文章摘要
朱宗斌,陶剑锋,葛辉良,郑佳.基于BPSO-KNN算法的被动声呐目标分类识别技术研究[J].声学技术,2019,38(2):219~223
基于BPSO-KNN算法的被动声呐目标分类识别技术研究
Passive sonar target classification and recognition technique based on BPSO-KNN algorithm
投稿时间:2018-01-09  修订日期:2018-02-28
DOI:10.16300/j.cnki.1000-3630.2019.02.018
中文关键词: 功率谱特征|被动声呐目标分类识别|特征选择|二进制粒子群最近邻算法
英文关键词: power spectrum characteristics|passive sonar target classification and recognition|feature selection|BPSO-KNN algorithm
基金项目:海军预研基金项目(30202)
作者单位E-mail
朱宗斌 中国船舶重工集团公司第七一五研究所, 浙江杭州 310023 zhuzongbin90@163.com 
陶剑锋 中国船舶重工集团公司第七一五研究所, 浙江杭州 310023  
葛辉良 中国船舶重工集团公司第七一五研究所, 浙江杭州 310023  
郑佳 中国船舶重工集团公司第七一五研究所, 浙江杭州 310023  
摘要点击次数: 849
全文下载次数: 545
中文摘要:
      以提取得到的被动声呐目标功率谱特征为基础,采用二进制粒子群(Binary Particle Swarm Optimization,BPSO)优化算法和k最近邻(k-Nearest Neighbor,KNN)分类算法相结合的BPSO-KNN算法进行特征选择和参数优化,分别用KNN分类算法和BPSO-KNN分类算法对实际得到的四类海上被动声呐目标进行分类识别。结果表明,BPSO-KNN算法可对提取的功率谱特征进行特征优化选择,并对KNN分类器进行参数优化,提高了对四类目标的分类精度。该算法在被动声呐目标分类识别方面有参考价值。
英文摘要:
      Based on the obtained power spectrum characteristics of passive sonar target, the BPSO-KNN algorithm combining binary particle swarm optimization (BPSO) algorithm and k-nearest neighbor (KNN) classification algorithm is used to carry out feature selection and parameter optimization. The comparative study is made for four types of passive sonar target recognition by using the KNN classification algorithm and the BPSO-KNN algorithm. Experimental results show that the BPSO-KNN is an effective method for both power spectrum characteristics reduction and KNN algorithm parameter optimization. And the classification accuracy of the four types of targets is improved, which shows that the algorithm has reference value in passive sonar target classification and recognition.
查看全文   查看/发表评论  下载PDF阅读器
关闭