将结合Barzilai-Borwein步长和非单调线搜索的梯度投影法用于压缩感知信号重构.分析了Barzilai-Borwein步长计算方法,结合其特点给出了非单调线搜索方法,为降低线搜索对算法性能的影响,引入了自适应的策略,最后给出了算法收敛性分析.实验结果表明,该算法能很好地重构不同稀疏度的信号,且在相同条件下,计算效率优于经典的基追踪法、正交匹配追踪和其他梯度投影法. A gradient projection method based on Barzilai-Borwein step and non-monotone line search is applied to compress sensing signal reconstruction.The Barzilai-Borwein step length calculation method is analyzed and a non-monotone line search method is given. In order to reduce the line search The adaptive strategy is introduced and the convergence of the algorithm is given.Experimental results show that the proposed algorithm can reconstruct signals with different sparsity well under the same conditions and the computational efficiency is better than Classical basis tracking method, orthogonal matching pursuit and other gradient projection methods.