Currently, there are lots of image reconstruction algorithms are used in literatures such as the least 1-norm, matching pursuit, iterative threshold and so on. While, these methods are suboptimization. The algorithm is proposed in this paper to resolve the multimodal optimization problem of the least 0-norm for compressive sensing image reconstruction. The result shows that the method is feasible.%近年来由Donoho和Candes等人提出的压缩感知图像处理有效地解决了图像高速采样与压缩重构之间的瓶颈问题,使得采样与压缩同时进行,并有效利用采样所得到的数据,用于后期的图像重构中.目前文献中使用的重构算法很多,如最优l1范数法、匹配追踪等贪婪算法、迭代阈值法等,但这些方法都是次优化算法,没有从压缩感知最初需要解决的问题出发.在此给出的算法是从压缩感知重构的最初需要解决的问题出发,寻找一种能够解决最优l0范数的多峰优化问题的算法.实验结果也证明了该方法的可行性.
展开▼
