earticle

영어

This paper proposes a method of automatic exposure that applies image histogram feature (HF) function, which is used under the condition of fast and extensively changing background lighting so as to take automatic exposure control of the high-speed camera. First, multi-spot metering is employed to fetch region of interest from the acquired images and to reduce the calculation amount of systematic metering. Then, by calculating HF function in the region of interest, large step length is selected and exposure time is roughly adjusted. Finally, fuzzy logic is used to calculate the accurately adjusted step length, and threshold limit is applied to search for the optimal exposure time with variable step length, in order to improve the accuracy and the stability of automatic exposure of high-speed cameras. According to the results, within 2ms, the brightness of a one-frame image is measured and exposure time is adjusted. Compared to the conventional automatic exposure methods that are based on the average brightness value, within 0 to 110ms, when light intensity of the light source repeatedly changes between 760lux and 23100lux, the image information entropy obtained by this method is increased by 48.38%, and the variance is reduced by 62.13%. As verified by dynamic experiments, on the premise of ensuring dynamic resolution of the camera, the method in this paper is able to rapidly acquire the optimal exposure time, to provide more favorable and more stable image details, and to offer reference for the subsequent automatic focusing, image recognition and target tracking.