Auto-focus And Auto-exposure Algorithm Based On Image Processing

In Digital Still Camera, focus and exposure are two important parameters that will determine the quality of the final picture. Auto-focus and auto-exposure are two key techniques.In traditional auto-focus algorithms, the circuit and move institutions are complex. Besides, the focusing is not intelligent. Therefore, in this dissertation after analyzing the image formation theory, we investigate a new algorithm based on the image processing technique, namely IFDA algorithm. By processing variable number of input blurred images, we can recover the sharp image. In IFDA algorithm, we think of the evaluate function as the criterion which judge whether the image is sharp or not. In this dissertation, we also compare the effect of three different optimization techniques that optimize the recovery image, and confirm the best one.Applying the BP neural network into the image processing technology, Auto Exposure algorithm is investigated. First we segment the image into five parts, and then the histogram of each part will give the brightness. Let the brightness information of each part be the input nerve cell of the BP neural network, we can get the suitable exposure value of the image. According to the exposure value, we can control the exposure of the Digital Still Camera availability.Simulink the Auto-focus and the Auto-exposure algorithm, the results show that the two algorithms are satisfactory.The main feature of this dissertation is that the realization of the two algorithms needs no additional device but image process technology.