img

官方微信

高级检索

中国沙漠 ›› 2020, Vol. 40 ›› Issue (6): 169-179.DOI: 10.7522/j.issn.1000-694X.2020.00027

• • 上一篇    下一篇

基于无人机可见光影像的荒漠植被分布格局研究方法

郝梦宇(), 秦龙君, 毛鹏, 罗婕纯一, 赵文利, 邱国玉()   

  1. 北京大学 深圳研究生院 环境与能源学院,广东 深圳 518055
  • 收稿日期:2020-01-13 修回日期:2020-04-15 出版日期:2020-12-09 发布日期:2020-12-09
  • 通讯作者: 邱国玉
  • 作者简介:邱国玉(E-mail: qiugy@pkusz.edu.cn
    郝梦宇(1994—),女,北京人,硕士研究生,主要从事无人机遥感生态研究。E-mail: haomy@pku.edu.cn
  • 基金资助:
    科技部科技基础资源调查专项(2017FY100206-03);深圳市技术创新计划技术攻关项目(JCYJ20180504165440088)

Unmanned aerial vehicle UAV based methodology for spatial distribution pattern analysis of desert vegetation

Mengyu Hao(), Longjun Qin, Peng Mao, Jiechunyi Luo, Wenli Zhao, Guoyu Qiu()   

  1. School of Environment and Energy,Shenzhen Graduate School,Peking University,Shenzhen 518055,Guangdong,China
  • Received:2020-01-13 Revised:2020-04-15 Online:2020-12-09 Published:2020-12-09
  • Contact: Guoyu Qiu

摘要:

植被调查常用的地面样方法,需要大量人力和时间。同时,由于标准样方的尺度通常只有100 m2,许多植被信息可能无法被揭示出来。为此,本研究开发了一套基于无人机技术的荒漠植被样方调查方法,在此基础上进一步开发出一套植被空间分布格局分析方法,并以内蒙古巴彦淖尔荒漠区典型灌木群落为例,进行了实证研究。具体研究步骤是:首先,通过无人机低空航拍,获取研究区1.44×104 m2的可见光影像。其次,经图像处理,进行植被种类识别与分类。最后,利用得到的不同植被类型分布矢量图进行多尺度随机样方布设和植株位置提取,实现基于植被信息完全统计的样方分析和点格局分析。结果表明:分类过程精度高,样方尺度和数量比传统的地面调查扩大数十至百倍,植被信息数据也成倍增加,实现了流程自动化。据我们所知,这是第一个提出基于无人机进行荒漠植被分布格局方法的研究。本研究结果可应用于多种植被和生态学调查,由于可操作性强,该方法也可以推广应用于大范围的荒漠植被调查。

关键词: 无人机, 荒漠植被, 空间分布, 样方分析, 格局

Abstract:

Ground-based quadrat sampling is one of the commonly used vegetation survey methods, which is a time consuming and labor intensive process. The scale and amount of the quadrat are usually samll, and lots of vegetation information may not be fully revealed, too. To overcome these challenges, an UAV based methodology was proposed and was then applied to the spatial distribution pattern study in desert region: Taking the shrubs in the desert region in Bayannur as example, high spatial resolution RGB images were acquired by the UAV. After image preprocessing and object-oriented classification, the vegetation species were identified and then classified. The species distribution vector graphics were used for the multi-scale random quadrat layout and the vegetation location extracting. Then, the multi-scale quadrat analysis and the point pattern analysis based on complete statistics of vegetation information were carried out. The results showed that the classification processis of high accuracy. The quadrats data of vegetation information with enlarged scale and increased quantity could be processed automatically and then the multi-aspects ecological analysis could be conducted. Therefore, this method shows a great potential to be applied in a larger scale spatial distribution pattern analysis of plant population in desert reginon.

Key words: unmanned aerial vehicle, desert vegetation, spatial distribution, quadrat analysis, pattern

中图分类号: