Individual-Based Remote Sensing of Canopy Trees

Pioneering work over the past decade has demonstrated the potential to map and monitor individual canopy trees using a wide range of sensor data, including digital airborne images, high-density terrestrial laser scanning, drone lidar, airborne imaging spectroscopy and multispectral satellite observations. Studies on Barro Colorado Island have identified canopy individuals of 11 species of tropical trees in 7 families. Improvements in spatial resolution, spectral resolution, spectral range, temporal frequency, and radiometric accuracy will increase the number of species that can be mapped and monitored using remotely sensed data. Over the next decade, whole-tree segmentation algorithms using very-high-density point clouds from lidar and digital photogrammetry will augment ground-based tree inventories, expanding the size of field-inventory plots and the types of measurements that are collected. Constellations of small satellites will enable analysis of tree populations throughout entire species ranges. Critical to all of these efforts is the acquisition of high-quality and spatially representative training data, as well as investment in automated algorithms with clearly defined sources of uncertainty.