Research On Image Denoising And Enhancement Technology Based On Inter-frame Informatio

Time domain denoising algorithm is widely used in image processing system is used to improve imaging quality,under the static scene time domain filtering algorithm can effectively improve the image signal-to-noise ratio,but the existing time-domain filtering algorithm is difficult to guarantee the movement state of image denoising ability,in removing noise and easy to produce the tail phenomenon,and use of motion compensation temporal filtering algorithm and tend to noise In view of the above problems,this paper carries out research on the algorithm of video image denoising and enhancement.The main contents of this paper are as follows:Firstly,the noise characteristics of the image are analyzed,the imaging model of the target in the image is derived,the method and technology of image feature extraction are studied,and the feature extraction method based on image scale invariance is analyzed.Then,the principle and limitation of time domain filtering algorithm are studied,the theory of feature sparsity is proposed,and the feasibility of feature extraction algorithm for video motion region is demonstrated.According to the difference of different frequency band Gaussian difference images,a high frequency noise suppression strategy based on spatial filtering is proposed.Combining time-domain filtering with motion compensation and high-frequency noise suppression,a time-space domain video denoising algorithm with motion compensation is proposed.Based on the high frequency characteristics of DOG image studied in the time domain denoising algorithm,the high frequency features of the image were studied.The study found that the high frequency information could significantly enhance the details of the image after the signal-to-noise ratio(SNR)was improved,and an image accumulative enhancement algorithm based on inter-frame high frequency information was proposed for this feature.Experimental results show that compared with traditional image enhancement algorithms,the proposed algorithm can enhance image details more effectively.High frequency information can also come from different sensors,analyze the imaging characteristics of infrared images,put forward the principle of infrared region of interest extraction based on the image high frequency characteristics,so as to fuse infrared and visible images,to solve the problem that the spectrum detection range of a single visible sensor is limited.And an improved image fusion algorithm based on color space transformation is proposed to further enhance the image information.Finally,the proposed image denoising algorithm based on inter-frame information is transplanted into ZYNQ hardware platform,including hardware transplantation optimization of algorithm,data flow processing,module parameterization design,etc.,and the realization of the hardware system is tested to verify the effectiveness of system functions.