Gao Y.,University of Morelia |
Mas J.F.,University of Morelia |
Kerle N.,International Institute for Geoinformation Science and Earth Observation ITC |
Pacheco J.A.N.,University of Morelia
International Journal of Remote Sensing | Year: 2011
Image segmentation is a preliminary and critical step in object-based image classification. Its proper evaluation ensures that the best segmentation is used in image classification. In this article, image segmentations with nine different parameter settings were carried out with a multi-spectral Landsat imagery and the segmentation results were evaluated with an objective function that aims at maximizing homogeneity within segments and separability between neighbouring segments. The segmented images were classified into eight land-cover classes and the classifications were evaluated with independent ground data comprising 600 randomly distributed points. The accuracy assessment results presented similar distribution as that of the objective function values, that is segmentations with the highest objective function values also resulted in the highest classification accuracies. This result shows that image segmentation has a direct effect on the classification accuracy; the objective function not only worked on a single band image as proved by Espindola et al. (2006, Parameter selection for region-growing image segmentation algorithms using spatial autocorrelation. International Journal of Remote Sensing, 27, pp. 3035-3040) but also on multi-spectral imagery as tested in this, and is indeed an effective way to determine the optimal segmentation parameters. McNemar's test (z2 1/4 10.27) shows that with the optimal segmentation, object-based classification achieved accuracy significantly higher than that of the pixel-based classification, with 99% significance level. © 2011 Taylor & Francis.
Anitha K.,Mahatma Gandhi University |
Anitha K.,Salim Ali Center for Ornithology and Natural History |
Joseph S.,Mahatma Gandhi University |
Joseph S.,International Institute for Geoinformation Science and Earth Observation ITC |
And 3 more authors.
Ecological Complexity | Year: 2010
The structure, function, and ecosystem services of tropical forest depend on its species richness, diversity, dominance, and the patterns of changes in the assemblages of tree populations over time. Long-term data from permanent vegetation plots have yielded a wealth of data on the species diversity and dynamics of tree populations, but such studies have only rarely been undertaken in tropical forest landscapes that support large human populations. Thus, anthropogenic drivers and their impacts on species diversity and community structure of tropical forests are not well understood. Here we present data on species diversity, community composition, and regeneration status of tropical forests in a human-dominated landscape in the Western Ghats of southern India. Enumeration of 40 plots (50m × 20 m) results a total of 106 species of trees, 76 species of saplings and 79 species of seedlings. Detrended Correspondence Analysis ordination of the tree populations yielded five dominant groups, along disturbance and altitudinal gradients on the first and second axes respectively. Abundant species of the area such as Albizia amara, Nothopegia racemosa and Pleiospermum alatum had relatively few individuals in recruiting size classes. Our data indicate probable replacement of rare, localized, and old-growth 'specialists' by disturbance-adapted generalists, if the degradation is continuing at the present scale. © 2010 Elsevier B.V.
Kuria Z.N.,University of Nairobi |
Kuria Z.N.,International Institute for Geoinformation Science and Earth Observation ITC |
Woldai T.,International Institute for Geoinformation Science and Earth Observation ITC |
Meer F.D.v.d.,International Institute for Geoinformation Science and Earth Observation ITC |
Barongo J.O.,University of Nairobi
Journal of African Earth Sciences | Year: 2010
Southern Kenya Rift has been known as a region of high geodynamic activity expressed by recent volcanism, geothermal activity and high rate of seismicity. The active faults that host these activities have not been investigated to determine their subsurface geometry, faulting intensity and constituents (fluids, sediments) for proper characterization of tectonic rift extension. Two different models of extension direction (E-W to ESE-WNW and NW-SE) have been proposed. However, they were based on limited field data and lacked subsurface investigations. In this research, we delineated active fault zones from ASTER image draped on ASTER DEM, together with relocated earthquakes. Subsequently, we combined field geologic mapping, electrical resistivity, ground magnetic traverses and aeromagnetic data to investigate the subsurface character of the active faults. Our results from structural studies identified four fault sets of different age and deformational styles, namely: normal N-S; dextral NW-SE; strike slip ENE-WSW; and sinistral NE-SW. The previous studies did not recognize the existence of the sinistral oblique slip NE-SW trending faults which were created under an E-W extension to counterbalance the NW-SE faults. The E-W extension has also been confirmed from focal mechanism solutions of the swarm earthquakes, which are located where all the four fault sets intersect. Our findings therefore, bridge the existing gap in opinion on neo-tectonic extension of the rift suggested by the earlier authors. Our results from resistivity survey show that the southern faults are in filled with fluid (0.05 and 0.2 Ωm), whereas fault zones to the north contain high resistivity (55-75 Ωm) material. The ground magnetic survey results have revealed faulting activity within active fault zones that do not contain fluids. In addition, the 2D inversion of the four aero-magnetic profiles (209 km long) revealed: major vertical to sub vertical faults (dipping 75-85° east or west); an uplifted, heavily fractured and deformed basin to the north (highly disturbed magnetic signatures) characteristic of on going active rifting; and a refined architecture of the asymmetry graben to the south with an intrarift horst, whose western graben is 4 km deep and eastern graben is much deeper (9 km), with a zone of significant break in magnetic signatures at that depth, interpreted as source of the hot springs south of Lake Magadi (a location confirmed near surface by ground magnetic and resistivity data sets). The magnetic sources to the north are shallow at 15 km depth compared to 22 km to the south. The loss of magnetism to the north is probably due to increased heat as a result of magmatic intrusion supporting active rifting model. Conclusively, the integrated approach employed in this research confirms that fault system delineated to the north is actively deforming under E-W normal extension and is a potential earthquake source probably related to magmatic intrusion, while the presence of fluids within the south fault zone reduce intensity of faulting activity and explains lack of earthquakes in a continental rift setting. © 2009 Elsevier Ltd. All rights reserved.