摘要
海水中声速剖面反演可以及时获取海洋环境信息,对于改善和提高水声设备的工作性能以及海洋的研究和开发都具有重要意义。针对声速剖面反演过程中,神经网络存在容易过早收敛的不足,该文提出了基于双种群约束QPSO-BP的声速剖面反演方法。利用Argo获取的温度和盐度数据,以2004—2017年经验正交函数所得到的特征向量和历史声速剖面作为训练样本,以BP神经网络反演声速剖面的模型作为基础,并在QPOS-BP算法基础上引入了双种群约束策略,通过与2018年6月和12月的实测声速剖面数据进行比较,双种群约束QPSO-BP声速预测模型在精度上比BP神经网络和QPSO-BP网络模型分别平均提高了35%和25%。结果表明,双种群约束QPSO-BP能有效提高声速剖面反演精度。
The inversion of sound velocity profile in seawater can obtain the information of marine environment in time, which is of great significance to improve the performance of underwater acoustic equipment and the research and development of the ocean. In order to solve the problem of premature convergence of neural network in the process of inversion of sound velocity profile, a new method of inversion of sound velocity profile based on quantum particle swarm optimization-back propagation(QPSO-BP)with two population constraints is proposed in this paper.Based on the temperature and salinity data obtained by Argo,the eigenvector and historical sound velocity profile obtained from empirical orthogonal function(EOF)in 2004-2017 were used as training samples based on the model of BP neural network inversion of sound speed profile.On the basis of QPSO-BP algorithm,a dual population constraint strategy was introduced in this paper.Compared with the measured sound speed profile data in June and December 2018,the accuracy of the prediction model of dual population constraint QPSO-BP was 35% and 25% higher than that of BP neural network and QPSO-BP network respectively.The results showed that QPSO-BP with two population constraints could effectively improve the inversion accuracy of sound velocity profile.
作者
孙佳龙
张杰
唐玥
孙苗
张正阳
张弛
SUN Jialong;ZHANG Jie;TANG Yue;SUN Miao;ZHANG Zhengyang;ZHANG Chi(Key Laboratory of Marine Environmental Information Technology,MNR,Tianjin 300171,China;School of Marine Technology and Geomatics,Jiangsu Ocean University,Lianyungang,Jiangsu 222005,China;Jiangsu Institute of Marine Resources Development,Lianyungang,Jiangsu 222001,China;Zhangjiagang Nanfeng Water Conservancy Management Service Station,Suzhou,Jiangsu 215600,China)
出处
《测绘科学》
CSCD
北大核心
2021年第8期127-134,共8页
Science of Surveying and Mapping
基金
国家自然科学基金项目(40974016)
自然资源部海洋信息技术创新中心开放基金课题、连云港高新区重点研发计划项目(ZD201905)
江苏省高校海洋科学技术优势学科建设项目、测绘工程国家一流本科专业建设点建设项目。
