Image restoration using Wigner distribution for night vision system

Night vision systems have become an important research area in recent years. Due to variations in weather conditions such as snow, fog, and rain, night images captured by camera may contain high level of noise. These conditions, in real life situations, may vary from no noise to extreme amount of noise corrupting images. Thus, ideal image restoration systems at night must consider various levels of noise and should have a technique to deal with wide range of noisy situations. In this paper, we have presented a new method that works well with different signal to noise ratios ranging from -1.58 dB to 20 dB. For moderate noise, Wigner distribution based algorithm gives good results, whereas for extreme amount of noise 2^nd order Wigner distribution is used. The performance of our restoration technique is evaluated using MSE criteria. The results show that our method is capable of dealing with the wide range of Gaussian noise and gives consistent performance throughout.