《计算机应用研究》|Application Research of Computers

基于改进区域生长的水岸线提取方法研究

Research on waterfront line extraction based on improved region growing method

免费全文下载 (已被下载 次)  
获取PDF全文
作者 郑又能,肖宇峰,寸超,向可均,张华,刘冉
机构 西南科技大学 信息工程学院,四川 绵阳 621000
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2020)06-057-1876-06
DOI 10.19734/j.issn.1001-3695.2018.12.0917
摘要 针对无人船在水面目标识别及视觉导航中涉及到的水岸分割问题,提出了一种基于改进区域生长的水岸线提取方法。首先在Lab颜色空间对图像进行阈值分割,得到水面种子点候选区域;然后通过构建最小二乘问题在种子点候选区域自动选取最优初始种子点,并基于图像标准差实现生长规则阈值自适应;最后依据初始种子点和生长规则进行区域生长,并对得到的水面区域进行边缘提取,从中分离出水岸分界线。采用该方法对水岸样本图像进行实验,通过相关系数和偏移误差对水岸线提取结果进行评价。实验结果表明,该方法能够提取出复杂环境下的水岸线,具有一定的鲁棒性,且基本满足实时性要求。
关键词 水岸线提取; 区域生长; 最小二乘法; 阈值
基金项目 国家核能开发科研项目([2016]1295)
国家自然科学基金资助项目(61601381)
西南科技大学博士基金资助项目(13zx7146)
四川省重点研发计划资助项目(19GJHZ0197)
本文URL http://www.arocmag.com/article/01-2020-06-057.html
英文标题 Research on waterfront line extraction based on improved region growing method
作者英文名 Zheng Youneng, Xiao Yufeng, Cun Chao, Xiang Kejun, Zhang Hua, Liu Ran
机构英文名 School of Information Engineering,Southwest University of Science & Technology,Mianyang Sichuan 621000,China
英文摘要 Concerning waterfront segmentation for unmanned ships when performing water surface target recognition and visual navigation tasks, this paper proposed a waterfront line extraction method based on improved region growing. First, the method segmented the image in the Lab color space to obtain a seed point candidate area. Then, it constructed the least squares problem to select the optimal initial seed point in the seed point candidate area, and got the threshold of growth rule by calculating the image standard deviation. Finally, it carried out the region growing, and extracted edge of the obtained water surface area, from which separating the waterfront line. It tested the waterfront sample images by this method, and evaluated the results by correlation coefficient and offset error. The experimental results show that the proposed method can extract the waterfront line in complex environment, and the method is robust and real-time.
英文关键词 waterfront line extraction; region growing; least squares; threshold
参考文献 查看稿件参考文献
 
收稿日期 2018/12/11
修回日期 2019/1/14
页码 1876-1881
中图分类号 TP391.41
文献标志码 A