The hyperspectral image destructor (HSI) is an essential preprocessing step for improving the performance of subsequent applications. For HSI, there is much global and local redundancy and correlation (RAC) in spatial / spectral dimensions. In addition, the reduction performance can be greatly improved if the RAC is used efficiently in the process of eliminating recess. In this paper, we propose a HSI HSI elimination method using the global and local RAC in spatial / spectral domains. Firstly, sparse coding is used to model the global RAC in the spatial domain and the local RAC in the spectral domain. Noise can be eliminated by scarce approximate data with learned dictionary. At this stage, only local RAC in the spectral domain is employed. It will cause spectral distortion. To compensate for local spectral RAC deficiency, low-rank restriction is used to deal with the global RAC in the spectral domain. Different sets of hyperspectral data are used to test the performance of the proposed method. The results of elimination of the degradation by the proposed method are superior to the results obtained by other methods of hyperspectral deenergization of the state of the art.
The recent breakthrough in sensor technology is a boon to hyperspectral remote sensing. Although hyperspectral images (HSI) are captured using these advanced sensors, they are highly prone to problems such as noise, high dimensionality of data and spectral mixing. Among them, noise is the biggest challenge that affects the quality of the captured image. In order to overcome this problem, the hyperspectral images are subjected to spatial preprocessing (denoising) before image analysis (Classification). In this article, the authors discuss a strategy based on shortage based that uses low bandpass filter matrices (AB filter) to effectively eliminate each band of HSI. Both subjective and objective evaluations are performed to test the efficiency of the proposed method. Subjective evaluations involve visual interpretation, whereas objective evaluations are concerned with the calculation of quality matrices such as the Signal to Noise Index (PSNR) and the Structural Similarity Index (SSIM) to different noise variances. In addition to these, the recessed image is followed by a shortage based classification using the Orthogonal Combination Search (OMP) to evaluate the effect of various waste disposal techniques on sorting. The classification indices obtained without and with the preprocessing are compared to highlight the potential of the proposed method. The experiment is performed on the Indian Pines standard data set. Using 10% of the training set, a significant improvement in overall accuracy (84.21%) was obtained with the proposed method, compared to the other existing techniques.
Denoising is a key task in hyperspectral image processing (HSI) that can improve the performance of classification, unmixing, and other subsequent applications. In an HSI, there is a large amount of local and global redundancy in its spatial domain that can be used to preserve detail and texture. In addition, the spectral domain correlation is another valuable property that can be used to obtain good results. Therefore, in this paper, we have proposed a new HSI Denoising scheme that exploits composite spatial-spectral information using a non-local technique (NLT). First, a specific way of extracting patches is used to effectively exploit spatial-spectral knowledge. We then use a framework with compounding regularization models to implement the destructor. A number of HSI datasets are used in our evaluation experiments and the results demonstrate that the proposed algorithm outperforms other later generation HSI Denoising methods.