欢迎访问林业科学,今天是

林业科学 ›› 2005, Vol. 41 ›› Issue (2): 16-20.doi: 10.11707/j.1001-7488.20050203

• 论文及研究报告 • 上一篇    下一篇

基于图像理解的树木直径抽取技术

王雪峰 张超 唐守正   

  1. 中国林业科学研究院资源信息研究所,北京100091
  • 收稿日期:2003-02-09 修回日期:1900-01-01 出版日期:2005-03-25 发布日期:2005-03-25

A Technology of Gathering Forest Diameter Based on Image Understanding

Wang Xuefeng,Zhang Chao,Tang Shouzheng   

  1. Research Institute of Forestry Resource Information Technique, CAF Beijing100091
  • Received:2003-02-09 Revised:1900-01-01 Online:2005-03-25 Published:2005-03-25

摘要:

介绍由摄像机图像抽取林木直径的方法、步骤,将度量误差模型算法应用到参数的求解当中,提出适合于林业野外作业的内、外参数分离策略,最后对文中算法、策略进行实际验证。结果表明:1)从图像中重构林木直径是可行的;2 )由度量误差模型算法求解参数是一种很好的算法;3)将摄像机、内外参数分离,不仅减轻了繁重的林业野外作业,同时保证了三维重建精度;4 )单独考虑摄像机焦距,增加了实际操作上的灵活性

关键词: 林木调查, 直径量测, 立体视觉, 图像理解, 三维重建

Abstract:

The approach that gathers forestry diameter by camera image is introduced in detail and the error-in-measure models algorithm is successfully applied to the parameters resolving. While the inner parameters and outer parameters decomposing maneuver that is fit to forestry fieldwork is provided. The algorithm and maneuver in this paper are illustrated finally. Results prove:1) The way that reconstructing forestry diameter from camera images is available; 2) It's a good algorithm that parameters are resolved by error-in-measure models algorithm; 3) Decomposing camera inner parameters and outer parameters not only decreases the heavy forestry field work but also ensures the 3D reconstructing precision; 4) Considering the camera focus length independently, the flexibility of practical operation is improved.

Key words: forest inventory, forest diameter measuring, stereo vision, image understanding, 3D-reconstruction