Vegetation Change Detection And Automatic Disturbance Identify Method Using Knowledge-based Temporal And Spatial Characteristics

The current remote sensing change detection method is based on the comparison of remote sensing image classification,the time continuity is relatively poor and lack of the change process information.Given this,in this paper,carrying out the research on the identification of vegetation change and disturbance in China based on 500 m 8-day composite Moderate Resolution Imaging Spectroradiometer(MODIS)and other data from 2001 to 2015,the main contents are as follows:1.The remote sensing index were evaluated from four basic covering types of vegetation,impervious surface,bare soil and water body,which showed that Optimized Soil Adjusted Vegetation Index,Biophysical Composition Index and Land Surface Water Index could express the information of vegetation,impervious surface,and water bodies better,but the relative research of bare soil index is less.In the frame of Brightness-Darkness-Greenness model,the bare soil index is constructed according to the characteristics of high brightness and lower darkness.The bare soil index,which is constructed,can extract the bare soil information without the water mask,and also expressed the bare degree of land surface.2.The knowledge-based temporal and spatial characteristics of vegetation cover change are established.The change of vegetation cover is divided into three levels:change/unchange,change to non-vegetation/vegetation and specific change type depending on whether or not the change is occurring.Secondly,the characteristics of vegetation change were analyzed from the years of remote sensing indices of Optimized Soil Adjusted Vegetation Index,Biophysical Composition Index,Land Surface Water Index and Bare Soil Index,so thata feature table corresponding to the three levels of vegetation change was created.The knowledge-based temporal and spatial characteristics of vegetation change was constructed at multiple dimensions and multiple levels,and the classification of vegetation change can be achived without setting threshold value,which has good generality.3.The method of automatic identification of vegetation change and disturbance was established.Based on JM distance,the temporal similarity of vegetation indices was calculated and the vegetation change was indicated.Using the Logistic function to fit the temporal similarity curves,the vegetation change process is divided into mutation type,gradient type,disturbance type,mixing type and stable type.Determining the change area by using the change of temporal similarity,based on the knowledge-based temporal and spatial characteristics of vegetation cover change and vegetation change process,the antomatic identification method of vegetation cover type change,growth change and recovered after disturbance is constructed.The change detection accuracy of vegetation cover type based on reference point is 86.82%,Kappa coefficient of 0.8396,and the change time detected is more consistent with the actual change time.The method can detect the vegetation change type,at the same time describe the vegetaiton change process information,and realize the continuous monitoring of vegetation change.4.Based on the automatic identification method of vegetation change and disturbance,the vegetation change and disturbance area in China was acquired from 2001?2015.The results revealed that the change of vegetation in China was dominated by growth change,followed by vegetation cover types change,and again for disturbance recovery.More than half of the changing regions belong to mutation.The vegetation growth change area was 1,164,535 km2,in which was grow better mainly,and was mainly distributed the warm temperate deciduous broad-leaved forest zone and the temperate Southern forest steppe region.The area of vegetation cover type change was 417,355 km2,more than Two-thirds of them are afforestation,mainly distributed in the Loess Plateau and Northwest Regions,followed by is urbanization,covering the type of vegation change 10.34%,showing a feature that from Beijing,Tianjin and Hebei-Shandong Peninsula-Yangtze River delta-Hercynian Economic zone along the coastal zone and the provincial capital cities as the center of the expansion.The area of 19,035 km2 was recovered after disturbance and concentrated in the Greater Khingan Mountains and the arid/semi-arid regions of northwest.