利用区域信息的航拍图像分割

杨名宇; 李刚

doi:10.3788/CO.20140705.0779

利用区域信息的航拍图像分割

doi: 10.3788/CO.20140705.0779

杨名宇^1,2, ,,
李刚^1,2

1.
中国科学院长春光学精密机械与物理研究所, 吉林长春 130033;
2.
中国科学院航空光学成像与测量重点实验室, 吉林长春 130033

基金项目:

国家自然科学基金资助项目（No.61308099）

详细信息

作者简介:
杨名宇（1983- ），男，吉林松原人，博士，助理研究员，2012年于中国科学院自动化研究所获得博士学位，主要从事航空图像处理方面的研究。

通讯作者:
杨名宇，E-mail：ymy1983@163.com

中图分类号: TP391.4
计量
- 文章访问数: 2001
- HTML全文浏览量: 667
- PDF下载量: 574
- 被引次数: 0
出版历程
- 收稿日期: 2014-04-14
- 修回日期: 2014-06-16
- 刊出日期: 2014-09-25

Aerial image segmentation with region information

YANG Ming-yu^{1,2
, ,},
LI Gang^1,2

1.
Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun 130033, China;
2.
Key Laboratory of Airborne Optical Imaging and Measurement, Chinese Academy of Sciences, Changchun 130033, China

摘要

摘要: 提出一种利用区域信息的航拍图像分割模型。针对GAC模型和Chan-Vese模型存在的不足，提出一种符号压力函数，该符号压力函数可以有效地增大模型的作用范围。与Chan-Vese模型相比，新模型不受初始条件的限制，进一步增大了模型的作用范围。新模型利用了图像的区域信息，可以同时将目标的内外边界分割出来。在新模型中，水平集函数不必初始化为符号距离函数，节省了计算开销。与传统的基于水平集方法的模型相比，新模型不含曲率项，实现简单。实验结果表明，与GAC模型和Chan-Vese模型相比，新模型的分割精度高于3%，分割速度快6倍以上。
- 图像分割 /
- 航拍图像 /
- 区域信息
Abstract: Image segmentation based on level set method is one of the most widely used methods in segmentation domain. Due to a large field of view for aerial image, traditional methods usually can not obtain a global segmentation, and segmentation result is often very poor. In this paper, a new segmentation method with region information for aerial image is proposed. First, a new signed press function is proposed to enhance the capture range, which also can obtain global segmentation result. Second, compared with Chan-Vese model, proposed model is not limited by initial condition, and can enhance the capture range further. Third, proposed model utilizes the region information of an image, which can automatically segment inside and outside of an object simultaneously. In the proposed model, level set function doesn't need to be initialized to signed distance function, which can reduce a lot of computation cost. Moreover, the proposed model has no curvature item compared with traditional level set method, and is easy for numerical implementation. Experiment results demonstrate that proposed model has a higher accuracy of 3%, and 6 times faster than GAC model and Chan-Vese model.
- image segmentation /
- aerial image /
- region information

HTML全文

参考文献(1)

[1] KASS M, WITKIN A, TERZOPOULOS D. Snakes:active contour models[J]. International J. Computer Vision, 1988, 1(4):321-332.
[2] 王醒策, 张美霞, 武仲科, 等. 基于全局LBF水平集模型的脑血管层次粗分割[J]. 光学精密工程, 2013, 21(12):3283-3297. WANG X C, ZHANG M X, WU ZH K, et al. Level coarse brain vessel segmentation based on global LBF model[J]. Opt. Precision Eng., 2013, 21(12):3283-3297.(in Chinese)
[3] 王卫星, 苏培垠. 基于颜色、梯度矢量流活动轮廓及支持向量机实现白细胞的提取和分类[J]. 光学精密工程, 2012, 20(12):2781-2790. WANG W X, SU P Y. Blood cell image segmentation on color and GVF snake for Leukocyte classification on SVM[J]. Opt. Precision Eng., 2012, 20(12):2781-2790.(in Chinese)
[4] LI CH M, KAO C Y, GORE J C, et al. Minimization of region-scalable fitting energy for image segmentation[J]. IEEE Trans Image Processing, 2008, 17(10):1940-1949.
[5] 王成艳, 刘晶红, 楚广生, 等. 图像场景区分于航空摄像机自动调光方法研究[J]. 液晶与显示, 2013, 28(6):948-954. WANG CH Y, LIU J H, CHU G SH, et al. Scene distinguish and aerial camera auto-dimming[J]. Chinese J. Liquid Crystal and Displays, 2013, 28(6):948-854.(in Chinese)
[6] CASELLES V, KIMMEL R, SAPIRO G. Geodeisic active contours[J]. International J. Computer Vision, 1997, 22(1):61-79.
[7] CHAN F, VESE L. Active contours without edges[J]. IEEE Trans Image Processing, 2001, 10(2):266-177.
[8] SETHIAN J A. Level Set Methods and Fast Marching Methods:Evolving Interfaces in Computational Geometry, Fluid Mechanics, Computer Vision, and Materials Science[M]. London:Cambridge University Press, 1999.
[9] OSHER S, SETHIAN J A. Fronts propagating with curvature dependent speed: algorithms based on Hamilton-Jacobi formulation[J]. J. Computer Physics, 1988, 79(1):12-49.
[10] MALLADI R, SETHIAN J A, VERNURI B C. Shape modeling with front propagation:a level set approach[J]. IEEE Trans Pattern Analysis and Machine Intelligence, 1995, 17(2):158-175.
[11] MUMFORD D, SHAH J. Optimal approximation by piecewise smooth functions and associated variational problems[J]. Comm. Pure Appl. Math., 1989, 42(5):577-685.
[12] XU C Y, YEZZI A, PRINCE J L. On the relationship between parametric and geometric active contours[C]. Processing of 34^th Asilomar Conference on Signals Systems and Computer, 2000:483-489.
[13] PERONA P, MALIK J. Scale-space and edge detection using anisotropic diffusion[J]. IEEE Trans Pattern Analysis and Machine Intelligence, 1990, 12(7):629-640.
[14] ZHANG K H, ZHANG L, SONG H H, et al. Active contours with selective local or global segmentation:a new formulation and level set method[J]. Image and Vision Computing, 2010, 28(2):668-676.