Research On Color Night Vision Imaging Methods

Traditional low-light night vision and infrared thermal imaging cameras have successfully expanded the human eye's spectral response range,and can distinguish targets in a dark environment,but the resulting images are mostly monochrome images,due to the color level that human eyes can distinguish It is several hundred times the gray level.For a long time,scientific researchers have devoted to the colorization of black and white night vision images in order to use the visual characteristics of human eyes to more effectively obtain the target scene information.Therefore,the research of color night vision technology has important theoretical and practical significance,and it is the current research focus in the field of night vision at home and abroad.In order to achieve true-color imaging under low-light conditions,this paper develops a true-color night vision imaging system based on a CMOS sensor,and proposes an effective improvement scheme on this basis.First,design the overall scheme of the color night vision system,determine the sensor signals,select the image processing chip and the coding chip,design the circuit schematic diagrams for these three main chips,make the hardware circuit board,develop the sensor module,FPGA module and interface module;second,Write code to drive the sensor,and perform targeted data recovery and decoding based on the output data format and image frame composition;after that,transplant related image processing algorithms to FPGA,mainly demosaicing algorithms and images based on Bayer interpolation Enhance the denoising algorithm,the white balance algorithm for color reproduction,and the gamma correction algorithm to increase the contrast of the image,and optimize the traditional dehazing algorithm to remove the fogging of the image and increase the color saturation of the image.The camera module was jointly tuned to develop a true-color night vision imaging system,and it was demonstrated for imaging and performance testing.The results show that the selfdeveloped color night vision imaging system output image is consistent with the color of the real scene,and can be imaged in real time under the illumination of 10-2lx,which meets the requirements of low-light night vision imaging.At the same time,the subjective and objective methods are used to evaluate the output image quality of the system.Three different target images are collected under different illumination and gain.As the illumination decreases,the peak signal-to-noise ratio and structural similarity gradually decrease,and the root mean square The error,chromatic aberration,and chromaticity aberration gradually increase,adjusting the gain can effectively improve the image quality,and the peak signal-to-noise ratio is significantly improved.Combining image enhancement and denoising algorithms can significantly improve the brightness and smoothness of the image.Finally,this paper explores three improved solutions of black and white + color dual sensors,color filter film based on Bayer sparse array,and near-infrared transmissive film to further improve the sensitivity and imaging quality of self-developed cameras.