Detall
Conferència "Virtual Reality Geological Studio (VRGS): An integrated approach to digital outcrop modelling. "
A càrrec de avid HODGETTS (School of Earth, Atmospheric and Environmental Sciences, University of Manchester, Regne Unit)
Organitzat conjuntament CSIC-UB
Data: 04/11/2010
Hora: 12:00 h
Lloc: Sala d'actes de l'Institut Jaume Almera
Descripció:
Digital outcrop data collection is changing the way geologists do fieldwork. The advent of cheap GPS systems and more expensive techniques such as LIDAR allow us to collect spatially referenced data rapidly and accurately. These new data collection techniques in no way replace traditional fieldwork approaches, but used in conjunction with them are incredibly powerful. LiDAR data provides and convenient, rapid and accurate method of mapping the topography of an outcrop at a very high resolution (>1.0cm) and at a high data density (scanners collect data at rates of 10's to 100's thousands of points per second). Integrating the LiDAR data with digital images allows photo-realistic models of the outcrop to be generated which may then be used in the laboratory for analysis, or taken back into the field and used as very high resolution 3D base-models for further outcrop data collection. Though the basic processing of LiDAR data is straight forward, detail analysis of the data
In-house software called VRGS (Virtual Reality Geological Studio) has been developed at the University of Manchester to aid in the interpretation and analysis of LiDAR and other digital data, and allow traditional field data collection methodologies and data types to be integrated and digitised into with the digital dataset. This integrated dataset then facilitates a whole new range of approaches which help analyse and interpret the data. Both manual and automated approaches to the mapping of geological features within laser scan data will be presented, using a variety of surface attributes such as surface orientation, roughness and curvature. Given the resolution and extent of LiDAR datasets manual interpretation can be very time-consuming, so a variety of geometric and statistical techniques are being developed in order to help reduce interpretation time, these include the use of Artificial Neural Networks (ANN) to classify the datasets into categories of similar orientation and at