欢迎访问林业科学,今天是

林业科学 ›› 2015, Vol. 51 ›› Issue (11): 76-82.doi: 10.11707/j.1001-7488.20151110

• 论文与研究报告 • 上一篇    下一篇

基于人工火场的EOS/MODIS林火辐射增量研究

罗永明1, 孙涵1,2, 刘诚2, 戎志国2, 钟仕全1, 何立1, 陈燕丽1, 黄永璘1   

  1. 1. 国家卫星气象中心遥感应用试验基地 广西气象减灾研究所 南宁 530022;
    2. 国家卫星气象中心 北京 100081
  • 收稿日期:2014-10-24 修回日期:2015-07-20 出版日期:2015-11-25 发布日期:2015-12-08
  • 通讯作者: 钟仕全
  • 基金资助:
    国家863项目 "多源卫星遥感高时空分辨率森林火情监测预测评估系统研究" (2006AA12Z125);广西青年基金项目 "基于RS与GIS的广西森林火险等级预报关键技术研究"(2014GXNSFBA118224)。

Analysis on Radiation Increment of the Artificial Fire Based on EOS/MODIS

Luo Yongming1, Sun Han1,2, Liu Cheng2, Rong Zhiguo2, Zhong Shiquan1, He Li1, Chen Yanli1, Huang Yonglin1   

  1. 1. Remote Sensing Application and Validation Base of National Satellite Meteorological Center Guangxi Meteorological Disaster Mitigation Institute Nanning 530022;
    2. National Satellite Meteorological Center Beijing 100081
  • Received:2014-10-24 Revised:2015-07-20 Online:2015-11-25 Published:2015-12-08

摘要: [目的]林火引起的辐射增量是林火遥感监测技术中非常重要的物理量,开展人工火场地面与卫星同步观测试验,计算人工火场在EOS/MODIS第20,21,22,31和32波段上产生的辐射增量,为林火监测与研究提供参考。[方法]通过地面人工火场观测试验与卫星同步获取火场辐射参数,并收集同步的EOS/MODIS卫星数据。以30 m分辨率TM数据提取地表背景数据,将火场周围分为植被、裸地和水体3个覆盖类型,并将其作为1 km分辨率EOS/MODIS像元的端元。根据线性混合理论,构建无火场影响情况下EOS/MODIS第20,21,22,31和32波段的辐射估算模型,实现人工火场影响上述波段辐射增量的计算,获得辐射增量值。[结果] 1)根据线性混合理论建立的EOS/MODIS像元地物辐亮度估算模型,其样本相关系数均大于0.99; 2) EOS/MODIS第20,21和22波段对地表热辐射反应非常明感,经模型估算,亮温增量在第20波段为9.35 K,第21波段为7.49 K,第22波段为8.00 K;第31和32波段对高温热源不敏感,增量分别仅为1.13和0.83 K。[结论] 1)卫星能观测到地面 ≥200 m2(即火场面积与像元面积比为1/5000)的火场。研究结果对于误码识别和深入研究林火识别与面积估算模型具有重要意义; 2)卫星观测存在目标物的像元分割现象,本次试验中主火场在13号像元,但14号像元第20波段亮温增量也达4.08 K,表明火场很可能落在2个像元之间,14号像元也受火场影响。这种现象往往造成火点多判,高估火场面积,应在研究和应用中引起重视; 3)本文研究方法和结论适用于国产FY-3等中分辨率遥感数据。

关键词: 辐射增量, 人工火场, MODIS

Abstract: [Objective] Radiation increment caused by forest fire is a vital parameter in using the remote sensing technique. In this paper, to provide reference for forest fire monitoring and research, the ground and artificial satelite syncharonous observation experiment are carried out, and then, radiation increments of MODIS bands at No.20,21,22,31 and 32 are calculated. [Method] Radiation parameters were acquired through the ground observation experiment of artificial fire scene and satellite synchronous observation. First, surface background data were extracted using TM data of 30 meters resolution. Second, the land around the artificial fire was divided into three categories of vegetation, bare and water which were taken as pixel endmember of MODIS data of 1 kilometers resolution. Finally, according to the linear mixed theory, radiation estimation models of MODIS channel No.20,21,22,31 and 32 were constructed without fire scene influence, then the above channel's radiation increments of artificial fire scene were calculated. [Result] 1) Correlation coefficients of MODIS radiance estimation model samples are all over 0.99; 2) MODIS radiance estimation model is used to calculate bright temperature increment. It shows that MODIS bands No.20, 21 and 22 are highly sensitive to surface thermal radiation with increment of 9.35, 7.49 and 8.00 K respectively while bands No. 31 and 32 are insensitive to high-temperature heat source with increment of 1.13 and 0.83 K respectively.[Conclusion]1) Fire scene area of 200 m2 (ratio of fire scene area to pixel size is 1:5000) can be detected, it is of great significance for error identification, fire detection and area estimation in further study; 2) Pixel segmentation phenomenon exists in satellite observation and these phenomenon could easily lead to over-calculated fire points and overestimate its area. In our experiment, artificial fire is mainly within No.13 pixels but No.14 pixel is also influenced by the fire with its radiation increment of 4.08 K. Thus, the artificial fire may be within No.13 and 14 pixel simultaneously. 3)Methods and conclusions of this paper are suitable for mid-resolution remote sensing data, such as FY-3.

Key words: radiation increment, artificial fire, MODIS

中图分类号: