Shadow Detection in Single Image using Color Spaces

Abstract: A shadow basically is a sort of image produced when light is blocked. A shadow generally takes up most of the space behind an opaque item with light right in front of it. Shadows often introduce errors in the performance of computer vision algorithms, such as object detection and tracking. For better performance, there is need to detect shadow from image and remove it to get shadow free image. This paper proposes a simple method to detect shadow from images. In this paper LAB color space is used to detect shadow. Here L means lightness, A and B are two different color dimensions. Working with the Lab color space includes all of colors in the spectrum, as well as colors outside of human perception. This paper discusses all the steps to be carried out in achieving the desired objective i.e to detect the shadow from image by using some simple steps.

Keywords: Shadow detection, shadow removal, LAB color space, illumination.