Forest Resource Change Monitoring Based On The Fusion Of Medium And High Spatial Resolution Remote Sensing Images

With the rapid development of remote sensing technology as well as the increasing demand for forest resources change monitoring accuracy,high spatial resolution remote sensing image has become the main data sources of forest resources change monitoring.However,high spatial resolution remote sensing image is limited by its sensor hardware as well as the problem of "Spatial-temporal Contradiction".In addition,the disadvantages(air,cloud pollution,etc.)has influence on Sensor imaging.It leads to the lack of high-quality and high spatial resolution remote sense image,which limits its application in the forestry field.Therefore,this paper will do research on spatial-temporal fusion technology of mono temporal remote sensing image,and using the existing medium spatial resolution remote sensing images and high spatial resolution remote sensing image(abbreviated as" Medium and high spatial resolution remote sensing images")to generate high quality fusion image with high spatial resolution in missing time.High spatial resolution fusion image is applied to forest resources change monitoring,making observation analysis and evaluation about spatial distribution and utilization of forest resources,timely mastering current situation of forest resources and dynamic change of forest covering and providing important basis for forestry planning and policy implementation.This paper takes the junction area between Jingde County and Jingxian County,Xuancheng City,Anhui Province as the research area.By using the spatial and temporal adaptive reflectance fusion model as well as the Spatiotemporal Satellite Image Fusion Through One-Pair Image Learning,this paper constructs spatial-temporal fusion framework of medium and high spatial resolution remote sensing image based on the date of mono temporal Landsat 8 OLI and GF-2 MSS image pair,which generate the high spatial resolution fusion image in October 9,2017.By using the method of quantitative and qualitative analysis,the image is conducted quality assessment.Finally,using the generated high spatial resolution fusion image and GF-2 MSS image on January 16,2017,the thematic map of forest resource cover change in the study area is obtained by post classification comparison method,which can realize the application of forest resource change monitoring.The main results and conclusions are as follows:(1)By using the STARFM algorithm to fuse Landsat 8 OLI and GF-2 MSS image,the research shows that it exists "light spot" phenomenon in the high spatial resolution fusion image generated by this algorithm.The larger the sliding window,the more "light spot " on the high spatial resolution fusion image,which has impact on discrimination of land class.The high spatial resolution fusion image generated by this algorithm has large difference from GF-2 MSS image.(2)By using the SSIF algorithm to fuse Landsat 8 OLI image and GF-2 MSS image,it is found that high spatial resolution fusion image generated by this algorithm has no"light spot" phenomenon.What's more,the quantitative evaluation index of each band is higher than that of STARFM algorithm,which means that SSIF algorithm can effectively reduce the transmission error between Landsat 8 OLI and GF-2 MSS image information.Compared with STARFM algorithm,data adaptability and image fidelity of SSIF algorithm is better.(3)This paper proposes the classification method based on combining adaptive threshold wavelet denoising with random forest,which can significantly improve the overall classification accuracy and Kappa coefficient of random forest classification method.It can also reduce salt and pepper noise in classified image.(4)In this paper,the forest resources change in the study area is monitored by the method of classification and comparison based on the combination of adaptive threshold wavelet denoising and random forest.According to the accuracy evaluation results,the overall accuracy of the forest resources change monitoring results in the study area is 80.2%,and the Kappa coefficient is 0.78.The research shows that compared with STARFM algorithm,SSIF algorithm can more effectively solve the " Spatial-temporal Contradiction" problem of GF-2 remote sensing image.The monitoring results of forest resources change can provide assistant decision-making service for forestry management department.