Classification of color hazy images

For image dehazing first thing to be analysed that whether the image is hazy or non-hazy and for this one cannot depend on perception of human, so for determining the image state as hazy or non-hazy must be solved first. In this research work a new method is proposed to judge whether an image is hazy or non-hazy and for further classifying the hazy image in some definite categories. Based on the difference between the hazy and non-hazy images, the four parameters i.e. standard deviation, max, min and mean is used for making decision of the category for hazy images and non-hazy images and for further classification, obtaining a high rate of recognition. The proposed new method is very simple, feasible, and effective and it is proved by the experimental results. The proposed new method for real-time and batch processing helps in the judgement of image state which will further help in the process of image dehazing.