湿地遥感研究进展

Application of Remote Sensing Technology to Wetland Research

湿地在维持生态平衡、保护生物多样性等方面发挥着重要作用。由于各种自然、人为因素的影响,目前越来越多的湿地转化为农业和城市用地,湿地的质量退化和数量减少引起人们广泛关注。遥感技术为湿地研究提供强有力的技术支撑。重点从遥感影像数据源、湿地分类、湿地调查3个方面论述遥感技术在湿地研究中的应用,同时指出目前湿地遥感研究中存在的不足,并展望其未来发展。

Wetland is one of the most important ecosystems, and it has high social benefit, economic benefit and scientific research value. However, wetland resources are taking on a heavy pressure because of various natural and anthropogenic factors. The degradation of the wetland quality and quantity has aroused widespread concerns. To conserve and manage wetland resources, it is important to monitor wetlands and their adjacent up- lands. Satellite remote sensing has several advantages, such as saving time and labor, multi-temporal, multi-platform, containing large amount of information and so on, for monitoring wetland resources, especially for large geographic areas. This review summarizes the literature on satellite remote sensing of wetlands, in- cluding the data source of remote sensing images used in wetland study, remote sensing classification methods of wetland, the survey of wetland, wetlands ecology and the survey of wetlands environment. Nowadays, Land- sat TM, Landsat MSS, and SPOT images are the major satellite images that have been used in wetlands re- search; Other images including NOAA AVHRR, IRS-1B LISS-II, MODIS images and radar images, and JERS-1, ERS-1 and RADARSAT images. Early work with satellite imagery used visual interpretation for clas- sification, which is still used widely today. The most commonly used computer classification methods are unsu- pervised classification and supervised classification. It is difficult to make great progress on improving the ac- curacy of remote sensing classification because of ＂the same things with different spectrums＂ and ＂different things with the same spectrums＂. However, the appearance of some new algorithms（decision tree, support vec- tormachine, BP neural network） as well as the use of ancillary data （soil data, elevation or topography data） im- prove the satellite remote sensing classification of wetlands to some extent. Also the integrated use of multiple classifications becomes a new trend, which can also increase the accuracy of the satellite remote sensing classi- fication to some extent. Remote sensing is also used widely in the surveying of wetland resources, estimating vegetation biomass of wetland, assessmenting wetland ecosystem health and so on, which can save time and la- bor greatly. At the same time, this paper points out five shortcomings existed in wetland research by remote sensing technology and prospects its future development from six aspects.