摘要
针对水下较复杂图像处理,提出了一种基于双层孔洞型掩膜的Chan-Vese改进模型。模型通过增加双层孔洞掩膜,改善传统Chan-Vese模型在处理水下复杂图像孔洞区域时的不足,从而提高分割效果,进而实现目标标定。在实验中,对多种水下图像进行分割实验,在此分割基础上使用HSV图像提取特定颜色目标物,实现目标标定。实验结果表明,与传统Chan-Vese模型相比,上述模型分割效果明显,具有更好的鲁棒性和适应性,更适用于目标提取与标定,能够更好地处理水下图像中的孔洞和噪声,可以为水下较复杂图像处理及目标标定提供一种有效的解决方案。
For underwater complex image processing,an improved Chan-Vese model based on a double-layer hole-type mask is proposed in this paper.An improved model was proposes based on the dual-layer porous mask for underwater image processing,which model enhances the traditional Chan-Vese model's limitations in handling hole regions in complex underwater images,thereby improving the segmentation effect and achieving target calibration.In the experimental section,various underwater images were subjected to segmentation experiments.Based on this segmentation,the HSV image was used to extract specific colored objects for target calibration.The experimental results demonstrate that compared to the traditional Chan-Vese model,this model achieves significant segmentation improvement and exhibits better robustness and adaptability.It is more suitable for target extraction and calibration,effectively handling holes and noise in underwater images.Thus,it provides an effective solution for the processing of complex underwater images and target calibration.
作者
王鑫
杨晨宇
刘丹
张惠宇
WANG Xin;YANG Chen-yu;LIU Dan;ZHANG Hui-yu(North University of China,Key Laboratory of Instrumentation Science&Dynamic Measurement,Ministry of Education,Taiyuan Shanxi 030051,China)
出处
《计算机仿真》
2025年第5期372-377,共6页
Computer Simulation
基金
山西省重点研发计划(2021020101010)
山西省回国留学人员科研资助项目(2022-144)
中北大学研究生科技专项资金(20221865)
成都信息工程大学四川协同创新中心“基于退化光声混合信道的水下航行器集群通信技术研究”项目(CXPAQ202205)。
关键词
双层孔洞掩膜
水下图像分割
目标标定
Double-layer porous mask
Underwater image segmentation
Target calibration
作者简介
王鑫(1984-),男(汉族),内蒙古林西人,讲师,主要研究领域为机器人构型与随动控制;杨晨宇(1998-),女(汉族),甘肃平凉人,硕士研究生,主要研究领域为水下潜航器自主控制;刘丹(1988-),男(汉族),山西忻州人,教授,硕士研究生导师,主要研究领域为水下无人系统和多智能体协同控制;张惠宇(1998-),男(汉族),山西晋城人,硕士研究生,主要研究领域为水下多机器人协同控制研究。
