Research And Design Of The License Plate Location Technology

License plate location is a key step in the license plate automatic identification system, it also play a vital role in the license plate character recognition. Currently, many researchers use three sub-values of the HSI color model to process image in the license plate location, calculate three sub-values at the same time will affect the efficiency of the program,at the same time, a single color model is not able to meet all the environmental conditions, And some scholars simply use H of the HSI color model to process the image, due to the light factors, which could cause some error for extract the image interesting area. In response to these issues, this thesis mainly research and discussion the license plate image binary and edge detection, and made the corresponding theoretical analysis and related experimental. The main work is as follows:(1) License plate image binary. By comparing and analysis of relevant color model, this paper introduces the HSI color model and the YUV color model, according to the S value distribution of image pixel saturation, we option of using H of the HSI color model or Y of the YUV color model to binary the license plate image, through a combination of two color models, we overcome the error caused by a single color model.(2) License plate edge detection. In the edge detection, analysis and comparison of the common edge detection operators, and analysis some thesis, in this thesis, we improvement the Prewitt edge detection operator, change it from the square to the cube of the difference. In order to enhance the image edge, at the same time, overcome the disadvantage of the Roberts operator susceptible to the noise .(3) This thesis also made some corresponding theoretical analysis and made the relevant experiment for the other relevant steps of the license plate location.In mathematical morphology, this paper use the expansion operation and opening operation of the mathematical morphology in the image processing. In the license plate location, this thesis use gray level characteristics of jump up and down the license plate characters to determine the boundary, and then use the vertical projection, scan every column of the license plate image, and then determine the license plate area Left and right borders, ultimately locate the license plate region.This thesis implement the relevant algorithm in Visual C++ 6.0 programming environment. Through the test results indicate that the algorithm can accurately positioning plate.