A New Method of License Plate Location Based on Multi-threshold Quantization Segmentation of Color Image

Abstract

In the complex reality environment, License plate image may contain multiple plates. At this point, simply using grayscale or color image processing only is very difficult to achieve fast and accurate positioning. By full use of the multiple conditions such as color, edge features and veins, a new method of multi-objective, integrated color of license plate location has been put forward. Firstly, make the multi-color threshold quantization segmentation of color image, edge detection and morphology of its deal to get a candidate region, integration of the color code matrix diagram and quantify the length and width of the plate area fixed ratio to get the second candidate region, then in the second candidate region, judging the edge points on both sides of the diagonal with the color quantization value, verify the authenticity of the candidate regions. Through progressively filter the candidate region, and ultimately locate each license plate. The algorithm is full integration the advantages of the method based on color image processing, simple and effective, and have good positioning accuracy and engineering practicality.