Research on the Auto-Exposure Method of an Aerial TDI Camera Based on Scene Prediction

Abstract

Aerial TDI cameras can obtain images with a high sensitivity, high resolution, and wide dynamic range under low-illumination conditions. Under the condition of low illumination, exposure is an important parameter that affects the image quality of the camera. Auto-exposure aims to obtain the information of target scene in advance and uses it as the basis for determining exposure parameters, which can avoid the loss of image information caused by overexposure or underexposure. The auto-exposure of TDI CCD is usually difficult, as the shooting mode of TDI CCD is a push and sweep mode, which can only take one image of the target scene, and the output image is not repetitive, so it is difficult to obtain an image of the target scene in advance. At present, the common method is to add an additional sensor to collect the feature information of the target scene in advance; however, this increases the complexity of the system. Therefore, this paper proposes that the camera uses TDI CCD to collect the first N rows of data of the target scene in advance as the basis, and then adjusts and determines the exposure parameters with the median gray value as the target; thus, without adding additional sensors, the auto-exposure of TDI CCD can be realized. To evaluate the effect of auto-exposure, three methods of image power spectrum variance, image histogram, and image information entropy are used. The test results show that, after auto-exposure, the variance in the image power spectrum increased by 0.4362, the entropy of image information increased by 1.7064, and the distribution of the image histogram was more uniform than that before auto-exposure. This shows that the effect of auto-exposure is good, and it has better scene adaptation, allowing for it to meet the requirements of auto-exposure imaging under low-illumination conditions in aerial TDI cameras.