Study And Application Of Automatic Image Processing Technology For The Second ID Card System

For the Second ID Card, we always address the technology research of automatic processing for digital portrait images. It is a key for the Second ID card to how to better the effect of images automatically and let them conform to the technical standard of the ID card that the police department publicized. So the thesis works on adjustments of rotation, color, zoom, shear and background mostly, which based on the pose, color and location of the portrait in the image. And the researches of the article have been done as follows:Firstly, about the adjustment of rotation, a method, based on SUSAN corner detection algorithm, is presented in the thesis. And the rotated image is rectified exactly by the symmetry of the portrait image and the invariant to rotation of shoulder corner. Then the detail of images is preserved by the approximately geometrical processing based on the bicubic interpolation algorithm.Secondly, on the adjustment of color, the processing from RGB color space to HSI color space can strengthen the information structure, and reduce the loss of color information. Then, in HSI space, the method based on adaptive histogram equalization (AHE) is introduced to adjust the saturation and intensity. And then, the contrast is regulated further.Thirdly, concerning the rotation and the shear of portrait images, the face area can be detected accurately and rapidly by an effective algorithm of the cascade classifier based on Adaboost learning method. So the adjustment of zoom is realized. And then,the normal shear is achieved ground on the projection method. Afterwards, according to SUSAN edge detection, a new edge detection algorithm of portrait is discussed to process the background rapidly.Finally, the automatic image processing proposed in this thesis is formed to an integrate system. And with example of a number of digital photos collected, the results are given, analyzed and evaluated.