According to Nyquist theorem

the sampling of harmonic signal in a power system will result in a very large number of data

while revolutionary compressed sensing has broken through the limit of traditional theory. On the premise that the original signal is sparse

only a few sampling data will be sufficient to realize the reconstruction of the original signal. Based on an analysis of the existing greed matching pursuit and adaptive algorithm and the characteristics of harmonic signal

this paper proposes a new sparsity adaptive and compressive sampling matching pursuit algorithm. This algorithm is able to adaptively adjust the step to approach the original signal and achieve the goal of harmonic detection by signal proxy and backtracking

even if the sparsity of signal is unknown. The simulation results show that the proposed algorithm performs well in harmonic detection. This work is supported by National Natural Science Foundation of China (No. 51207110) and Shanghai Science and Technology Committee (No. 12ZR1451300).