Extraction Of Peat Bog Based On Multi-source Remote Sensing Images

Peat bogs are an important carbon sink in terrestrial ecosystems,an important component of wetland resources,and play an important role in maintaining the balance of nature and the global carbon cycle,its strong carbon sequestration capacity and huge carbon emissions on the impact of global climate change has become a hot spot for scholars at home and abroad.In order to promote carbon sequestration and greenhouse gas emission reduction in the world,the investigation of organic carbon pool in peatland in China has been carried out.The traditional survey method of peat bog can not meet the need of monitoring peat bog in time effect How to use remote sensing technology to acquire the spatial position of peat bog quickly and effectively,and master the spatial distribution of peat bog,it is of great theoretical value and practical significance to strengthen the research on the protection and management of peat bogs in China.Based on the data of Landsat-8 OLI_TIRS image,Sentinel-1A SLC radar image and DEM,combined with the ground sample and the investigation data of peat swamp carbon pool in the study area,the authors studied the typical distribution area of peat swamp in the Inner Mongolia of Yakeshi,through the selection of feature space and classification algorithm,the theory and method of extracting peat swamp information are studied in order to extract the spatial distribution of peat swamp accurately.The main contents and conclusions of this paper are as follows:(1)According to the formation mechanism of peat bog,the temporal phase,ratio,terrain,surface temperature and radar backscattering coefficient of peat bog were selected by remote sensing data.The results show that the peat swamp extraction based on multi-feature fusion data is superior to that based on single feature data,and the extraction of peat swamp is affected by the amount of feature data The spatial position of peat bog can be extracted accurately by analyzing and optimizing the characteristics of peat bog.(2)To select the best classification algorithm for the remote sensing extraction of peat bogs by using maximum likelihood,random forest algorithm,Support vector machine method and KNN to classify the land cover types in the study area.The results show that the accuracy of the four classification methods based on random forest is the highest,the overall classification accuracy is 87.32%,the user accuracy of peat bog is 80.54%,and the maximum likelihood classification accuracy is 81.87%,the user precision of peat bog is 67.35%,because the classification algorithm is simple and time-consuming,it is more suitable for extracting peat bog in large area,but it has lower precision The accuracy of the comprehensive evaluation of Support Vector Machine and KNN is over 75%,which meets the accuracy requirement of remote sensing extraction of wetland types and can also be used to extract peat swamp information.(3)Taking the Landsat-8 OLI_TIRS and Sentinel-IA SLC remote sensing images as the main remote sensing data sources and introducing the characteristics of land surface temperature,through comparing the effects of optical image,radar image and land surface temperature on the extraction of peat swamp,the characteristic factors and data sources suitable for peat swamp extraction were studied.The results show that the Landsat-8 image has abundant spectral information,easy data acquisition,fast processing speed and large image coverage area,but due to the abundant vegetation on the peatland surface,the single optical image is easily disturbed by the vegetation on the ground surface,which leads to low precision Sentinel-1A radar image and land surface temperature information can be used to monitor land surface humidity and temperature,which can effectively identify peat bog and improve classification accuracy.