The Refocusing Research Of Images

Traditional ordinary cameras can only peceeive information of single plane. We must select the parameters of the traditional ordinary cameras carefully before we use them, and we must confirm that the focusing place within the scope of our vision should be a plane. The calculation cameras which can refocus after shoot and realize the depth of field has a bright application prospect in recent years. Depending on its smaller bulk and the well portablity, micro arrays camera can be applied to the handheld and tablet devices. Many experts speak highly of it and consider it to be the sign of the second revolution of the camera digital technology. So the refocusing research of micro arrays camera will be a important subjuct in the computer vision areas. But there are few people are dedicated this study which have not been widely aplied. This paper study amounts of theorys and algorithms in this atomosphere. We finished following work in this paper:1) we reviewe the theories of the camera imaging model and camera calibration, especially the knowledge of Zhang’s calibration. We calibrate 9 micro cameras of the 3×3 arrays camera.2) We introduce the classification of the stereo matching algorithms, and analyse the theory and steps of the graph cut. We build an energy function according to the the compatibility constraints in grayscale and constraints in depth smoothness of adjacent pixels of the 9 images.3) We solve the energy function we builded by using the algorithm of graph cut, and get the depth information of each pixelof the image. We refocus the image finally.The research of refocusing based on arrays camera is still relatively new field, and it is in exploration constantly. As a result of the limitation of hardware technology, the refocusing algorithm has certain deviation because of the limit of the hardware. In order to get better experimental result, we will verify the algorithm on the arrays camera which equipped different focal cameras.