提出了种并行的稀疏分解算法，以解决目前的稀疏分解算法计算量大、耗时多，不适宜用作战场声目标实时识别的缺点。-Kinds of parallel sparse decomposition algorithm to solve the sparse decomposition algorithm large, time-consuming, and not suitable for real-time recognition of the shortcomings of the battlefield acoustic target.