Advanced techniques for hyperspectral/LiDAR sensor data integration

点击翻译按钮获取中文摘要

Integrated hyperspectral and LiDAR data will be an essential tool for a wide variety of environmental and transportation engineering applications. LiDAR data provide primarily the 3D geometry of the scanned environment while hyperspectral sensors deliver data on the radiometric aspects of the surveyed objects. Multi-sensor integration has been one of the active research topics for several years. However, many challenges have not been solved yet; particularly, when data collected from LiDAR and hyperspectral sensors are integrated. This is due to the differences of the sensor data collection techniques and how the data collected is processed. This proposed research program focuses on the development of innovative techniques and technologies for multi-sensor integration and data processing. The program main objectives are: a) Develop low-cost light-weight 3D-hyperspectral mapping system with platform-switching capability for indoor and outdoor mapping; b) Develop new approaches and algorithms for indoor 3D modelling and mapping; and c) Develop genetic algorithms for point cloud classification and feature extraction. The developed all-in-one 3D-hyperspectral mapping system will allow users to switch collection modes (i.e. from airborne, terrestrial, and mobile platforms) with a "plug-and-go" concept. An innovative software will be developed to automatically integrate, for the first time, data captured by LiDAR, and hyperspectral sensors. A new approach will be developed for indoor mapping and 3D modelling without any dependence on indoor infrastructure (e.g. WiFi signal) or GNSS signals. New algorithms will be further developed to further improve the automation of information extraction based on the data collected by the multi-sensor systems. The techniques developed will serve in variety of applications in emerging environment and transportation Monitoring and Modelling (ETMM). Primarily focus will be given to smart city 3D modelling, noise mapping, indoor mapping for mining, and road quality detection. The proposed research program is the first of its kind at inventing a light-weight 3D-hyperspectral mapping system with platform switching capability associated with the corresponding data processing algorithms for knowledge discovery. The proposed research work will serve as a core for a number of fundamental research studies yet to come to facilitate the use of LiDAR and hyperspectral data processing for engineering applications. The research will potentially support technology transfer and commercialisation to the industry and consolidates the training of HQP.