Research On Noise Detection And Correction Method Of SCMOS Camera

The sCMOS camera is a scientific camera that combines the advantages of low output noise,high sensitivity,high resolution and high frame rate.It plays an important role in medical diagnosis based on image,microscopic imaging,astronomical observation and high-definition intelligent monitoring.Obviously,the high performance of sCMOS cameras not only depends on the technical advantages of their hardware design,but also depends on the algorithm optimization in software control.Since sCMOS cameras have high requirements in image quality,they are also extremely sensitive to noise compared to ordinary cameras.As the primary factor affecting image quality,image noise has been receiving much attention in the scientific community.Although some filtering methods can suppress the noise interference to a certain extent,these methods often destroy the details of the image,and it is difficult to meet the image quality requirements of the sCMOS camera.Therefore,in order to make sCMOS cameras play their role better in scientific research,it is of great academic significance and application value to develop a noise detection and correction method with high accuracy,low image destruction and strong adaptive ability.The main research work and results of this thesis are as follows:1.The fixed pattern noise detection and correction method for sCMOS camera is studied.According to the cause and characteristics of fixed pattern noise,a multi-image cross-reference detection idea is proposed.The fixed pattern noise adaptive detection method in dark scene and illuminated scene is designed.Based on the relative stability of the fixed pattern noise,the detection and correction processes are separated,and the noise coordinates are stored after the detection is completed,and only the pixels at the coordinate points are corrected later.The experimental results show that the detection accuracy of the fixed pattern noise of this method is as high as 99.13%.Compared with the existing methods,the corrected image peak signal to noise ratio is improved by 7%,and the correction time is shortened by 50%to 0.5 second.In the following,the image quality and the correction speed are effectively improved.2.The random noise detection and correction method of sCMOS camera in static scene is studied.According to the characteristics of random noise and the relative stability of continuous shooting images of high frame rate cameras in static scenes,a random noise detection method based on cross-referencing multiple images is proposed.For each noise of the target image,a statistical random noise correction algorithm is designed.By statistically analyzing the values of the corresponding coordinate points in multiple continuous shooting images,the accurate value is determined and used to replace the noise of the target image.Compared with the currently used filtering algorithms,this method achieves the optimal original image restoration effect,it has outstanding performance in the correction accuracy of random noise,and it is more suitable for image denoising of sCMOS cameras in static scenes.3.The secondary development algorithm of HK_A5100_MINI software was designed.The software is compatible with sCMOS camera devices and has data import capabilities for image information,allowing two optional noise correction modes.The software can greatly facilitate the user to obtain and analyze the image information.The process time of the software can be as low as 0.4 second,which is an important supplement and extension to the function of the sCMOS camera.