Real-time LiDAR Processing for Tree Metrics
So much geospatial data is now being collected worldwide that the main challenges service providers now face is insufficient storage space and processing power. Due to limited capacities on the upstream and survey sector, it is the data users and solution-focused providers that are forced to tackle the problem. Half of a data analyst?s time is now used for data download, data harmonisation and data pre-processing, resulting in little time for data interpretation and turning large volumes of raw data into digestible information and knowledge (Giuliani, Dubois, and Lacroix 2013). By processing in real time, on board mobile, aerial or space-borne platforms, we are able to (a) speed up the delivery of services, (b) reduce the bandwidth requirements for data download from platform, and (c) significantly open access to EO data by offering direct downloads of analysis ready data products without the need for end-user clients to obtain third party software and analysts. Carbomap has been prototyping a hardware module, instant-Omega (i-?) that enables on-board real-time processing of laser scanning data with API access for open innovation of applications. In this paper we present results where the i-? has been used to process real-time tree metrics of urban trees. We will describe how the i-? combines data from a Riegl VUX and Applanix IMU to formulate geo-coded point clouds, and then identifies individual trees in the data, before extracting tree height and diameter at breast height (dbh, a key metric of tree size). We will present results and accuracy estimation, and describe some of the challenges of differentiating trees from street furniture. Furthermore, we will explain the API interface and describe some of the other possible applications, such as tree and forest metrics from VUX systems flown on UAVs.