No description available.

No description available.

No description available.

No description available.

A huge effort has been made in the development of image classification systemswith the objective of creating high-quality thematic maps and to establishprecise inventories about land cover use. The peculiarities of Remote SensingImages (RSIs) combined with the traditional image classification challengesmake RSI classification a hard task. Many of the problems are related to therepresentation scale of the data, and to both the size and therepresentativeness of used training set.In this work, we addressed four research issues in order to develop effectivesolutions for interactive classification of remote sensing images.The first research issue concerns the fact that image descriptorsproposed in the literature achieve good results in various applications, butmany of them have never been used in remote sensing classification tasks.We have tested twelve descriptors that encodespectral/color properties and seven texture descriptors. We have also proposeda methodology based on the K-Nearest Neighbor (KNN) classifier for evaluationof descriptors in classification context. Experiments demonstrate that JointAuto-Correlogram (JAC), Color Bitmap, Invariant Steerable Pyramid Decomposition(SID), and Quantized Compound Change Histogram (QCCH) yield the best results incoffee and pasture recognition tasks.The second research issue refers to the problem of selecting the scaleof segmentation for object-based remote sensing classification. Recentlyproposed methods exploit features extracted from segmented objects to improvehigh-resolution image classification. However, the definition of the scale ofsegmentation is a challenging task. We have proposedtwo multiscale classification approaches based on boosting of weak classifiers.The first approach, Multiscale Classifier (MSC), builds a strongclassifier that combines features extracted from multiple scales ofsegmentation. The other, Hierarchical Multiscale Classifier (HMSC), exploits thehierarchical topology of segmented regions to improve training efficiencywithout accuracy loss when compared to the MSC. Experiments show that it isbetter to use multiple scales than use only one segmentation scale result. Wehave also analyzed and discussed about the correlation among the useddescriptors and the scales of segmentation.The third research issue concerns the selection of training examples and therefinement of classification results through multiscale segmentation. We have proposed an approach forinteractive multiscale classification of remote sensing images.It is an active learning strategy that allows the classification resultrefinement by the user along iterations. Experimentalresults show that the combination of scales produces better results thanisolated scales in a relevance feedback process. Furthermore, the interactivemethod achieves good results with few user interactions. The proposed methodneeds only a small portion of the training set to build classifiers that are asstrong as the ones generated by a supervised method that uses the whole availabletraining set.The fourth research issue refers to the problem of extracting features of ahierarchy of regions for multiscale classification. We have proposed a strategythat exploits the existing relationships among regions in a hierarchy. Thisapproach, called BoW-Propagation, exploits the bag-of-visual-word model topropagate features along multiple scales. We also extend this idea topropagate histogram-based global descriptors, the H-Propagation method. The proposedmethods speed up the feature extraction process and yield good results when compared with globallow-level extraction approaches.

Vegetation community patterns and processes are indicators and integrators of climate. Recently, scientists have shown that climate change is most pronounced in circumpolar regions. Arctic ecosystems have traditionally been sequestering carbon and accumulating large carbon stores. However, given enhanced warming in the Arctic, the potential exists for intensified global climate change if these ecosystems transition from sinks to sources of atmospheric CO2. In the Mid and High Arctic, ecosystems exhibit extreme levels of spatial heterogeneity, particularly at landscape scales. High spatial-resolution (e.g., 4m) remote sensing data capture heterogeneous vegetation patterns of the Arctic landscape and have the potential to model ecosystem biophysical properties and CO2 fluxes. The following conditions are required to model arctic ecosystem processes: (i) unique spectral signatures that correspond to variations in the landscape pattern; (ii) models that transform remote sensing data into derivative values pertaining to the landscape; and (iii) field measures of the variables to calibrate and validate the models. First, this research creates an ecosystem classification scheme through ordination, clustering, and spectral-separability of ground cover data to generate ecologically meaningful and spectrally distinct image classifications. Classifications had overall accuracies between 69% - 79% and Kappa values of 0.54 - 0.69. Secondly, biophysical variable models of percent vegetation cover, aboveground biomass, and soil moisture are calibrated and validated using a k-fold cross-validation linear bivariate regression methodology. Percent vegetation cover and percent soil moisture produce the strongest and most consistent results (r2 ≥ 0.84 and 0.73) across both study sites. Finally, in situ CO2 exchange rate data, an NDVI model for each component flux, which explains between 42% and 95% of the variation at each site, is generated. Analysis of coincidence indicates that a single model for each component flux can be applied, independent of site. This research begins to fill a gap in the application of high spatial-resolution remote sensing data for modelling Arctic ecosystem biophysical variables and carbon dioxide exchange, particularly in the Canadian Arctic. The results of this research also indicate high levels of functional convergence in ecosystem-level structure and function within Arctic landscapes. / Thesis (Ph.D, Geography) -- Queen's University, 2013-01-03 22:24:20.157

