Automated forest inventory using the iPad Pro LiDAR scanner

Forest inventory is essential for forest management. Manual methods based on tools such as measuring tapes and calipers are still popular because of the high cost of most automatic methods. In 2020, Apple released the new iPad Pro equipped with a LiDAR sensor. Different studies have validated the iPad LiDAR as a viable low-cost alternative, but no developed application has been released yet. In this master thesis, we created a mobile application, a server, and a workflow to automatically process iPad LiDAR data. The application can extract the DBH (Diameter at Breast Height), the position of the trees, the surface of the scanned area, keep a history of the evolution of the trees, and compute derived forest attributes. The solution has been evaluated based on data collected in the Bois de Lauzelle forest in Louvain-la-Neuve. The workflow includes the Hough transformation to detect trees, RANSAC to extract DBHs, and the Hungarian algorithm to match the rescanned stands. Our evaluation shows that the application achieved a detection rate of 100%, an RMSE of 3.72 cm, and a halved survey time with respect to the manual method. These results confirm the interest of taking advantage of the iPad-LiDAR for forest inventory and demonstrate the efficacy of the various algorithms used in the workflow.