结合背景建模的水下运动发光目标阈值分割方法

王嘉; 关夏威; 张昊; 付少波; 张宇昂; 宋清华

doi:10.11993/j.issn.2096-3920.2024-0107

结合背景建模的水下运动发光目标阈值分割方法

doi: 10.11993/j.issn.2096-3920.2024-0107

武汉第二船舶设计研究所, 湖北武汉, 430205

基金项目: 湖北省青年拔尖人才计划项目资助.

详细信息

作者简介:
王嘉：王　嘉(1996-), 男, 硕士, 工程师, 主要研究方向为水下无人系统

中图分类号: TJ630.34; TP29
计量
- 文章访问数: 16
- HTML全文浏览量: 12
- PDF下载量: 0
- 被引次数: 0
出版历程
- 收稿日期: 2024-06-05
- 修回日期: 2024-09-17
- 录用日期: 2024-09-26
- 网络出版日期: 2025-01-13

Threshold Segmentation Method for Underwater Moving Luminous Targets Combined with Background Modeling

Wuhan Second Ship Design and Research Institute, Wuhan 430205, China

摘要

摘要: 针对水下运动发光目标的视觉定位提出了一种结合背景建模的自适应阈值分割方法。首先基于高斯混合模型进行背景建模, 粗略筛选出图像中具有动态特性的若干区域; 然后在HSV色彩空间中对这些区域内像素的特征进行分析, 以此判别区域内是否存在目标; 得到包含有目标的区域后应用Otsu法计算区域内的分割阈值; 最后, 基于该阈值对包含目标区域进行二值化处理实现目标提取。该方法为缓解水下复杂场景中亮度变化和杂物干扰给视觉定位造成的影响, 充分利用了目标的运动状态以及颜色、亮度属性, 将背景建模和阈值分割相结合, 提高了分割的精准度和稳定性。试验结果表明, 该算法针对亮度变化、光晕模糊、气泡干扰等水下视觉定位中的常见问题具有较强的适应性, 且不依赖于初始参数的选取, 较为适合于工程应用。
- 运动发光目标 /
- 背景建模 /
- 阈值分割 /
- HSV /
- Otsu法
Abstract: This paper presents an adaptive threshold segmentation algorithm combined with background modeling for the visual positioning of underwater moving luminous targets. Initially, background modeling is conducted based on the Gaussian Mixture Model to roughly screen out several regions with dynamic characteristics in the image. Subsequently, the features of the pixels within these regions are analyzed in the HSV color space to determine the presence of a target. After identifying regions containing the target, the Otsu method is applied to calculate the segmentation threshold within these regions. Finally, binarization processing is performed on the regions containing the target based on the calculated threshold to achieve target extraction. The algorithm is designed to alleviate the impact of brightness variations and clutter interference on visual positioning in complex underwater scenes. It fully utilizes the target’s motion state as well as its color and brightness attributes, combining background modeling with threshold segmentation to enhance the precision and stability of the segmentation process. Experimental results indicate that the algorithm demonstrates strong adaptability to common issues in underwater visual positioning, such as changes in brightness, halo blur, and bubble interference, and it is not dependent on the selection of initial parameters, making it suitable for engineering applications.
- moving luminous target /
- background modeling /
- threshold segmentation /
- HSV /
- Otsu method

HTML全文

图 1 绿色点光源在不同色彩空间中的成像

Figure 1. Images of green point light in different color spaces

下载: 全尺寸图片幻灯片

图 2 有气泡干扰情况下几种背景建模方法的处理结果

Figure 2. Processing results of several background modeling methods under the interference of bubbles

下载: 全尺寸图片幻灯片

图 3 Otsu法和三角法处理结果

Figure 3. Processing results of Otsu method and triangle method

下载: 全尺寸图片幻灯片

图 4 文中算法流程图

Figure 4. Flow chart of the algorithm in this paper

下载: 全尺寸图片幻灯片

图 5 不同算法对于绿色点光源目标的分割效果对比

Figure 5. Comparison of segmentation for green point light source targets using different algorithms

下载: 全尺寸图片幻灯片

参考文献(18)

