TEMPLATE MATCHING TECHNIQUE BASED ON CORRELATION ANALYSIS FOR RECOGNITION OF NIGERIAN VEHICLE LICENSE PLATE USING EDGE DETECTION

Abstract

Population increase in urban areas of Nigeria has led to a corresponding increase in the number of automobiles used for transportation. This has also led to an increase in traffic problems encountered on Nigerian roads. These problems include vehicle tracking and policing, control of restricted areas, vehicle theft, traffic offenders, enforcement of vehicle speed limits, and some more. Vehicle identification by the number plate is an image processing and computer vision technique that extracts the number plate from vehicles on the highway, process it, and identify the vehicle by the number plate. Vehicle license plate detection and recognition using edge detection and Template Matching is an image processing application that is useful for security purpose on Nigerian roads. The system developed here is capable of processing a captured image of a vehicle and process it to extract the license plate and recognize it. Sobel edge detector is used to preprocess the image. Connected components analysis is used to identify plate objects before character segmentation. Template matching technique is used to recognize the vehicle plate characters. In all, a total of 150 vehicle images were used in the experiment. The license plate location stage has a success rate of 98.6%, character segmentation is 98.6%, while plate recognition success rate is 97.9%. the system can be used for real time system and is developed with Matlab.