Moving Object’s Cast Shadow Removal

Abstract

Moving object’s cast shadow removal is the active and challenging research area in computer vision and digital image processing. Removing cast shadow is the primary step for object tracking and segmentation system. Shadow cause false classification and shape distortion. So, the system is proposed to remove the moving object’s cast shadow in the outdoor environment. In the system, background modeling is initially performed by averaging the movement regions. After background modeling, foreground objects are extracted using the background subtraction method. Frame normalization, V channel of HSV color space, Otsu thresholding approach, morphological processing and region properties are employed to detect the cast shadow region. After shadow detection, the shadow is removed based on previous background model. Experimental results show the proposed method can detect and remove cast shadow region properly.