Research On Image Enhancement Algorithm Based On Pulse Coupled Neural Networks

As one of the important means of transmitting information and media, image contains very important information. Digital image processing technique develops very rapidly, more widely applied. Digital image processing technique is widely used in many fields- remote sensing, biomedicine, communication, industry, aero-space, military affairs, security etc, brings great economic and social benefits.Image enhancement is an attractive area of digital image processing technique. Its purpose is to handle image processing for a particular application is more suitable than the original image characteristics of human visual recognition system or machine. Application of image enhancement algorithm is targeted, there is no general enhancement algorithms. The result of image enhancement including evaluation of subjective feeling, as well as the evaluation parameters based on objective evaluation.The paper introduces the basic algorithm for image enhancement and explains the principles of these algorithms, presents the application of the best in their respective scenes, and makes simulation experiment about median filtering and histogram equalization firstly. Then studies the artificial neural networks(ANN) and analyzes principles and brief analysis of several common principles of neural networks as well as the application of image enhancement. And then presents based on the traditional PCNN model of the image enhancement algorithm contrary to the model and principles of PCNN. Finally presents an improved PCNN model which was designed based on the traditional PCNN model. Integrator leakage of the input domain and connected domain was removed, the model parameters were reduced, the original model in several important features was maintained to some extent. Studies the image enhancement algorithms with an improved PCNN model make simulation experiment about these two algorithms. Experiment results show that the PCNN can effectively remove the noise, the image peak signal to noise ratio(PSNR) has been increased significantly, with the median filter to compare not only the incremental increase in PSNR, and can maintain a good image detail, thus improve the subjective image quality, and made the results better.