Climate warming has the potential to alter the structure and function of Arctic ecosystems in ways that are not fully understood. Polygonal terrain is a widespread permafrost feature of Arctic landscapes that is likely to be impacted by warming ground temperatures. This is of particular relevance in the uplands in the Mackenzie Delta region, where high-centre ice wedge polygon fields comprise 10% of the terrestrial landscape, and mean annual ground temperatures have increased between 1 and 2°C over the last 40 years (Burn and Kokelj 2009). I used broad-scale airphoto analysis and fine-scale field studies to investigate the impacts and possible trajectories of ice wedge degradation in the upland tundra north of Inuvik, NWT. Field investigations were undertaken to characterize biotic and abiotic conditions and feedbacks in stable and degrading high-centre polygons. Field surveys were conducted along transects which crossed three polygon micropositions (centres, edges and troughs) and targeted a degradation sequence from stable troughs to ice wedge melt ponds. I measured surface microtopography, active layer depth, water depth, plant community composition, soil gravimetric moisture, late winter snow depth, and shallow annual ground temperatures. Field data showed that ice wedge degradation drove increases in soil moisture, standing water depth, ground surface collapse, ground temperature, and active layer thaw and snow pack compared to stable troughs. These changing abiotic conditions drove the shift from mesic upland tundra plant communities to unvegetated melt ponds. Interactions between abiotic and biotic factors in degrading troughs increase ground temperature and contribute to positive feedbacks for ice wedge degradation. Analysis of broad-scale factors affecting ice wedge degradation involved the mapping of high-centre polygon distribution across the study area and the distribution of ice wedge melt ponds using high-resolution aerial photographs from 2004. Recent changes in melt pond area were also mapped using imagery dating from 1972. Thermokarst activity in polygonal terrain adjacent to anthropogenic disturbances was also assessed. Polygon fields were more abundant and larger in the northern part of the study area, where ground temperature conditions were most favourable for ice wedge formation. Spatial variation in polygonal terrain density was also related to topography, drainage, and the distribution of lacustrine sediments. Melt pond mapping and assessment of thermokarst at anthropogenic disturbances showed that ice wedges at higher latitudes are more susceptible to degradation primarily because these areas are underlain by larger and more abundant ice wedges. Melt pond mapping confirmed that the polygonal fields north of 69.4°N have shown both large increases and decreases in area, and that polygons in the south have been relatively stable in recent decades. The increased thaw sensitivity of polygonal terrain at higher latitudes has implications for soil carbon dynamics, terrestrial ecosystems, and the planning and maintenance of infrastructure as air and ground temperatures continue to increase. / Graduate / 0329 / 0372 / 0388

