The Change Analysis Of Forest Resource From Landsat Images In Renshou County

Forest resource is a general term of forest land and all the forest organism grown on the land.Forest is the main resource about forest resource,and it also includes understory plants,microorganisms and other resources.The forest area and the forest volume are the main indicators to reflect the amount of forest resource.Both them are basic indicators that reflect the whole forest size and the level of forest resources about a country or a region.They are also the important basis to reflect the abundance about the forest resources and measure the pros and cons of ecological environment.Forest resource has a better reflection in forestry production.So it's important to know the changes about forest volume in time to monitor the forest resources.Renshou county is chosen as study area in this article.Combined the Forest resource inventory data in 2014 to preprocessing images in 1992,2002 and 2014 like geometric correction,atmospheric correction,coordinate transformation and cropped to three different periods of Landsat images.Then using the maximum likelihood method to classify images.Extracting RS factors from these bands.Filter arguments by correlation analysis.Use multiple linear regression analysis to establish volume estimation model in study area.And use this model obtaining the distribution map of forest resources and trends.Combined test samples from ground truth data to calculate the volume model accuracy.The main conclusions are as follows:(1)The area change:Overall the 22 years,the land type changed obviously in Renshou County.The Non-forest area decreased 16.03%,the forest land area increased 22.38%,and the arbor forest land area increased 46.30%.(2)Model building and accuracy assessment:Use these 19 factors from Landsat image in 2014.In these 19 factors,vegetation index(NDVI,RVI,DVI),band Near infrared and band ratio B5/B3 and(B5+B2)/B4 are significantly correlated to forest volume.Other factors are not.Chose 6 factors to establish linear regression model in 2014.Used the remained 30 samples to test the model accuracy and get the overall accuracy is 92.06%.The model formula is:Y=50.161*NDVI+23.979*RVI+0.001*DVI-0.001*B5+0.996*B5/B3?1.018*(B5+B2)/B4-29.265;(3)Forest volume estimation result and spatial distribution:Estimated the forest volume of three images,the estimation results are 606485m3,1275341m3 and 2109673m3.Using the forest resource distribution map of images to extrapolate about three periods.We know that the main forest resource of the image in year 1992 is only located Longquan Mountains area near the Heilongtan reservoir.The overall forest distribution get together.The other regions have the less amount forest volume.It described that the arbor forest distributed unevenly in Renshou County in 1992.In 2002,forest volume gradually increases in Longquan Mountain in Renshou County and the forest volume grown rapidly in the southeast regions,the Wangyang Town.The other regions increased slowly.In 2014,the forest volume is growing rapidly in Renshou County.