Study on the key technology of spectral reflectance reconstruction based on a single pixel detector

In order to realize an accurate reconstruction of images, this article explores image reconstruction from a spectrum point of view. Combined with a principal component analysis method and compressed sensing algorithm, this article puts forward a spectral reflectance reconstruction algorithm based on a single pixel detector, and compares the different modulation matrixes for DMD in the experiment. From subjective evaluation and three objective evaluation methods: color difference, peak signal-to-noise ratio and structural similarity, we compare and analyze the reconstruction effects of two modulation methods: random modulation and Hadamard modulation. The experimental results show that the color difference of the spectral reflectance reconstruction method of a single pixel detector based on random modulation is smaller, the similarity of the structure is higher, and the peak signal-to-noise ratio is greater than Hadamard modulation. To sum up, the spectral reflectance reconstruction algorithm of a single pixel detector based on random modulation is better than Hadamard modulation, which is more conducive to an accurate reconstruction of spectral reflectance.