A block sparse compressive sensing based infrared image reconstruction method and system. The method comprises: a block sparse representation step, representing continuous infrared small target image sequences as signals having block sparse characteristics (S1); a temporal correlation modeling step, modeling, in a Bayesian framework, the temporal correlation of data between the continuous infrared small target image sequences (S2); a signal reconstruction and output step, reconstructing the infrared small target image using a sparse learning algorithm, and outputting the final estimated value of the signals (S3). The method and the system, during the compressive sensing construction process, utilize the correlation of continuous signals in time to reconstruct the signals, and can improve the reconstruction precision of single-frame signals, and substantially reduce the running time of the algorithm.
展开▼