Extraction of vehicle license plate number using license plate calibration

摘要

License plate number extracting from a vehicle license plate (VLP) image is considered to be the key stage for vehicle license plate recognition in automated transport system. It significantly controls the performance of an automated transport system. This paper presents an effective scheme to extract the VLP number from a VLP image. The presented algorithm first extracts out the edges of the VLP image using image enhancement and Sobel operator, using Hough transform to calibrate distorted VLP images, then using horizontal boundaries detection and vertical boundaries detection to shrink the VLP region, and finally extracts the plate number out from the original VLP image. The experiment results show that the presented scheme can extract the VLP number effectively and efficiently even in distorted VLP images.