Application Research On The Image Processing In The Surveillance Of The Customs

The background of the application of the image processing in the surveillance of the Customs is presented, and then the technologies and problems in the area are discussed and analyzed. At last, our scheme for those problems is brought forward. We also discuss key tasks of the present surveillance of the Customs—the loading and unloading of the ware, in which the main problems present are the recognition of the number of the container and the vehicle license plate. The two tasks have high similarity, so we discuss our algorithm for the recognition of the container number.Our scheme comes forth as following: Firstly, the image captured in the real world is gray-scaled to accelerate our algorithm, secondly the median filter is used to decrease the noise of the noise which comes from the system of imaging, and thirdly the max-variance algorithm is used to binary our image. Based on the binary image, top-hat operation is used to detect the structure of the character, and some mathematical morphology is used to remove the disturbance, and then projection is used to locate the region of the interest area. After the location, projection is also used to segment the character, which is sent to the recognition system after size normalization.For the recognition of the number, we discussed two schemes: one is for the Chinese, and the other is for the letters and the numeric. For the Chinese, the principal component analysis and radial basis function are used to get over its complex pattern. At the same time, particle swarm optimization is used to accelerate the train speed of the neural network. And for the letters and the numeric, four different classifiers are designed to increase the precision of the recognition. And Bayes algorithm is used to integrate the results of the four classifiers. The experimental results show that good results can be obtained by the scheme we put forward.At the end of the paper we summarize and expect our work.