| In the field of aerospace optical remote sensing,improving the dynamic range and the sensitivity of imaging,especially for the targets under complex or weak illumination,is always the goal pursued by technicians.In view of the digital TDI CMOS imaging technology,we explore the breakthrough in the imaging link which can effectively improve the imaging performance,and achieve higher dynamic range and higher sensitivity.This has important theoretical and practical significance.Image sensor is the central link of photoelectric conversion of remote sensing camera.With the development of CMOS image sensor,its performance in sensitivity and noise suppression has been greatly developed.The gap between CMOS and CCD is getting smaller and smaller.In some indexes,CMOS even exceeds CCD.However,in the digital TDI application,CMOS has an obvious weakness that the multi-stage images accumulation also causes the readout and quantization noise superimposed.Compared to the original image,the digital TDI can improve the gray level and the signal-to-noise ratio to a certain extent,but the dynamic range of the imaging and dark target detection ability is still limited.Considering the definition,to improve the dynamic range of camera,direct measures include increasing the number of full well charge or reducing the noise of the pixel.While to improve the sensitivity of the camera,direct measures include increasing the quantum efficiency or the photosensitive area of the pixel.However,limited by the current CMOS manufacturing process and application conditions,the level of these indicators can hardly meet the higher requirements.Therefore,it is necessary to find other measures beyond the sensor.The accumulation process of digital TDI can get many different levels of images,which are similar to the multi exposure images of the same scene.This provides valuable materials for the research and application of high performance imaging algorithm based on image processing.Focusing on the high dynamic and high-sensitivity imaging technology of the digital TDI CMOS,this article mainly carried out the following work.(1)Research on evaluation and enhancement of multi-stage images.Aiming at the best detail and texture feature of images,we take the entropy and the average gradient as the evaluation index of the multi-stage image selection.The Entropy Maximized Adaptive Histogram Equalization(EMAHE)algorithm is proposed to solve the problem of unclear enhance effect.According to the specific characteristics of the image,it can select optimal parameters automatically,and has good stability and wide applicability.(2)Research on selection of fusion algorithm and weighting factor for enhanced images.In order to further transmit and integrate the detail and texture features obtained from the multi-stage images enhancement,image fusion algorithm and weighting factor selection method are studied.We propose the image fusion method based on gradient pyramid decomposition,and take the entropy of the enhanced image of each decomposition layer in each direction as the fusion weighting factor.So that we can get the fused image with higher entropy and average gradient,and give it more information and clearer texture features.(3)Research on high-sensitivity imaging method based on shifted Binning.The method based on shifted Binning to improve the imaging sensitivity of digital TDI is proposed,which take the real-time image processing as a means.This method can greatly improve the imaging sensitivity and signal-to-noise ratio,and maintain the imaging resolution at a certain level.This method plays a significant role in improving the target detection ability of camera in weak light condition.Experiments show that the proposed high-dynamic-range and high-sensitivity imaging theory and method,can effectively improve the target detection ability of digital TDI CMOS in high dynamic scenes and weak light conditions.These also provide a good reference for other related imaging technology research. |
PDF Full Text Request