[1]	李宇波, 朱效洲, 卢惠民, 等. 视觉里程计技术综述[J]. 计算机应用研究, 2012, 29(8): 2801-2805. doi: 10.3969/j.issn.1001-3695.2012.08.001 LI Y B, ZHU X Z, LU H M, et al. Review on visual odometry technology[J]. Application Research of Computers, 2012, 29(8): 2801-2805. doi: 10.3969/j.issn.1001-3695.2012.08.001
[2]	张辉. 水的光学特性对水下光学成像质量影响的分析[J]. 电子测试, 2013(20): 261-262. doi: 10.3969/j.issn.1000-8519.2013.20.125 Zhang H. Analysis of optical properties of water on the influence of optical imaging quality under water[J]. Electronic Test, 2013(20): 261-262. doi: 10.3969/j.issn.1000-8519.2013.20.125
[3]	向文豪. 基于高斯混合模型的海面舰船目标检测[J]. 舰船科学技术, 2024, 46(1): 148-151. XIANG W H. Surface ship target detection based on Gaussian mixture model[J]. Ship Science and Technology, 2024, 46(1): 148-151.
[4]	王秀芬, 王汇源, 王松. 基于背景差分法和显著性图的海底目标检测方法[J]. 山东大学学报(工学版), 2011, 41(1): 12-16. WANG X F, WANG H Y, WANG S. Underwater object detection based on background subtraction and a saliency map[J]. Journal of Shandong University(Engineering Science), 2011, 41(1): 12-16.
[5]	雷飞, 黄文路, 王雪丽. 基于YUV颜色空间码本模型的水下运动目标检测[J]. 计算机与应用化学, 2014, 31(4): 416-420. LEI F, HUANG W L, WANG X L. Underwater moving targets detection in YUV color space[J]. Computers and Applied Chemistry, 2014, 31(4): 416-420.
[6]	徐武, 王欣达, 高寒, 等. 融合改进人工蜂群与Otsu的图像分割算法[J]. 计算机仿真, 2023, 40(6): 229-233. doi: 10.3969/j.issn.1006-9348.2023.06.042 XU W, WANG X D, GAO H, et al. Image segmentation algorithm by fusion of improved artificial bee colony and Otsu[J]. Computer Simulation, 2023, 40(6): 229-233. doi: 10.3969/j.issn.1006-9348.2023.06.042
[7]	徐硕, 姜言清, 李晔, 等. 智能水下机器人自主回收的双目视觉定位[J]. 哈尔滨工程大学学报, 2022, 43(8): 1084-1090. doi: 10.11990/jheu.202106079 XU S, JIANG Y Q, LI Y, et al. A stereo vision localization method for autonomous recovery of autonomous underwater vehicle[J]. Journal of Harbin Engineering University, 2022, 43(8): 1084-1090. doi: 10.11990/jheu.202106079
[8]	STAUFFER C, GRIMSON W E L. Adaptive background mixture models for real-time tracking[C//IEEE Computer Vision & Pattern Recognition. Fort Collins, USA: IEEE, 1999.
[9]	余鹏, 封举富. 基于高斯混合模型的纹理图像分割[J]. 中国图象图形学报, 2005, 10(3): 281-285. doi: 10.3969/j.issn.1006-8961.2005.03.003 YU P, FENG J F. Texture image segmentation based on Gaussian mixture models[J]. Journal of Image and Graphics, 2005, 10(3): 281-285. doi: 10.3969/j.issn.1006-8961.2005.03.003
[10]	彭明阳, 王建华, 闻祥鑫, 等. 结合HSV空间的水面图像特征水岸线检测[J]. 中国图象图形学报, 2018, 23(4): 526-533. doi: 10.11834/jig.170498 PENG M Y, WANG J H, WEN X X, et al. Shoreline detection method by combining HSV spatial water image feature[J]. Journal of Image and Graphics, 2018, 23(4): 526-533. doi: 10.11834/jig.170498
[11]	KAEWTRAKULPONG P, BOWDEN R. An improved adaptive background mixture model for real-time tracking with shadow detection[J]. Kluwer Academic Publishers, 2002.
[12]	ZIVKOVIC Z, HEIJDEN F V D. Efficient adaptive density estimation per image pixel for the task of background subtraction[J]. Pattern Recognition Letters, 2006, 27(7): 773-780. doi: 10.1016/j.patrec.2005.11.005
[13]	SAGI Z. Background subtractor CNT[CP/OL]. [2024-03-03]. https://github.com/sagi-z/BackgroundSubtractorCNT.
[14]	韩思奇, 王蕾. 图像分割的阈值法综述[J]. 系统工程与电子技术, 2002, 24(6): 91-94. doi: 10.3321/j.issn:1001-506X.2002.06.027 HAN S Q, WANG L. A Survey of Thresholding Methods for Image Segmentation[J]. Systems Engineering and Electronics, 2002, 24(6): 91-94. doi: 10.3321/j.issn:1001-506X.2002.06.027
[15]	王嘉, 关夏威, 付少波, 等. 基于先验信息的光照鲁棒性水下视觉定位方法[J]. 舰船科学技术, 2024, 46(10): 98-101. WANG J, GUAN X W, FU S B, et al. A robust underwater visual positioning method for brightness based on prior information[J]. Ship Science and Technology, 2024, 46(10): 98-101.
[16]	OTSU N. A threshold selection method from Gray-Level histograms[J]. IEEE Transactions on Systems, Man, and Cybernetics, 1979, 9(1): 62-66. doi: 10.1109/TSMC.1979.4310076
[17]	ZACK G W, ROGERS W E, LATT S A. Automatic measurement of sister chromatid exchange frequency[J]. Journal of Histochemistry & Cytochemistry Official Journal of the Histochemistry Society, 1977, 25(7): 741-753.
[18]	FISHER R B. Change detection in color images[EB/OL]. [2024-02-21]. https://www.semanticscholar.org/paper/Change-Detection-in-Color-Images-Fisher/0314ed6d26f63deeec2bb251771c3c17d07818c1.