针对高光谱分类中对光谱信息和空间信息利用不足的问题,提出了一种基于稀疏表示模型和自回归模型相结合的分类算法。该算法利用稀疏表示模型和自回归模型,设计联合字典:在光谱维上,利用稀疏表示模型将高光谱的每个光谱向量表示为字典中训练样本的稀疏线性组合;在空间维上,利用自回归模型对每个光谱向量的8邻域进行约束。针对不同样本分别构造一个字典,在减少计算量的同时减小重构误差,最后在最小重构误差和邻域相关性的约束下求解稀疏表示问题,以最小重构误差为准则实现高光谱数据的分类。仿真结果表明,该方法能够有效地提高高光谱数据的分类精度。 Aiming at the problem of insufficient use of spectral information and spatial information in hyperspectral classification, a classification algorithm based on sparse representation model and autoregressive model is proposed. The algorithm uses a sparse representation model and an autoregressive model to design a federated dictionary. In the spectral dimension, each spectral vector of the hyperspectral spectrum is expressed as a sparse linear combination of the training samples in the dictionary using a sparse representation model. In the spatial dimension, The regression model constrains 8 neighborhoods of each spectral vector. For each sample, a dictionary is constructed to reduce the reconstruction error while reducing the computational complexity. Finally, the sparse representation problem is solved under the constraint of the minimum reconstruction error and the neighborhood correlation, and the hyperspectral data is achieved by using the minimum reconstruction error as the criterion Classification. Simulation results show that this method can effectively improve the classification accuracy of hyperspectral data.