The maximum likelihood detection algorithm produces the best bit error performance in optical spatial modulation systems, but the complexity of the decoding process limits its application in practical settings. To address this deficiency, a transmission signal with sparse characteristics was generated by mapping an optical spatial modulation laser vector and a pulse position modulation vector together. A compressed sensing-based optical spatial modulation signal detection method was proposed based on the resulting transmission signal with sparse characteristics and the utilization of an orthogonal matching pursuit algorithm. Simulation results show that this method can greatly reduce computational complexity at the expense of a small increase in the performance error. In comparison with the maximum likelihood detection algorithm, the complexity of the proposed method is reduced by 99.54% when 64 lasers are used. Moreover, this approach is more suitable for wireless optical communication systems with large-scale lasers, due to the introduction of sparsity.