排序方式: 共有3条查询结果,搜索用时 15 毫秒
1
1.
2.
针对高分辨率遥感影像地物信息复杂、目标识别率低等问题,提出了一种联合矩阵低秩逼近的稀疏表示遥感影像目标识别方法。对原始遥感影像进行Radon变换,将处理过后的遥感影像进行低秩和稀疏分解,得到具有低秩性和稀疏性的两部分信息;通过K-SVD算法分别对这两部分信息进行字典学习,构建稀疏表示的判别字典;通过稀疏表示求解算法求解出待分类的目标在判别字典上的稀疏系数,根据稀疏系数最大准则对目标进行分类识别。在Uc Merced数据集上选取具有代表性的线性和非线性子集分别进行实验,结果表明所提算法与传统的SRC、SVM、MLC和KNN等分类识别算法相比,在采样比例为1/16、稀疏度为5时,识别率在线性子集上能够提高10%、在非线性子集上能够提高5%,表明所提方法具有较好的识别效果。 相似文献
3.
In this paper we study the problem of blind channel identification in chaotic communications. An adaptive algorithm is proposed, which exploits the boundness property of chaotic signals. Compared with the EKF-based approach, the proposed algorithm achieves a great complexity gain but at the expense of a slight accuracy degradation.However, our approach enjoys the important advantage that it does not require the a priori information such as nonlinearity of chaotic dynamics and the variances of measurement noise and the coefficient model noise. In addition,our approach is applicable to the ARMA system. 相似文献
1