Study On Auto-focusing Technology For Photoelectric Theodolite

Variation in image content can lead to a decrease in the auto-focusing ability of control system because of the changes of sharpness evaluation value, which is very common in real-time image measurement equipment, such as photoelectric measuring equipment, radio observations, etc. During the process that the photoelectric theodolite tracking moving target, the distance between the camera lens and the object is changing all the time, so are the size and shape of the target. Therefore, the real-time auto-focusing is necessary to satisfy the image sharpness quality. At present, manual focusing is mostly used in the photoelectric theodolite system, which evaluating image sharpness /blurriness depending on human and the system has low degree of automation and intelligent. In conclusion, the study on real-time and reliable auto-focusing has important significance and practical value.This paper researches into auto-focusing based on digital image processing technology, including the sharpness evaluation algorithm and the auto-focusing search strategy. The sharpness evaluation algorithm is used to measure the sharpness/blurriness with a criterion function. And the auto-focusing search strategy is used to determine the focus motor position providing the sharpest focus.Three main parts of this paper are as follows:Firstly, the traditional image sharpness evaluation algorithms are summarized.And the performance of the traditional sharpness evaluation functions based on spatial domain is tested and analized.Secondly, a simple but effective real-time sharpness evaluation algorithm is proposed which is independent of image content. Combined with the distribution of natural image gradients, the algorithm makes the most of the diversity in edge width and edge intensity distribution among images of different sharpness. It focuses on the overall property that reflects the sharpness of images and is robust to various image content.In the end, an improved hill-climbing search strategy is presented combined with the sharpness evaluation algorithm independent of image content. By making full use of the relations between the sharpness evaluation values and the image quality, a self-adaptive step magnitude is applied to reach the place near the focus position roughly. Moreover, the algorithm can judge reliably whether the image is sharpnessenough to meet the quality demand.Experimental results of natural scenes with different contents demonstrate the accuracy, real-time performance and applicability of the algorithm. The sharpness evaluation value of any natural image is always in a range of 0 and 1. Two natural images with different contents but equivalent sharpness values get the same sharpness degree of human vision. Conclusion: The sharpness evaluation value corresponds well with human visual assessment and can reflect the degree of image sharpness or blurriness. The proposed improved hill-climbing search strategy can locate the focus position rapidly and accurately.