Video Processing: Number Plate Extraction From The Moving Vehicle Using Frame Selection

Abstract

In the recent days the vehicles are utilized to a great extent due to increase in human needs and growth in population. Hence, control of vehicles is becoming a serious problem. Automatic vehicle identification systems are used for the purpose of effective control of vehicles. The Automatic Number Plate Extraction (ANPE) system plays a significant role in Transportation System. Automatic Number Plate Recognition System is an image processing technology that identifies vehicles by their number plates. It only considers Indian vehicle for Number Plate Extraction. It does not consider Military vehicles. In existing system the number plate information is extracted from vehicle’s image using morphological operations, canny vertical edge detection thresholding, and connected component analysis. The input image is first preprocessed using iterative bilateral filters and adaptive histogram equalization. In the proposed work the number plate extracts from the video rather than an image. Nevertheless, in recognizing the number plate, image from the video, unwanted frames bring about confusion for recognizing the number plate. The top five frames are selected using super resolution process. To address this problem, a memorability based frame selection algorithm that enables automatic selection of memorable frames for number plate feature extraction is introduced. In a video, multiple images of a number plate are available with different illumination, position, viewpoint, and camera distance.