A HYBRID RECOGNITION METHOD THROUGH VIDEO SURVEILLANCE USING COMBINED HARRIS-EDGE

Abstract

This paper discuss Intelligent Transportation System (ITS) which is incorporated with computer vision techniques in recognition of vehicles for typical fast growth cities in development countries. Two algorithm are combined i.e. Harris and Edge Prewitt to increase accuration of vehicles’ detection. Verification is conducted using Support Vector Machine (SVM) and Histograms of Oriented Gradiens (HOG) as its descriptor. Six pieces of experimental videos from a wide variety of road conditions are used for processing. It shows that percentage average for sensitivity is 75.73%, specificity 90.1% precision 56.0% and with accuration of 77.73%.

Keywords

• intelligent transportation system (its), vehicles’ detection, interest point, harris, edge prewitt, hog, svm