Dynamic Shadow Detection and Removal in Evening Time for Vehicle Counting System

Shadow detection and removal are an important process for the further progression in computer vision method, i.e., tracking, recognition, classification, etc. Vehicle counting system in traffic management is not able to count exact number of vehicles without shadow detection and removal especially during evening time. Due to strength of shadow in evening time, the shadow detection is quite difficult in traffic especially in evening time. Many researchers have published paper on shadow detection, but for daylight/noon time dataset. Hence, need arises for determination of shadow detection in evening time. In this paper, HSV color space-based non-deterministic model with background updating concept is proposed for shadow detection during evening traffic dataset. Investigational outcomes suggest that the proposed method has been able to detect dynamic shadows accurately.