Technique for Anomaly Information Reconstruction in Airborne Radiation Data Based on Dictionary Learning Algorithm

This paper introduced the K-SVD algorithm of machine learning to Decompose and Reconstruct the airborne radiation data. According to the sparse representation method, the inaccuracy of gridding interpolation due to the mismatch of data dimension was reduced and then a self-adaptive dictionary was constructed via training the original data. After reconstruction by dictionary learning, the increase in terms of two indicators peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM) as well as the noise signal extracted which obeyed a stripe-like distribution along the sideline testified that the KSVD encoding learning processing technique had effective results in reconstructing and de-noising airborne radiation data.