Author_Institution :
Sch. of Geogr. & Remote Sensing, Beijing Normal Univ., Beijing, China
Abstract :
Crofton weed, which was introduced to the southwest of China in the 1940s via the border between China and Burma, now has a seriously damaging effect on the biodiversity and the living environment in those regions, including Yunnan, Sichuan, Guizhou, Guangxi, Tibet, Chongqing, and Hubei. To help monitor this situation effectively, there have been many researches that utilize remote sensing images to identify the spread of Crofton weed. However, most of these researches found it difficult to identify Crofton weed because it is mixed and entangled with a large number of other vegetation. In our research, data processing includes three steps: 1) Retrieving the spectral feature values of Crofton weed at the different growth stages through long time regional sampling; 2) Comparing the NDVI of Crofton weeds, forests, crops, water bodies, residential areas in the typical sampling regions by using the time series data of remote sensing images; and finally, 3) Classifying the remote sensing images of the research regions according to the distinctive features of NDVI curves based on the time series data. The outcome of following these three data processing steps is a relatively accurate classification result, which produces a high-accuracy distribution map of the Crofton weed.
Keywords :
crops; ecology; geophysical image processing; image classification; remote sensing; time series; Crofton weeds; MODIS NDVI; NDVI curves; biodiversity; crops; damaging effect; data processing; forests; high-accuracy distribution map; living environment; long time series; regional sampling; remote sensing image classification; remote sensing images; residential areas; sampling regions; spatial distribution; spectral feature value; time series data; vegetation; water body; Agriculture; Cities and towns; Monitoring; Remote sensing; Roads; Time series analysis; Vegetation mapping; Crofton weeds; MODIS; NDVI; Remote sensing; Spatial distribution; Time series;