提出一种新的压缩传感稀疏信号重构算法——残差最小化追踪(residual minimization pursuit,RMP).残差最小化追踪RMP每次迭代选择残差信号在测量矩阵的正交投影绝对值最大的元素来检测支持集,然后求解支持集上的最小二乘解更新稀疏信号.另外,提出两种扩展残差最小化追踪RMP算法,算法每次迭代选择多个元素来检测支持集.实验结果表明,残差最小化追踪RMP稀疏重构性能优于正交匹配追踪OMP算法. A new compression sensing sparse signal reconstruction algorithm, residual minimization pursuit (RMP), was proposed.Residual minimization tracking (RMP) was performed at each iteration of the RMP to select the residual signal in the orthogonal projection of the measurement matrix The largest element to detect the support set, and then solve the sparse signal updating least-squares solution on the support set.In addition, two extended residual minimization tracking RMP algorithm is proposed, the algorithm selects multiple elements per iteration to detect the support set.Experiment The results show that residual minimization tracking RMP sparse reconstruction performance is better than orthogonal matching pursuit OMP algorithm.