Abstract :
This paper discussed the image reconstructions using the improved compressive sensing algorithm based on the wavelet transform. First, by using the wavelet transform, the image is decomposed into low-frequency wavelet coefficients and high-frequency wavelet coefficients; Second, keep the low-frequency wavelet coefficients unchanged, use the measurement matrix to measure the high-frequency wavelet coefficients, and then combine with the unchanged low-frequency coefficients to reconstruct the image. Compared with the traditional compressive sensing algorithm, the improved compressive sensing algorithm can effectively improve the peak signal-to-noise ratio (PSNR) of the reconstructed images. Moreover, in the improved compressive sensing algorithm, we use different matrices such as Gaussian random matrix, Bernoulli matrix, Toeplitz matrix, and Hadamard matrix to reconstruct the images. Comparisons show that the reconstructed images can achieve high PSNR when using Hadamard matrix.
