of this method. These results are important to image’s choosing automatically, and this will help we to manage the images we get through adaptive optical telescopes base on the ground.