Canopy Density Estimation Model Of Three Different Data Sources Of Southwest Sichuan Evergreen Broad-leaved Forest

Canopy density refers to the ratio of the area of projection of the trees canopy in the forest to the ground in direct sunlight and the total area of the forest.It is the reflection of stand density indicator.It is not only an important survey factor in the investigation of the forest resources,and can reflect the forest environment and structure.In forest management,canopy density is the important index of small class divisions,determining the intensity of tending felling,and it is also important factor of the forest stock volume estimation by using the remote sensing images. Located in famous strong rainfall area of Sichuan,Southwest Sichuan Natural Evergreen Broad-leaved Forest's rainfall is abundant,which is of great significance to keep regional soil and water conservation,at the same time,the region is also an important part of the upper and middle Yangtze river's ecological barrier.Compared with the evenaged forest,the stand structure and composition of the natural evergreen broad-leaved forest is more complex,and the forest ecological system of which is more stable,whose self-adjustment ability and ability to resist disturbance is stronger,and the ecological benefit of regional role is more prominent.But over the years,the scholars'study of Southwest Sichuan Natural Evergreen Broad-leaved Forest is less,and more focuses on the study of the related surface vegetation parameters in a single scale,the multi-scale on the related research is less.In this research,the evergreen broad-leaved forest in Shang Li town,Yu Cheng district was studied.Based on the remote sensing factors extracted from Pleiades-1,SPOT-5 and Landsat-8 images data source,combined with the measured ground canopy density obtained from the hemispherical photography method,the study established quantitative estimating model of canopy density in southwest Sichuan evergreen broad-leaved forest under 2m,15m and 30m scales by partial least squares regression analysis,and the obtained different estimating model were comparative analyzed.(1)Based on Canopy density estimation Pleiades-1 image,integrated relevant factors,selected normalized difference vegetation index (NDVI),the ratio vegetation index (RVI),vertical vegetation index (PVI),elevation,and the Shadow abundance.Canopy density estimation model established by the least squares method,and fitting 2M scale Canopy density maps of the study area.Model: Canopy density=0.54121*RVI+0.544514*NDVI+0.479604PVI+0.318437ELEVATION-0.242535*SF+ 6.86777,the accuracy was 85.8%.(2)Estimate based on Canopy density SPOT-5 image,integrated relevant factors,selected the NDVI,RVI,PVI,elevation and slope.Canopy density estimation model established by the least squares method,and fitting 10M-scale study area Canopy density distribution.Model: Canopy density=0.157989*NDVI+0.136496*RVI+0.153188PVI+0.318437ELEVATION+0.128691*SL POE+8.02016,the accuracy was 81.8%.(3)Based on Canopy density estimation Landsat-8 images,select the NDVI,RVI,PVI,variance,el evation and shadow abundance (SF),established by the least squares method to estimate Canopy density model and intends to get together 30M scale Canopy density maps of the study area.M odel: Canopy density=0.129587*RVI+0.17575*PVI+0.153112NDVI+0.091987ELEVATION+0.06703*VAR +0.061918*SL-0.075168SF+7.34522,the accuracy was 75.8%.(4)Its spatial resolution is high,which may reduce the existence of mixed pixels,and improve the classification accuracy.And the higher spectral resolution,the more accurate surface information reflected by single band will be,and the richer information reflected by the image will be.This paper did a lot of detailed analysis and comparison of three kinds of images in the canopy density inversion problem in the respects of spatial esolution,spectral resolution,texture information,vegetation index and so on.For the canopy density estimation of the Sichuan southwest of vergreen broad-leaved forest,it is thinked that the higher spatial resolution image data,the higher the inversion accuracy will be,which is to say in terms of the accuracy of the canopy density inversion,the Pleiades-1 images have more advantages.(5)By comparing the 3 kinds of model of canopy density distribution map,it has been found that canopy density distribution was the more broken which is based on Pleiades-1 image,but canopy density maps which is based on Landsat-8 images obtained relatively continuous,canopy density distribution which is based on SPOT-5 images between the two,that has a direct relationship with the retrieval of canopy density scale.