The majority of existing methods for delineating trees from LIDAR point cloud use a region growing approach. Seed points representing the highest point of the trees (tree-peaks) are detected in the point cloud. The remaining points are iteratively assigned to one of the seed points, thus growing the region representing trees. The tree-peak detection methodologies are based on local geometry analysis, identifying locally highest points within some appropriately sized neighborhood as tree-peaks. Such approaches result in false positives over a tree with large crown and false negatives with multiple trees in close proximity. Hence, we advocate non-local approach to tree-peak detection which analyzes mutual relation between points in the point cloud. Initially, we assume that each point in the point cloud could possibly be a true peak. Then a series of filters applied on the point cloud removes false peaks, eventually, producing superior results over the state of the art.
Copyright is held by the author.
The author granted permission for the file to be printed and for the text to be copied and pasted.
Supervisor or Senior Supervisor
Thesis advisor: Zhang, Hao
Member of collection