Forest Resources Volume Estimate Of Pingnan Using Remote Sampling Technology

Remotely sensed sampling investigation is a method that set by the basic methods of general sampling investigation and uses means of remote sensing to get the area or volume of plots. In this paper, based on the theories of stratified sampling technology and remote sensing of forest stock volume, CBERS-02B remote sensing images of 19.5m resolution in Pingnan County were used as the basic data source, forest resource of Pingnan County was divided into 4 layers (eucalyptus, fir, pine and hardwood) according to the area statistics of major tree species and sampling principles. Combined with sample plots'investigation and data analysis for accurate positioning on the ground, as well as forest distribution map, area statistics of major tree species and other information, to build models of volumes of eucalyptus, fir, pine and hardwood and canopy density estimation by relevant mathematical methods (such as average residue, ridge regression, multiple linear regression etc.), then to estimate the volumes of the plots that laid by stratified sampling through solving the volume model and canopy density model simultaneously. Therefore, the total forest stock volume of Pingnan County can be obtained according to sampling. The main conclusions are as follows:(1) The optimal factors for volume estimation model are:altitude, canopy density, angle of slope, direction of slope, B1, B2, B3, B4, NDVI, RVI, (B4+B1)/B3; the optimal factors for canopy density estimation model are:altitude, angle of slope, direction of slope, B1, B2, B3, B4, NDVI, B4/B1.(2) The volume estimation models fitted through remotely sensed estimation theory of the four layers are: Yeucalyptus=-1289.168-0.055X1-0.485X2+59.354X3-4.588X4-0.026X5-6.404X6+9.651X7+859 .464X8-3354.937X9+7.155X10+274.129X11; Yfir=-3519.077+0.017X1+0.874X2+103.655X3+2.713X4+0.422X5-6.720X6+3.227X7+3904 .609X8-9046.742X9+4.930X10-190.668X11; Ypine=975.350-0.02X1i+0.310X2+75.522X3+1.063X4-0.138x5+1.010X6-0.396X7-973.590X8 +2525.813X9-0.021X10-52.064X11; Yhardwood=1775.453-0.02X1+0.047X2+52.832X3+7X4-0.151X5+20.132X6-23.250X7-1128.1 O8x8+6809.533X9+2.544X10-446.258X11. The canopy density estimate models are: x3(eucalyptus)=0.483+0.001X1-0.002X2+0.0005X4+0.0009X5+0.007X6-0.004X7+0.636X9+0.6 36X10-1.14X11-0.010X12; x3(fir)=6.757+0.0003X1+0.009X2+0.065X4-0.002X5+0.023X6-0.068X7+19.324X9+0. 069X10-3.841X11-O.024X12; x3(pine松)=5.980+0.0002X1+0.002X2-0.026X4-0.002X5+0.0001X6+0.031X7+6.778X9-0.024X10-1.907X11-2.033X12; x3(hardwood)=-2.560+0.0002X1+0.0001X2+0.010X4+0.0008X5+0.008X6-0.014X7-2.817 X9-0.027X10+1.807X11+0.945X12.In the models, X1,X2,X3, X4, X5, X6, X7, X8, X9, X10, X11, X12, in turn represent altitude, angle of slope, canopy density, B1, B2, B3, B4, RVI, NDVI, direction of slope, the value of B4+B1/B3, and the value of B4/B1.(3) The average precisions of each volume estimation model are 90.16% for eucalyptus,88.29% for fir,88.69% for pine and 81.70% for hardwood; the average precisions of each canopy density estimation model are 80.09% for eucalyptus,81.15% for fir,81.27% for pine and 79.32% for hardwood.(4) Through solving the volume model and canopy density model simultaneously, the volumes of sample plots that extracted by stratified sampling are obtained, then we can get the total forest stock volume of Pingnan County according to sampling principles, it is 6854634.265 cubic meters, the precision of the sampling is 90.12%. The result shows that the precision of estimation by remotely sensed sampling is 78.96%, compared with 5663204 cubic meters (from the forest management inventory)Paper was supported by forestry research & special public service sectors "stand structure and growth simulation studies (201104028)".