Study on the key technology of reconstruction spectral reflectance based on the algorithm of compressive sensing

According to the low reconstruction efficiency and precision, a kind of spectral reflectance reconstruction method based on the algorithm of compressed sensing is provided. It can make full use of the sparse characteristics of spectral reflectance to improve the reconstruction precision and efficiency. In this paper, the first three principal components of the spectral reflectance with high contribution is obtained by the method of principal components analysis based on the analysis of the algorithm of compressive sensing and the least squares method. The dimension of the high dimensional spectral reflectance is reduced. The high dimensional spectral reflectance image is reconstructed by the iterative threshold method of the algorithm of compressive sensing. The simulation of the spectral reflectance reconstruction is simulated by the method of principal components analysis and algorithm of compressive sensing through the MATLAB software simulation platform. From the simulation, we can conclude that the reconstruction accuracy and efficiency by the algorithm of compressive sensing are better than the one by the method of pseudo inverse. The reconstruction accuracy is affected by the selection of training sample set, the sampling interval, and the iteration number. The reconstruction accuracy decreases as the increase of the sampling interval and the decrease of the iteration number and the contribution of the first three principal components. The reconstruction accuracy increases as the increase of the iteration number. The higher similarity between the selected training samples set and testing samples, the better representative of the training samples and reconstruction accuracy. The spectral reflectance reconstruction method based on the algorithm of compressed sensing can make full use of the sparse characteristics of spectral reflectance to improve the reconstruction accuracy and efficiency.