Glacier margin change monitoring on the basis of multi-temporal 3D point clouds

点击翻译按钮获取中文摘要

Global warming has resulted in a significant worldwide mass loss of glaciers, ice caps and ice sheets in recent decades. A wide range of glacier monitoring techniques on local or global scales have been developed in the past. Glacier monitoring often focuses on the measurement of the glacier margin retreat as a good indicator for changes, which can be obtained from monoscopic satellite, airborne or terrestrial imagery. While clean glacier ice and snow can easily be segmented in multi-spectral images, the reliable detection of the actual margin of glaciers covered by debris or rock glaciers is challenging and is a major contributor to uncertainty of glacier inventories.The goal of the proposed project is the development of methods to reliably map the changes of the margins of such debris-covered glaciers. The key point of our proposal is to perform change detection for debris-covered glaciers on the basis of multi-temporal 3D point clouds rather than images. These surface models will be derived from stereoscopic imagery by state-of-the-art photogrammetric image matching techniques and allow for the analysis of height changes induced by glacier retreat or advance. The segmentation of regions with significant height changes will then indicate the actual glacier margins. The problems arising in detecting glacier margins in 2D images will thus be solved by a 4D approach via change analysis in consecutive surface models. The methods will be developed with respect to different scales, ranging from mid resolution satellite imagery over high resolution satellite imagery and aerial imagery down to UAV (unmanned aerial vehicle) images and will be validated in two pilot studies.