Abstract:

Recurring to the SPSS and the model building and spatial analysis tools of ERDAS IMAGINE9.0 & ArcGIS9.2, this paper adopted TM images and 1∶100 000 relief map as the data source, extracted RS factors such as pixel value of six wave bands and theirs linear & non-linear combination from TM images using the buffer area of GPS points，and extracted GIS factors such as elevation, slope and aspect from relief map. RS factors & GIS factors were set as independent variables and field investigated forest volume data were set as dependent variable to build the multi-regression model. Sample data were selected by standard deviation law and factor variables were selected by the method of principal components analysis, stepwise and enter regression analysis. The forecast precision of the model is 87.35%. The grey images of each variable in the model were masked by the forest land extracted from the TM images of Shandong Province and got the masking image of each variable. The paper calculated the total forest volume of Shandong Province by the complicated operation of all the masking images using the multi-regression model. The result shows that the total forest volume of Shandong Province is 62 035 300 m³. The study provided an effective way to estimate the forest volume of large area.

Key words: 3S technology, Shandong Province, forest volume model, estimation