Work Package 3: Multi-temporal Analysis and Monitoring

WP3.1: Development of Advanced Change Detection Techniques

Objective: This WP aims at developing advanced automatic techniques for change detection analysis. The work will focus on the analysis of time-varying variables (WP 3.2 & WP 3.3) by processing heterogeneous data (optical data at different resolution, SAR data) and exploiting auxiliary data, for a range of applications.

Description:
A) Development of accurate registration and pan-sharpening methods for the analysis of spectral and spatial heterogeneous multi-source data.


B) Development of CD techniques for the analysis of multi-source data (SAR, optical and auxiliary data) to detect multiple types of changes from time series heterogeneous RS images.


C) Development of techniques for the contextual analysis of images of high geometrical resolution in order to obtain highly detailed change detection maps. The analysis will include hierarchical Markov random fields, irregular lattices, region-based analysis and mathematical morphology approaches for the contextual analysis.


D) Improving the change detection analysis for hyperspectral images based on the reference data acquired during the field campaigns in WP1. Definition of benchmark data for future change detection analysis.

WP3.2: Multi-temporal geological analysis

Objective: The objective is to monitor geological units and is divided into (A) Volcanic monitoring (B) Glacial monitoring,and (C) Fluvial and mass movement monitoring.

Description:
(A) Volcanic monitoring (IES, SPL, BGS, ZGIS: G.B.M. Pedersen)

Both the Hekla and Öræfajökull area has been affected by volcanic eruptions in the time period covered by our data archive. Four mixed Hekla eruptions (1970, 1980-1981, 1991, 2000) have produced lava and tephra deposited in the Hekla area, while the Öræfajökull area has been affected by ash from the Grímsvötn eruptions (1983,1998, 2004, 2011).


1) Lava flow thickness and volume calculation of lava flows from the four Hekla eruptions by differencing historical DEMs created based on stereophotogrammetry of the aerial images. The necessary software and expertise for stereophotogrammetry of historical imagery does already exist in the department of IES. This will provide unprecedented precision of the lava flow volume and provide a detailed spatial understanding of the variability of the lava flow thickness with respect to the subsurface (Stevens et al. 2001; Wadge et al. 2006, Neri et al. 2008). This data will be compared with the assessment of the lava flow morphology (WP2.2), but willalso provide very critical data for lava flow modelling(WP4).


2) Mapping the changing spatial extent of the tephra cover through change detection techniques (see WP3.1), which can be important for rapid response mapping. We will correlate the multi-temporal tephra cover to mapped geologic units (WP 2.2) and to vegetation changes (WP3.3) in order to assess the post-eruptive redistribution of tephra, the spatial preservation of tephra layers and the environmental impact of the tephra.

(B) Glacial Monitoring (IES, SPL, ZGIS: G. Aðalgeirdóttir & G.B.M. Pedersen)
1) Estimation of glacial volume loss in the second half of 20th century for decadal time periods from 1940 to 2010 by differencing historical DEMs created based on stereophotogrammetry of the aerial images.
2) Use automatic change detection techniques developed in WP3.1 to investigate (A) Changes in glacial outlines (B) Changes in firn-line location (C) Changes in glacial reflectance and (D) Changes in crevasse fields.

(C) Fluvial and mass movement monitoring (ZGIS, IES,SPL, DBS: D.Hölbling)

The research areas are exposed to significant mass movements and fluvial changes, which may impact infrastructure. Furthermore, landslide events are potential hazards, and it is crucial to get an overview of landslide frequency, the mass movement rate and the spatial correlation to geologic units and vegetation. The following approaches are planned:


1) Map mass movements and fluvial changes based on change detection techniques (WP3.1).


2) Apply semi-automated OBIA methods to map mass movements developed by ZGIS (e.g. Hölbling et al., 2012; Hölbling et al., accepted; Eisank et al., 2014). Differentiation of mass movement types (Hölbling et al., 2012). The applied change detection techniques from WP3.1 will be compared with the results from the semi-automated OBIA classificationand to existing mass movement data sets in Iceland. The developed methods can be applied for the regular update of existing landslide inventory maps.

WP3.3: Multi-temporal vegetation analysis

Objective: 
The aim of this WP is to detect and analyse changes in vegetation cover and distribution of plant communities.

Description: 
The vegetation map generated in WP 2.3 will serve as a basis for assessing past changes in plant communities and vegetation cover by using the historical RS data archive.


The analysis aims at detecting the effects of land use on vegetation, such as arming, grazing, land reclamation and tourism. They include changes in (i) total vegetation cover, (ii) plant group/species composition, (iii) woodland distribution, (iv) lupin dispersion, (v) biomass changes using vegetation indexes.

EMMIRS Communication site for WP3

Loading...

 

Loading...