摘要
为快速辨别海底底质类型和海底目标,在分析Kohonen自组织特征映射网络(Self-Organizing Feature Map,SOFM)和学习向量量化(Learning Vector Quantization,LVQ)算法的基础上,提出一种SOFM算法与改进的LVQ算法相结合的混合神经网络分类方法.利用这种分类方法,对预处理后的多波束测深系统获取的反向散射强度数据进行训练分类.通过对在实验区域提取的检测样本的分类结果进行比较分析,表明该方法是可行、有效的,而且在底质类型特征相近的情况下,具有较好的分类效果.
To identify the types of seabed sediments and target quickly, a hybrid neural network classifi- cation method, which combines the Self-Organizing Feature Map （SOFM） algorithm with the modified Learning Vector Quantization （LVQ） algorithm, is proposed by analyzing SOFM and LVQ algorithms de- veloped by Kohonen. The method is used to train and classify the preprocessed seabed backscatter strength data obtained by the multibeam system. By the comparison and analysis on classification results of the test samples in the experimental area, it shows that the method is feasible and effective. In the Case of similar seabed sediments, this method is of good classification effect.
出处
《上海海事大学学报》
北大核心
2013年第4期27-30,共4页
Journal of Shanghai Maritime University
基金
上海海事大学研究生创新基金(yc2012034)
关键词
底质分类
反向散射强度
自组织特征映射
学习向量量化
seabed classification
backscatter strength
self-organizing feature map
learning vector quantization
