地球信息科学学报 ›› 2017, Vol. 19 ›› Issue (4): 549-558.doi: 10.3724/SP.J.1047.2017.00549

• 遥感科学与应用技术 • 上一篇    下一篇

基于NDVI阈值法的森林冰冻受灾范围精确提取

王学成1,2(), 杨飞1,*(), 高星1, 李丽3   

  1. 1. 中国科学院地理科学与资源研究所 资源与环境信息系统国家重点实验室,北京 100101
    2. 中国科学院大学,北京 100049
    3. 西安科技大学 测绘科学与技术学院,西安 710054
  • 收稿日期:2016-09-12 修回日期:2016-12-14 出版日期:2017-04-20 发布日期:2017-04-20
  • 作者简介:

    作者简介:王学成(1991-),男,山东青岛人,硕士生,主要从事基于RS和GIS的森林资源调查方面研究。E-mail:wangxc.15s@igsnrr.ac.cn

  • 基金资助:
    国家自然科学基金“冰雪冻灾干扰后亚热带森林生态系统恢复力的动态诊断”(41301607);资源与环境信息系统国家重点实验室青年人才培养基金项目“森林生态系统恢复力的遥感监测”;西藏生态专项课题“西藏生态环境大数据规范与制图展示”

Precise Extraction of Damaged Forest Range Caused by Ice-snow Frozen Disaster Based on the NDVI Threshold Method

WANG Xuecheng1,2(), YANG Fei1,*(), GAO Xing1, LI Li3   

  1. 1. Institute of Geographic Sciences and Natural Resources Research, CAS, State Key Laboratory of Resources and Environmental Information System, Beijing 100101, China
    2. University of Chinese Academy of Sciences, Beijing 100049, China
    3. College of Surveying Engineering, Xian University of Science and Technology, Xi’an 710054, China
  • Received:2016-09-12 Revised:2016-12-14 Online:2017-04-20 Published:2017-04-20
  • Contact: YANG Fei

摘要:

快速提取森林冰雪受灾范围,有利于准确掌握森林受灾情况,为此类灾害性气候事件防灾减灾、森林资源管理和生态保护提供科学依据。本文利用2001-2007年NDVI数据,提取灾前植被NDVI参考值和正常波动范围,结合2008年NDVI数据提取冰雪冻灾范围。该方法弥补了基于单一时相的传统方法(NDVI差值法)忽略植被指数正常波动的问题,分像元提取植被NDVI正常波动范围,使提取结果更加客观合理。与传统方法提取结果对比,省级尺度的验证结果相同(即森林受灾率均为34.72%,而实地调查森林受灾率35.3%),但2种方法在县市行政单元提取的森林受灾率相差较大。NDVI阈值法提取的森林冰雪受灾范围主要分布于湖南省南部地区,北部地区分布相对较少,而传统方法提取结果主要分布于湖南省北部地区,南部地区分布相对较少。根据实地考察资料显示,相比于传统方法,NDVI阈值法提取结果与实际森林冰雪冻灾空间分布信息更接近,精度更高,更适合于区域大尺度提取森林冰雪受灾范围。

关键词: 森林, 冰雪冻灾, NDVI阈值, 信息提取, 湖南省

Abstract:

Extraction of damaged forest range caused by ice-snow frozen disaster is good for knowing the relevant regional disaster information in time, and it can provide scientific support for disaster prevention and protection of forest resources and ecosystem. We extract pre-disaster plant NDVI reference value and the threshold of normal change with the time series data of 2001-2008. We attained the results of the spatial distribution of Hunan forest disturbed by ice-snow frozen disaster with NDVI data in 2008. The NDVI threshold method can make up for the defect of traditional method based on single-temporal images, which doesn’t take the normal change of vegetation index into consideration. The NDVI threshold method helps extracting different normal change threshold for each pixel, causing the results extracted by NDVI threshold method is more objective and reasonable. Contrasted with the results extracted by the traditional method, the percentage of damaged forest according to two methods have significant difference at county level, although the rates forest disasters with two methods are all 34.74% (the real rate of forest disaster is 35.3%) at provincial level. The forest snow disaster is mainly distributed in southern Hunan province and less in northern Hunan province using the NDVI threshold method, but the results using traditional method is contrary compared to the NDVI threshold method. According to field survey data, the spatial distribution of forest snow disaster using the method of NDVI threshold is more closed to real results compared to traditional method and its extracting accuracy is higher. Therefore, the NDVI threshold method is more suitable for extracting the spatial distribution information of forest snow disaster at large regional scale.

Key words: forests, snow disaster, NDVI threshold, information extraction, Hunan province