Vehicle segmentation from video using color and edges

Intelligent Transportation System (ITS) starts to play more important role in alleviating traffic problems. The key issue lies in the collection of traffic information. This information will be fed into traffic analytic. In this regard, camera sensor is the typical sensor acquiring video signals containing traffic videos. Later on, video processing algorithm such as object segmentation is used to detect moving vehicle objects. Robust vehicle segmentation technique is very crucial as the algorithm must handle the changing environments such as lighting condition and shadow, vehicle dimension and traffic conditions. In this paper, we present the improved color-based statistical background subtraction and edge based information technique. Experimental results showed that our proposed algorithm can improve vehicle segmentation accuracy as it can overcome the shadow problem and works well with low contrast images when compared to other well known methods.