License Plate Location Method Based on Integrated Features and Marker-Controlled Watershed Algorithm

摘要

The License plate location is an important stage in Intelligent Transportation System (ITS). It is considered to be the most crucial step of an Automatic License Plate Recognition (ALPR) system. A new image segmentation method based on integrated features and Marker-Controlled Watershed algorithm was proposed to locate and split the plate region of vehicle images from complicated environment. Firstly, the color license plate image was converted to grayscale image, and Top-hat transform was applied to restrain background and eliminate variant illumination to enhance the contrast of the image, and then Marker-Controlled Watershed algorithm was used to segment the preprocessed images to get candidate regions, lastly, a new distinguishing method based on integrated features of the plate was proposed to locate the accurate plate region. Experimental results show that this new method is simple, effective and robust to interference characters and complicated background.