Research On Evaluation Function And Window Construction Algorithm In Auto-focus System

With the rapid development of digital imaging technology,optical imaging devices have been widely used in industrial imaging,medical image analysis,remote sensing instruments and military photography.For optical applications,the core of imaging system is auto-focus.The task of auto-focusing is to automatically obtain the best image,which plays a crucial role in improving the image clarity.Auto-focus is the key link to affect the image clarity of the imaging system.The basic idea is to replace the tedious process of manual focusing with the automatic operation of the lens and ultimately find the correct focus position of the imaging system.The auto-focus system uses a setup program to run a micromotor that calculates the sharpness of the image to find the correct focus position and projects the sharpest image of the subject onto the image sensor.Auto-focus technology saves system time and can be found on most imaging devices today.In most cases,it helps to improve the quality of the images taken.Auto-focus systems generally fall into two categories:active and passive.The research of this paper focuses on passive auto-focus systems.Passive focus adjusts the lens to the proper position by analyzing the image itself without the need for additional equipment.The principle of passive focusing technology is to run the motor to obtain the image of the object at different focal positions,then calculate the sharpness value of the image,and finally push the lens to the corresponding focus position when the sharpness value is the maximum.Thanks to the advantages of fast response,convenient operation,and system intelligence,passive autofocus based on image processing has become one of the widely used methods in this field.In general,the auto-focus algorithm based on image processing is usually composed of three parts:sharpness evaluation,focus window construction and extremum search.The research work of this paper is also focused on these three aspects.The innovations of this paper mainly include:(1)The gray level co-occurrence matrix can reflect the gray change between pixels in the neighborhood of the image,and its contrast can measure the quality of imaging.Based on the above theories,a sharpness evaluation function based on the gray level co-occurrence matrix is proposed.After analyzing the experimental results of different types of image sequences,it can be seen that the sharpness evaluation function in this paper satisfies the requirements of the evaluation function and has a better performance in terms of noise resistance.(2)On the basis of qualitative evaluation indicators,quantitative evaluation indicators are added,and the combination of the two can more accurately evaluate the performance of the focus function.(3)The focusing window evaluation factor is used as one of the evaluation indexes of the focusing window construction algorithm.At the same time,the corresponding evaluation function curve is obtained by combining the evaluation function proposed in this paper with the first-order windowing method of difference shadow and several common window construction methods.(4)For the determination of the focus window,a first-order moment window construction method based on image difference is proposed.Gaussian blur can smooth the strong edge of the image,using a Gaussian blur to obtain a smooth image,and a difference map obtained by difference between the original image and the smooth image.In addition to the background of the original image,the difference image map removes the influence of the background area and highlights the details of the target area of the original image.Using the first moment to locate the center of the focus window in the difference map,the focused window can be obtained according to the preset window size.