Hyperspectral Imagery Classification And Anomaly Dection Based On Sparse Representation

Hyperspectral image classification and anomaly detection is a technology that capture spectral characteristics of ground objects for pattern classification and anomaly detection by aerial remote sensors. Compared with the traditional multi-spectral remote sensing, hyperspectral remote sensing can be a combination of spatial dimension and spectral reflection characteristic, creating favorable conditions for hyperspectral image processing technology. Hyperspectral image classification and anomaly detection based on sparse representation is the main research directions of this article, the main work and research results of this paper the following aspects:1. It studys the research status of hyperspectral image classification and anomaly detection, the characteristics of hyperspectral images is also analyzed.2. The related techniques of hyperspectral image data processing are introduced. Firstly, the basic principle of dimension reduction technique is described, and then some typical algorithms and their characteristics are introduced in this paper. The classification and anomaly detection of hyperspectral images are studied, and the supervised classification and unsupervised clustering algorithm based on machine learning are introduced in detail.3. The feature extraction and classification of hyperspectral image based on sparse representation is introduced in this paper. A sparsity and low-rank graph-based discriminant analysis (SLGDA) is proposed for the dimensionality reduction in hyperspectral imagery. Experimental results on several hyperspectral images show that the proposed framework has better performances than the traditional alternatives.4. In this paper, a sparse-based anomaly detector is proposed to remove outliers in the local background region before employing the RX algorithm. Experimental results demonstrate that the proposed detector improves the original local RX algorithm.