The Study Of Auto-Focus Algorithm Based On Digital Image Processing

Auto-focus is one of the key problems in computer vision, imaging systems and various precision instruments. In the last ten years, research on auto-focusing techniques has been focused on methods related to digital image processing on account of the maturing of digital imaging devices, and a lot of achievements have been accomplished. The auto-focus means based on image technology is different from those traditional auto-focus means entirely. The traditional ones must depend on some special assistant equipment. They use these equipments to measure distance or find focus and realize auto-focus. However, the auto-focus based on image technology is completed by image quality analysis. It can judge the imaging state from the current image quality, and then the system can adjust focus distance correctly supervised by imaging state. The image quality analysis is the key technology.First, the dissertation briefs the principles and research status of auto-focus in digital imaging systems. Then, the dissertation discussed the past and present state of image quality evaluation. Following that, two image quality analysis methods are presented in detail in the thesis:(1) One is the assessment of image quality through gray scale area. It describes a new method which evaluates quality of image through searching and contrasting gray scale area. In this method, image was treated as being composed of many different gray scale areas. Distortion of each gray scale area was computed. Then the image quality was assessed and the simulation results are presented. And it indicates that this algorithm works effectively.(2) The other one is the wavelet based image quality evaluation. It analyzes image from different frequency resolutions through wavelet theory, and analyzes the change of Wavelet Coefficients through statistics analysis, and then assesses the image quality.The realization of auto-focus, for the DFF (depth from focus), is similar with the 1-demension search, Fibonacci search, function approaching or MCS (mountain climbing servo), calculates the auto-focus function value, compares it with the previous one and decides the next position; for the DFD (depth from defocus), moves a special distance relative to the image quality level and system parameters directly. Depending on above auto-focus theories, the thesis puts forward an auto-focus system architecture and analyzes its performance, which is constructed by high speed DSP (Digital Signal Processor) and FPGA (Field Programmable Gate Array).