An Approach for Recognizing Vehicle Based on Appearance

In Intelligent Transportation System (ITS) vehicle detection from the road image is very important for drivers and pedestrians awareness. Detection of vehicles from the road area is important to prevent the accident or collisions. This paper proposes a method for improving traffic control and management system for different condition of roads. Four basic steps based on appearance are presented in our proposed method. Initially, to segment the object, we use pitch color from the road image. In the next step, labeling and filtering is applied on foreground objects for getting the region of interest (ROI). Then, for extracting features we use Haar-like feature on ROI. Finally, test images are classified by support vector machine (SVM) classifier for recognizing the vehicle, which provides high accuracy with a very fast and reliable detection rate. Various road images were used with a variety of conditions to test proposed method and results are presented to prove it effectiveness. Moreover, it will satisfy the needs of ITS.