Canopy Cover Estimation Based on LiDAR and Landsat 8 Data using Support Vector Regression

Indonesia has large areas of forest that spread in almost every island of Indonesia. Forest in Indonesia also have various types of ecosystem, therefore they have an important role to protect each element contained within the ecosystem. The forest monitoring system in Indonesia still using the traditional approach for monitoring forests areas. This paper aims to generate a prediction model using remote sensing data and support vector regression for the model to estimate forest cover, especially in Indonesia. Landsat 8 OLI reflectance value from each band was used to estimate forest canopy cover with the integration of LiDAR data. The prediction model of forest canopy cover was observed at R-2 = 0.6921 and RMSE = 0.1658 of canopy cover. In this case R-2 means the correlation between LiDAR point cloud with Landsat bands. The SVR kernel used in this study was radial basis function with parameter (Cost: 10, Gamma: 1 and Epsilon: 0.1).