一种改进的高分辨率遥感影像分割方法

高伟; 刘修国; 彭攀; 陈启浩

doi:10.3799/dqkx.2010.050

一种改进的高分辨率遥感影像分割方法

doi: 10.3799/dqkx.2010.050

高伟^{1, 2,},
刘修国^{1, 2},
彭攀^{1, 2},
陈启浩^{1, 2}

1.
中国地质大学信息工程学院，湖北武汉 430074
2.
地理信息系统软件及其应用教育部工程中心，湖北武汉 430074

基金项目:

教育部新世纪优秀人才计划资助项目 NCET-07-0772

国家重点“863”项目 2007AA120503

详细信息

作者简介:
高伟(1980-)，男，博士生，主要从事遥感数据处理与信息提取的科研和教学工作.E-mail: cuggaowei@hotmail.com

中图分类号: TP311
计量
- 文章访问数: 3143
- HTML全文浏览量: 104
- PDF下载量: 50
- 被引次数: 0
出版历程
- 收稿日期: 2010-01-15
- 刊出日期: 2010-05-01

An Improved Method of High-Resolution Remote Sense Image Segmentation

GAO Wei^{1, 2
,},
LIU Xiu-guo^{1, 2},
PENG Pan^{1, 2},
CHEN Qi-hao^{1, 2}

1.
Faculty of Information Engineering, China University of Geosciences, Wuhan 430074, China
2.
Engineering Research Center for GIS Software and Applications, Ministry of Education, Wuhan 430074, China

摘要

摘要: 分形网络演化算法是面向对象的遥感影像分类中比较成熟的一种构建对象的算法，但在分割效率上有待进一步提高，而四叉树分割是一种高效的图像分割方法.提出了一种基于四叉树预分割的分形网络演化构建对象的方法.实验证明，该方法基本不影响影像分割的效果，而且提高了形成初始对象的效率，较大程度上提高了整体的分割效率.
- 影像分割 /
- 分形网络演化 /
- 高空间分辨率 /
- 四叉树 /
- 地理信息系统
Abstract: Fractal net evolution approach (FNEA) is a relatively mature one among the object-oriented image segmentation algorithms, but its efficiency is to be improved. Quad-Tree segmentation is a kind of effective image segmentation method. In this paper, an improved object oriented multi-scale image segmentation method based on the quad-tree pre-segmentation and FNEA is introduced. The experiment shows that the effect of its segmentation result is almost the same as that of traditional FNEA method. Moreover, It saves time.
- image segmentation /
- fractal net evolution approach /
- high-nesolution remote sense /
- quad-tree /
- geographic information system (GIS)

HTML全文

图 1 第1次分割(a)和第2次分割(b)示意

Fig. 1. The result of the first segmentation (a) and the second segmentation (b)

下载: 全尺寸图片幻灯片

图 2 四叉树分割过程示意

Fig. 2. Sketch of the quad-tree segmentation process

下载: 全尺寸图片幻灯片

图 3 传统的分形网络与改进后的处理流程对比

Fig. 3. Compare between traditional and the improved segmentation process

下载: 全尺寸图片幻灯片

图 4 原始影像(a)、传统的FNEA分割结果(b)、四叉树阈值45尺度50形状参数0.2(c)和四叉树阈值60尺度50形状参数0.2(d)

Fig. 4. Original image (a) and the traditional segmentation results based on FNEA (b) and the result based on quad-tree (threshold: 45) (c) and the result based on quad-tree (threshold: 60) (d)

下载: 全尺寸图片幻灯片

表 1 实验分析

Table 1. Experience analysis

分割方法	时间(s)	参数
传统的FNEA分割方法	18.49	尺度参数：50，形状参数：0.2
基于四叉树预分割的FNEA分割方法	10.17	四叉树阈值：45
	10.17	尺度参数：50，形状参数：0.2
	8.95	四叉树阈值：60
	8.95	尺度参数：50，形状参数：0.2

下载: 导出CSV

参考文献(18)

