Object-based image analysis for forest species classification using Worldview-2 satellite imagery and airborne LiDAR data

摘要

It has been shown that new remote sensing technologies have the potential to complement deficiencies of conventional methods such as aerial photograph interpretation and field sampling as well as improve the accuracy, reduce costs, and increase the number of applications within various forest environments. Newly available high resolution spatial data such as small footprint, multiple-return, discrete airborne LiDAR data and WorldView-2 satellite imagery offer excellent opportunities to develop new and efficient ways of solving conventional problems in forestry. However, the development of a comprehensive procedure for deployment of these new remote sensing data to create forest mapping products that are comparable and/or superior in accuracy to conventional photo-interpreted maps poses big challenges. Proper use of high spatial resolution data with object-based image analysis approach and nonparametric classification method such as decision trees may offer an alternative to aerial photograph interpretation in support of forest classification and mapping. This study presented ways of processing airborne LiDAR data and WorldView-2 satellite imagery for object-based forest species classification using decision trees in the Strzelecki Ranges, one of the four major Victorian areas of cool temperate rainforest in Australia. The results showed the contribution of four new WorldView-2 image bands to forest classifications, and demonstrated that the integration of airborne LiDAR and eight WorldView-2 bands significantly improved the classification accuracy.