| Summary: | In this study the limits and potentials of the principle of photoclinometry for glacier elevation modelling is assessed. A combination of optical and laser remote sensing instruments are used to map the surface of different remote ice bodies. Loss and gain of a glaciers' mass is a good indicator for local climate. If mass-balances of glaciers are known on a world-wide scale, they can give a better understanding of global climate change. Due to the remoteness and harsh environments at which glaciers are located a mass-balance is difficult to maintain or even start-up. Therefore, the use of optical and laser remote sensing can be of great value. The focus of this thesis is to identify options in which optical and laser remote sensing instruments can reliably map the surface of a glacier. It will use the technique of photoclinometry, also known as shape-from-shading, onto optical acquisitions. Estimated surface and topographic information will be integrated with spaceborne laser altimetry. These methodologies will compose a workflow, which produces elevation estimates of a glaciers' surface. Snow has a homogeneous reflection pattern, therefore the perceived reflection of a surface is primarily dependent on the amount of received sunlight per unit area. In that sense brightness values, perceived by an optical spaceborne sensor, may directly be related to surface orientation. This relation is used in this study to estimate a surface model from an optical image. Estimation of this model is done through the relations of a Lambertian reflection function and inverse theory. Only small patches of imagery can be processed, as its estimation is limited in size due to accumulating elevation errors and computational limitations. The methodology was tested on the high mountain glacier of Zhadang, on the Tibetan Plateau. An optical acquisition from the Earth Observation (EO-1) satellite with the on board Advanced Land Imager (ALI) was used. The estimated surface model was assessed and validated with another elevation dataset and leveling transects from a surveying campaign. The resulting surface model needs transformation to be of use for elevation modelling. Different strategies are used to resolve the residual unknowns of the transformation. Registration of surface models was conducted on the icefield of Lomonosovfonna, Svalbard. Laser altimetric measurements from the Ice, Cloud and Elevation Satellite (ICESat) with the on board Geoscience Laser Altimeter System (GLAS) instrument are used for this registration problem. Both geometric and temporal relations are included into the registration model, but only under specific circumstances the registration is reliable. A profile based approach of photoclinometry is also investigated. It may give topographic correction to resolve errors due to miss-alignment of the ICESat GLAS laser profiler. However, assumptions on terrain geometry were too optimistic to resolve in a reliable correction. Geomatics Geoscience and Remote Sensing Civil Engineering and Geosciences
|
|---|