压缩感知中基于广义Jaccard系数的gOMP重构算法

发布时间：2018-09-01 17:01

【摘要】：为了解决信号重构性能差的问题,提出了一种基于广义Jaccard系数的广义正交匹配追踪(generalized orthogonal matching pursuit,g OM P)重构算法。该算法利用广义Jaccard系数相似性匹配准则替换g OM P算法中的内积度量准则,优化了通过感知矩阵来选择与残差余量最匹配原子的匹配方式。实验结果表明,该算法的重构成功率不仅高于g OMP算法,同时也高于OMP、St OMP等算法。
[Abstract]:In order to solve the problem of poor signal reconstruction performance, a generalized orthogonal matching tracing (generalized orthogonal matching pursuit,g OM P) reconstruction algorithm based on generalized Jaccard coefficients is proposed. In this algorithm, the generalized Jaccard coefficient similarity matching criterion is used to replace the inner product metric in the g OM P algorithm, and the matching method of the best matching atom with the residual residue is optimized by the perceptual matrix. Experimental results show that the reconstruction success rate of this algorithm is not only higher than that of g OMP algorithm, but also higher than that of OMP,St OMP algorithm.
【作者单位】：兰州交通大学自动化与电气工程学院;兰州交通大学光电技术与智能控制教育部重点实验室;兰州理工大学机电工程学院;西北工业大学航天学院;
【基金】：国家自然科学基金资助项目(51465034) 甘肃省高校基本科研业务费资助项目(213063)
【分类号】：TN911.7

【相似文献】