Research On The Stripe Noise Removal Method For Uncooled Infrared Images

With the development and wide application of uncooled infrared detectors,infrared imaging technology has been further developed.Due to the limitations of fabrication process and materials,a special kind of non-uniformity noise--stripe noise,can appear in the infrared imaging system,which greatly reduces the imaging quality of the infrared system and has some impact on the subsequent applications.To address the above problems,this topic studies the stripe noise removal methods for uncooled infrared images,and the research contents are as follows:(1)For the traditional algorithms,the approximate component after wavelet decomposition also contains stripe noise.This thesis proposes a stripe noise removal method for infrared images based on wavelet analysis and parameter estimation.First,the wavelet transform is used to decompose the image into components of different directions.Then the approximate and vertical components containing stripe noise are denoised separately,the approximate component is denoised by parameter estimation and the vertical component is denoised by a one-dimensional guide filtering.Finally,the wavelet reconstructs the image to obtain the denoised image.This method precisely removes stripe noise from the image and preserves as much detail as possible in the image.(2)To address the phenomenon that the vertical edge structure of the image is often lost in the guide filtering algorithm,an infrared image stripe noise removal method based on Huber penalty function and guide filtering was proposed.First,a smoothing structure based on Huber penalty function is used to smooth the image to obtain the smoothed part and the high-frequency part,where the stripe noise is mainly concentrated on the highfrequency part.Then,the high-frequency part is denoised using a one-dimensional guide filter.The method can effectively distinguish the vertical edge structure and stripe noise,and avoid the vertical edge structure from being erroneously removed.In order to better verify the advantages of the algorithm in this thesis,the typical stripe noise removal algorithms in recent years are compared with the two algorithms proposed in this thesis for experimental analysis.The experimental results show that the algorithm proposed in this thesis can better remove the stripe noise and retain the image details,which verifies the effectiveness of the algorithm proposed in this thesis.