Based on compressed sensing, the sparsity-feature of short-time power quality disturbance signals is not considered in the existing reconstruction methods, and thus we can further improve its reconstruction performance. To this end, a sparsity-feature based reconstruction method is proposed in this paper. Firstly, the signals are sampled according to the compressed sensing theory. Then, the sparsity-feature of short-time power quality disturbance signals, i.e., its sparsity in frequency-domain is even, is developed. With the developed feature, a reconstruction method referred to as double step-size sparsity adaptive matching pursuit (DS-SAMP) is proposed. Compared with the conventional sparsity adaptive matching pursuit (SAMP) algorithm, the analysis and simulation results show that the proposed method reduces the computational complexity and mean square error (MSE), improves the signal-to-noise ratio (SNR) and the probability of correct reconstruction.
