Infrared stripe noise correction algorithm based on multi-scale analysis and one-dimensional vector convolution

Aiming at the problem of stripe noise in uncooled infrared imaging system, a multi-scale analysis and one-dimensional vector horizontal convolution stripe noise non-uniformity correction algorithm is proposed. The algorithm uses the wavelet transform to extract the vertical component of the original infrared image, and uses the Gaussian kernel function to generate the one-dimensional vector to smooth the vertical component, and the wavelet reconstruction outputs the corrected image. The algorithm can accurately remove infrared noise without causing more troublesome "ghosting" problems. Multiple sets of different infrared image data were selected for simulation experiments, and compared with the current advanced infrared stripe non-uniform correction algorithm and qualitative and quantitative analysis, which proves that the algorithm proposed in this paper has a good visual effect, and the image quality evaluation parameters can also get good results