| Monitoring vegetation fraction changes by remote sensing ,which can provide scientific basis for the regional ecological construction and sustainable development,is an important part for regional ecological monitoring.Vegetation coverage as a significant index for measuring land vegetation cover is a primary parameter for the ecosystem description and assessment.The research about it is benefit for monitoring the regional vegetation fraction and natural ecological environment conditions.The Songhua Lake basin which is located in the central south of Jilin Province,is the crucial hydro-junction area of our province.The basin has three lakes national nature reserves of Jilin Songhua River.And there are many species of wild plants and animals in there,many of which are national protected species and endangered species.So it is essential to protect the existing ecological environment.Therefore,the research of Vegetation coverage of SongHua lake has great realistic meaning.The purpose of this study is to calculat the vegetation coverage of SongHua lake by the remote-sensing methods.It mainly includes:Firstly,the paper applied some methods for monitoring vegetation coverage by remote sensing technique.Among these methods, dimidiate model with the characteristics of wide application scope and does not depend on the real-test data can well reflect the vegetation coverage of the research areas in the macro.Thus we can grasp vegetation changes'state quickly,accurately and conveniently.The study selected Dimidiate model which was established based on Normal Deference Vegetation Index (NDVI) to complete calculative process.Secondly , the paper introduced the research progress of vegetation index ,comprehensive analysed the reasons for selectted NDVI,then presented the principle of dimidiate pixel method in detail and combined with vegetation index to give the calculation model of vegetation fraction.Next,taking 2000 and 2006 year's Landsat TM/ETM+ as data source,we calculated the vegetation coverage.Durning the calculation process, because the research data were got from different places,at different times and different types of sensors, before NDVI's calculation,we converted image DN(Digital Number)into corresponding target spectral radiance value through reflectivity calculation.Then we got parameters that used to calculate vegetation coverage by probability statistics the results,after that took them into dimidiate pixel model to calculate.In order to make the image of the vegetation coverage easy to analyze and compare,and make it intuitively reflect vegetation-covered condition of vegetation distribution among different coverage during the period, this paper adopt the unsupervised classification in ERDAS to classify the final gray image of vegetation fraction.For its own needs,the classified image would be further classified into five categories and assigned different colors. According to the classification results,the number of pixels weights of the SongHua lake basin's vegetation coverage in each grade were statistics.Based on it the characteristics of the SongHua lake vegetation variation in 2000 ~ 2006 was analysised.Finally,it was on the basis of calculative process and results analysis that we drew the conclusion and gave some suggestions.Through calculation we knew within the whole valley,compared with 2000,the 2006 vegetation coverage showed an up trend.But when the vegetation coverage between0.8 and 1,we found it degenerated in the Huifa river area;when vegetation coverage between 0 and 0.6,vegetation coverage of 2006 is better than 2000.While when the vegetation coverage between 0.6 and 1,vegetation coverage condition of 2006 is worsen than 2000.In a word,in the study area, the vegetation fraction is degenerating.The study provided reliable scientific data for evaluating the vegetation coverage.Meanwhile,it provided scientific evidence for taking more pertinent measures to protect the vegetation within the basin. |
