May 26, 2018
GeoHashTree library for storing and accessing multi-dimensional point clouds
A file format and library for storing and rapidly accessing point cloud data, in particular LIDAR data.
GeoHashTree organizes points into a tree structure for fast spatial access. The tree structure itself encodes the significant bits of at each node, so child nod es can omit them. The result is a smaller file than if all the points were store d with full precision. Each node includes statistical information about the chil dren below e.g. average/median Z value permitting fast overview generation. Ad ditional attributes are attached to the tree at parent nodes, below which all ch ildren share the attribute value. This reduces duplicate data storage further.
The advantage of a GeoHashTree file over a LAS file is fast access and filtering , since the tree encodes useful information at each node to speed searches over the full set of points in the file. LASZ zipped files can be smaller, but will b e less efficient at overviews, searching and sub-setting. GHT is a good working format for applications that will be filtering and querying large sets of LIDAR data.