[1]	Baatz, M., Schape, A., 2000. Multiresolution segmentation: an optimization approach for high quality multiscale image segmentation. In: Strobl, J., Blaschke, T., Griesebner, G., eds., Angewandte geographische informationsverarbeitung Ⅻ. Wichmann-Verlag. Heidelberg, 12-23.
[2]	Blaschke, T., Hay, G.J., 2001. Object-oriented image analysis and scale-space: theory and methods for modeling and evaluateng multi-scale landscape structure. International Archives of Photogrammetry and Remote Sensing, 34: 22-29.
[3]	Chen, Q.H., 2007. Research and realization of multi-sources remote sensing data on object oriented (Dissertation). China University of Geosciences, Wuhan (in Chinese).
[4]	Chen, Q.H., Liu, X.G., Gao, W., et al., 2009. A fast and efficient high spatial resolution multi-scale image segmentation strategy. The 2nd International Conference on Earth Observation for Global Changes, Chengdu.
[5]	Chen, Z., Zhao. Z., Gong, P., et al., 2006. A new process for the segmentation of high resolution remote sensing imagery. International Journal of Remote Sensing, 27(22): 4991-5001. doi: 10.1080/0143116060658131
[6]	Gao, W., 2006. Image fusion based on trous wavelet transfrom. Earth Science—Journal of China University of Geosciences, 31(Suppl. ): 132-135 (in Chinese with English abstract). http://en.cnki.com.cn/Article_en/CJFDTOTAL-DQKX200203024.htm
[7]	Hu, X.Y., Tao, C.V., Prenzel, B., 2005. Automatic segmentation of high-resolution satellite imagery by integrating texture, intensity and color features. Photogrammetric Engineering and Remote Sensing, 71(12): 1399-1406. doi: 10.14358/PERS.71.12.1399
[8]	Lu, G.X., Zhou, D.W., Wang, J.L., et al., 2002. Geological information extracting from remote sensing image in complex area: based on wavelet analysis for automatic image segmentation. Earth Science—Journal of China University of Geosciences, 27(1): 50-54 (in Chinese with English abstract). http://www.cqvip.com/main/zcps.aspx?c=1&id=5786777
[9]	Mo, D.K., Lin, H., 2005. A robust approach toward high resolution remote sensing image segmentation. In: China Society of Image and Graphics, ed., The 12th national conference on image and graphics (NCIG2005). Beijing, 244-249 (in Chinese).
[10]	Peng, P., Liu, X.G., Gao, W., et al., 2009. An improved strategy for object-oriented multi-scale remote sensing image segmentation, ICISE2009, Nanjing. doi: 10.1109/ICISE.2009.280
[11]	Ren, X.F., Ji, G.R., Ji, G.Y., 2007. Improved fractral image coding approach based on quadtree partition. Control & Automation, 23(10-3): 291-293 (in Chinese with English abstract).
[12]	Song, Y.B., Zhang, B.Q., Hao, Y.P., 2004. Quad-tree-basing structure image segment technology. Ordnance Industry Automation, 23(6): 63-64, 69 (in Chinese with English abstract). http://en.cnki.com.cn/Article_en/CJFDTOTAL-BGZD20040600U.htm
[13]	陈启浩, 2007. 面向对象的多源遥感数据分类技术研究与实现(硕士学位论文), 武汉: 中国地质大学.
[14]	高伟, 2006. IKONOS影像融合技术及其在土地监测中的应用. 地球科学——中国地质大学学报, 31(增刊): 132-135. https://www.cnki.com.cn/Article/CJFDTOTAL-CHKD200702046.htm
[15]	陆关祥, 周鼎武, 王居里, 等, 2002. 复杂结构构造区遥感图像的地质信息提取方法——基于小波变换的多层次图像分割. 地球科学——中国地质大学学报, 27(1): 50-54. https://www.cnki.com.cn/Article/CJFDTOTAL-DQKX200201010.htm
[16]	莫登奎, 林辉, 2005. 一种稳健的高分辨率遥感影像分割方法. 见: 中国国家图形学会编. 第十二届全国图象图形学学术会议. 北京, 244-249.
[17]	任秀芳, 姬光荣, 姬光玉, 2007. 基于四叉树分割的分形图像编码改进方法. 微计算机信息, 23(10-3): 291-293. https://www.cnki.com.cn/Article/CJFDTOTAL-WJSJ200730117.htm
[18]	宋宇彬, 张秉权, 郝永平, 2004. 基于四叉树的图像分割技术. 兵工自动化, 23(6): 63-64, 69. https://www.cnki.com.cn/Article/CJFDTOTAL-BGZD20040600U.htm