Varying Area Based Shadow Detection in Video Sequences

Shadows can be occurred when an object partially or totally obstructs direct light from a source of illumination. Shadows have their-own merits and demerits. Sometimes it will provide useful information about objects. But the presence of shadows may reduce the reliability of many computer vision algorithms, including segmentation, object detection, tracking, etc. Therefore, to improve the performance of Computer vision processes, shadow detection and removal is an important pre-processing step. Moving shadow detection has an important role in computer vision applications including video conference, vehicle tracking, and three-dimensional (3-D) object identification. This paper proposes a simple and efficient method to detect the moving shadow from video sequences. The basic principle employed in the approach is the property that the area of shadows will be changing from one frame to the other frames of the video. The performance of the technique is demonstrated and it is found to be efficient in detection and removal of shadows.