Page 237 - 应用声学2019年第4期
P. 237
第 38 卷 第 4 期 Vol. 38, No. 4
2019 年 7 月 Journal of Applied Acoustics July, 2019
⋄ 李启虎院士八十华诞学术论文 ⋄
基于超椭圆拟合的水下小目标分类 ∗
王 梁 1,2,3 田 杰 1,3† 黄海宁 1,3 薛山花 1,3
(1 中国科学院声学研究所 北京 100190)
(2 中国科学院大学 北京 100149)
(3 中国科学院先进水下信息技术重点实验室 北京 100190)
摘要 水下小目标分类技术在海底探测、水下考古等方面应用广泛。在实际的水下声图像中,小目标投影产生
的阴影区域通常在形状和尺寸方面显著于目标本身产生的亮区,故阴影分析算法对于目标的检测、识别和分
类均有重要的研究意义。该文采用超椭圆曲线拟合算法拟合目标阴影区域,通过控制超椭圆函数的几个参数
变化,实现不同的超椭圆曲线拟合不同的目标阴影形状,并将控制超椭圆曲线尺寸、形状和位置的参数作为特
征向量输入到分类器,通过对比多个分类器得出分类结果,证明了以拟合参数为特征的分类方法有效。
关键词 水下小目标,目标阴影,超椭圆拟合,目标分类
中图法分类号: TB566 文献标识码: A 文章编号: 1000-310X(2019)04-0697-08
DOI: 10.11684/j.issn.1000-310X.2019.04.029
Underwater small target classification based on superellipse fitting technology
WANG Liang 1,2,3 TIAN Jie 1,3 HUANG Haining 1,3 XUE Shanhua 1,3
(1 Institute of Acoustics, Chinese Academy of Sciences, Beijing 100190, China)
(2 University of Chinese Academy of Sciences, Beijing 100049, China)
(3 Key Laboratory of Science and Technology on Advanced Underwater Acoustic Signal Processing,
Chinese Academy of Sciences, Beijing 100190, China)
Abstract Underwater small target classification technology is widely used in seabed detection, archaeology
and so on. In the actual underwater acoustic image, the shape and size of shadow area, which produced by
underwater small target, are usually more significant than bright area generated by the target itself, so shadow
analysis algorithms have important research value in detection, identification and classification of target. In
this paper, the superellipse fitting algorithm is used to fit the shadow area of the target. By controlling several
parameters of superellipse function, different superellipse curves can be used to fit different target shadow
shape. The parameters controlling the size, shape and position of superellipse curve are input into the classifier
as feature vectors. By comparing the classification results of several classifiers, it is demonstrated that the
classification method characterized by fitting parameters is effective.
Key words Underwater small target, The shadow of target, Superellipse fitting, Target classification
2019-03-12 收稿; 2019-06-10 定稿
国家自然科学基金项目 (11174313),“问海计划”项目 (SQ2017WHZZB0701-3-2)
∗
作者简介: 王梁 (1993- ), 女, 天津人, 硕士研究生, 研究方向: 数字图像处理。
† 通讯作者 E-mail: tianjie@mail.ioa.ac.cn