The Research Of The Low-light Noise Reduction And Demoasic Algorithm Based On Adaptive Feedback

In low-light condition, how to acquire large-area high-resolution digital images isa key problem for civil and military applications. Low-light noise is an importantproblem in digital camera imaging. Most business camera has poor low-light imagingperformance, leading to obvious artificial noise in the image. So how to suppresslow-light noise of color image has become the focus in ascending the imagingperformance in low-light condition. In this paper, we analysis the basic principle ofadaptive control and the basic principle of system identification. In addition, wesummary the advantages and disadvantages of several classical image denoisingalgorithms and demosaicing algorithms. By using the ideas of adaptive control andsystem identification, this paper builds up the non-local low-light denoising anddemosaicing algorithm modal. The experimental results prove the effectiveness of theproposed modal.The study of the paper primarily contains four aspects:Firstly, this paper summarizes the development and research status of low-lightcolor camera, analyzes the characteristics of low-light image noise, and then expoundsthe imaging principle of camera and the Mosaic phenomenon. This paper alsosummarizes the advantages and disadvantages of several traditional denoisingalgorithms and traditional demosaicing algorithms.Secondly, this paper summarizes the basic principles, the basic models and themain research problems of the adaptive control theory, and then summarizes the basicprinciples, the research problems and the basic methods of system identification theory.Thirdly, this paper analyzes the traditional non-local mean image denoisingalgorithm, and proposes two kinds of improved algorithm modal. The one sets up theAdaptive Feedback Nonlocal Means Image Denoising Algorithm Based onMathematical Morphology modal, by using the idea of adaptive feedback forreference. The other sets up the A Doubly-Scale Adaptive Non-local Means ImageDenoising Algorithm modal, by cleverly using two dimension transformations.Experimental results demonstrate the superiority of the two proposed algorithm modalsto some other improved NLM algorithm modals.Finally, in this paper, through analyzing the characteristics of the poisson noise under low-light condition, for solving the problem of how to determine the statisticalmodel of the noise, we build up the Adaptive Feedback Non-Local Mean Low-LightImage Denoising Algorithm via Noise Classification, by using the idea of the adaptivecontrol and the residual error analysis approach for reference. And then, according tothe characteristics of the color filter array, this paper sets up A Color Image NoiseReduction Method Based on Non-local Mean Denoising Algorithm and RegionalAdaptive Gradient Interpolation Algorithm. At last, we get the finally low-lightdenoised color image, and reac h the goal of this paper.In this paper, through a large number of simulation experiments, we demonstratethe effectiveness of the several algorithm models mentioned above. Finally, the fulltext is summarized, and in the end the next step is prospected.