Screening Of Vegetation Indexes And Studies On Remote Sensing Monitoring Models For Species Multiplicity

Grassland vegetation is the most important renewable resource and raw material base for animal husbandry; it is also the natural green barrier of the earth, playing an important role in human's survival and development. While in recent years, most of pastures were over-grazed. In order to reasonably utilize the grassland resources, Qinghai Province Sanjiaocheng Sheep Breeding Farm had taken a serious of measures to guarantee the sustainable development of the grassland. Moreover, using remote sensing technique to evaluate species multiplicity and build evaluation model is a new researching trend in recent years. Compared with traditional method, it has many advantages. Meanwhile, it offers new ways for managers to better manage grassland resources.The main achievements of this research include:1) Local extant historical materials were collected and reorganized; local topographic maps and some pasture pictures and documents were processed with vector map generating. Local grassland resources were investigated for a better understanding of the present pasture condition and its utilization situation.2) After preliminary treatment, geometrical correction and radiation adjustment of remote sensing images, those with better effects were referred for more detailed investigations of specified monitoring region. Samples were chosen, plots were set, for the measurements of various ecological indexes.3) According to spectrum characteristics of various grassland vegetations, vegetation indexes were settled which are often applied in grassland monitoring nowadays. Through selecting vegetation indexes with the ratio of vegetation signals and soil noise to decide which can be better utilized in this region, we found that revised soil-vegetation indexes can only be applied in Achnatherum splendens steppes for better vegetation information extraction, in other words, soil noise disturbance was better eliminated in this way; extraction effect of vertical vegetation indexes is best as for other three grassland types.4) Elevation model was established with contour maps whose data has been converted into vector data using ERDAS IMAGE software. Afterwards, terrain analysis was carried on with this DEM model.5) Correlation relationships between remote sensing vegetation index and species multiplicity were analyzed, the results showed the best correlativity is between PVI and...