Shadow Detection and Extraction from Imagery Based on Total Variation

The shadow phenomenon on the imagery is treated as a special kind of image degradation,then,the following two restrictions: 1) shadow region on the imagery usually has low brightness;2) a great contrast are take into account usually exists between shadow region and non-shadow region on the imagery.On the basis of these analyses and the general principle of total variation,an TV model based algorithm of shadow detection and extraction on imagery is presented.After the experiments with several real images,it shows that the total variation based shadow detection algorithm is valid both for the color images and the black and white images.