Laser altimetry and satellite gravity surveys indicate that the St Elias Icefields are currently losing mass and are among the largest non-polar sea level contributors in the world. However, a poor understanding of glacier dynamics in the region is a major hurdle in evaluating regional variations in ice motion and the relationship between changing surface conditions and ice flux. This study combines in-situ dGPS measurements and advanced Radarsat-2 (RS-2) processing techniques to determine daily and seasonal ice velocities for the Kaskawulsh Glacier from summer 2009 to summer 2011. Three permanent dGPS stations were installed along the centreline of the glacier in 2009, with an additional permanent station on the South Arm in 2010. The Precise Point Positioning (PPP) method is used to process the dGPS data using high accuracy orbital reconstruction. RS-2 imagery was acquired on a 24-day cycle from January to March 2010, and from October to March 2010-2011 in a combination of ultra-fine and fine beam modes. Seasonal velocity regimes are readily identifiable in the dGPS results, with distinct variations in both horizontal velocity and vertical motion. The Spring Regime consists of an annual peak in horizontal velocity that corresponds closely with the onset of the melt season and progresses up-glacier, following the onset of melt at each station. The Summer Regime sees variable horizontal velocity and vertical uplift, superimposed on a long-term decline in motion. The Fall Regime sees a gradual slowing at all stations with little variation in horizontal velocity or vertical position. Rapid but short accelerations lasting up to 10 days were seen in the Winter regimes in both 2010 and 2011, occurring at various times throughout each regime. These events initiated at the Upper Station and progress down-glacier at propagation speeds up to 16,380 m day-1 and were accompanied by vertical uplift lasting for similar periods. Three velocity maps, one from the winter of 2010 and two from the fall/winter of 2011, produced from speckle tracking were validated by comparison with dGPS velocity, surface flow direction, and bedrock areas of zero motion, with an average velocity error of 2.0% and average difference in orientation of 4.3º.

The goal of quantifying the woody cover and biomass of tropical savannas, woodlands and forests using satellite data is becoming increasingly important, but limitations in current scientific understanding reduce the utility of the considerable quantity of satellite data currently being collected. The work contained in this thesis reduces this knowledgegap, using new field data and analysis methods to quantify changes using optical, radar and LiDAR data. The first paper shows that high-resolution optical data (Landsat & ASTER) can be used to track changes in woody vegetation in the Mbam Djerem National Park in Cameroon. The method correlates a satellite-derived vegetation index with field-measured canopy cover, and the paper concludes that forest encroached rapidly into savanna in the region from 1986-2006. Using the same study area, but with radar remote sensing data from 1996 and 2007 (ALOS PALSAR & JERS-1), the second paper shows that radar backscatter correlates well with field-measured aboveground biomass (AGB). This dataset confirms the woody encroachment within the park; however, in a larger area around the park, deforestation dominates. The AGB-radar relationships described above are expanded in the next paper to include field plots from Budongo Forest (Uganda), the Niassa Reserve (north Mozambique), and the Nhambita Community Project (central Mozambique). A consistent AGB-radar relationship is found in the combined dataset, with the RMSE for predicted AGB values for a site increasing by <30 %, compared with a site-specific equation, when using an AGB-radar equation derived from the three other sites. The study of the Nhambita site is extended in the following paper to assess the ability of radar to detect change over short time periods in this environment, as will be needed for REDD (Reducing Emissions from Deforestation and Degradation). Using radar mosaics from 2007 and 2009, areas known (from detailed ground data) to have been degraded decreased in AGB in the radar change detection, whereas areas of agroforestry and forest protection showed small increases.

U.S. Army Corps of Engineers (USACE) regulates fill of jurisdictional waters of the United States including wetlands. Recent USACE regulations set a threshold of impacts to wetlands at one-half acre. Impact area can be determined by Global Positioning System (GPS) measurement of wetland boundary and Geographic Information System (GIS) calculation of impact area. GPS point selection methods include (1) equal time interval, (2) transect and (3) intuition. Four two-acre shapes were measured with each GPS method and brought into GIS for area calculation. Analysis of variance and Root Mean Square Error analyses determine that the transect method is an inferior point selection method in terms of accuracy and efficiency.