ID: 57307
Title: Modeling alpha- and beta-diversity in a tropical forest from remotely sensed and spatial data
Author: J Luis Hernandez-Stefanoni, J Alberto Gallardo-Cruz, Jorge A Meave, Duccio Rocchini, Javier Bello-Pineda, J Omar Lopez-Martinez
Editor: F. D. van der Meer
Year: 2012
Publisher: Elsevier, Vol 19, October 2012
Source: Centre for Ecological Sciences
Reference: None
Subject: International Journal of Applied Earth Observation and Geoinformation
Keywords: Image texture, PCNM analysis, Regression kriging, Remote sensing, Species richness, Species turnover, Tropical dry forest
Abstract: Comprehensive information on species distribution and species composition patterns of plant communities is required for effective conservation and management of biodiversity. Remote sensing offers an inexpensive means of attaining complete spatial coverage for large area, at regular time intervals, and can therefore be extremely useful for estimating both species richness and spatial variation fo species composition (alpha-and beta-diversity). An essential step to map such attributes is to identify and understand their main drivers. We used remotely sensed data as asurrogate of plant productivity an dhabitat structure variables for explaining alpha-and beta-divresity, and evaluated the relative roles of productivity -habitat structure and spatial variables in explaining observed patterns of alpha-and beta-diversity by using a Principal Coordinates of Neighbor Matrices analysis. We also examined the relationship between remotely sensed and field data, in order to amp alpha- and beta-diversity at the landscape - level n the Yucatan Peninsula, using a regression kriging procedure. These two procedures integrate the relationship of species richness and spatial species turnover both with remotely sensed data and spatial structure. The empirical models so obtained can be used to predict species richness and variation in species composition, and they can be regarded as valuable tools not only for identifying areas with high local species richness (alpha-diversity), but also areas wit high species turnover (beta-diversity). Ultimately, information obtained in this way can help maximize the number of species preserved in a landscape.
Location: TE 15, Biologicalsciences Building, IISc
Literature cited 1: None
Literature cited 2: None
ID: 57306
Title: Land surface emissivity retrieval based on moisture index from LANDSAT TM satellite data over heterogeneous surfaces of Delhi city
Author: Javed Mallick, Chander Kumar Singh, S Shashtri, Atiqur Rahman, S Mukherjee
Editor: F. D. van der Meer
Year: 2012
Publisher: Elsevier, Vol 19, October 2012
Source: Centre for Ecological Sciences
Reference: None
Subject: International Journal of Applied Earth Observation and Geoinformation
Keywords: Emissivity, Normalized difference moisture index (NDMI), Surface temperature, Land use/land cover, LANDSAT TM, Urban Heat Island
Abstract: Emissivity and surface temperature enables better understanding of the overall urban land use/land cover classes and in turn helps in understanding the energy budget issues. In the present study it has been demonstrated that the notion of the assumed spectral emissivity (i.e. 1) induces errors in modeling the surface energy budget and urban climatology (micro-climate), especially over heterogeneous surface areas (urban) where emissivity is far smaller than unity.
An attempt has been made to derive emissivity by using normalized difference moisture index (NDMI). The emissivity at the satellite sensor channel in order to have least error in the surface temperature estimation. The estimated emissivity values over few land use/land cover (LULC) classes of LANDSAT TM have been compared with the literature values and field measurement emissivity data using infrared thermometer.
A strong correlation is observed between surface temperatures with NDMI over different LULC classes. A regression relation between these parameters has been estimated (Pearson ' s correlation of 0.938), indicating that surface temperatures can be predicted if NDMI values are known. The error in field data (in situ) and satellite derived surface temperature is within the range of 2-30C. The correlation coefficient between the satellite derived and field observed surface temperature is very high ? 0.942 (significant at p value = 0.01). The results suggest that the methodology is feasible to estimate NDMI, surface emissivity and surface temperature with reasonable accuracy over heterogeneous urban areas.
Location: TE 15, Biologicalsciences Building, IISc
Literature cited 1: None
Literature cited 2: None
ID: 57305
Title: The potential of WorldView - 2 for ortho-image production within the "Control with Remote Sensing Programme" of the European Commission
Author: Par Johan Astrand, Marco Bongiorni, Mattia Crespi, Francesca Fratarcangeli, Joanna Nowak Da Costa, Francesca Pieralice, Agnieszka Walczynska
Editor: F. D. van der Meer
Year: 2012
Publisher: Elsevier, Vol 19, October 2012
Source: Centre for Ecological Sciences
Reference: None
Subject: International Journal of Applied Earth Observation and Geoinformation
Keywords: Worldview-2 sensor, Image orientation, Accuracy analysis, DSM generation, Common Agricultural Policy, Global Monitoring for Environment and Security (GMES)
Abstract: The imagery acquired by the WorldView-2 sensor is of potential interest for the Control with Remote Sensing Programme of the European Commission, and therefore, should be assessed.
The horizontal accuracy of the ortho-images that can be derived from WorldView-2 imagery need to be considered since the Control with Remote Sensing guidelines that the one-dimensional Root Mean Square Error (RMSE) estimated on the external Check Points for any ortho-image should not exceed 2.5m in order to qualify WorldView -2 as a Very High Resolution prime sensor.
This work summarizes the results regarding the orientation tests for five totally overlapping WorldView-2 Panchromatic images acquired at the Maussane test site (Provence, southern France), two Pan-sharpened images acquired at the Cosenza test site (Southern Italy), and nine Pan-sharpened scenes also acquired at the Cosenza test site. Tests were carried out using the Geomatica 10.2 (PCI Geomatics), the ERDAS Imagine 2011, and the SISAR software using both a rigorous model and Rational Polynomial Functions (RPFs) model with Rational Polynomial Coefficient (RPCs).
The Hold-Out-Validation accuracy assessment method was considered by computing the RMSE of the residuals between estimated and reference positions of the Check Points for each horizontal component (East, North) by varying the number of Ground Control Points. In addition, the Leave-One-Out Cross Validation method was used to identify possible outliers.
The following conclusions regarding the orientation were drawn. The best orientation accuracy over both sites was reached using the RPFs model with the RPCs supplied by imagery metadata and by applying a shift refinement even with four Ground Control Points. Further, the orientation accuracy was practically software independent but displayed significant dependence on the off-andir angle (a higher accuracy for the lower off-nadir angle). Concerning the Coseza test size, it appeared more convenient to process the two long strips (up to 28 km) instead of the creesponding separate nine scenes, since for the former case the achieved orientation accuracy was practically the same, whereas the number of required Ground Control points was much lower.
In regards to the subsequent ortho-image validation, the already mentioned requirement, a one-dimensional RMSE below 2.5 m for eac horizontal component, can be reached provided that a good quality Digital Surface Model (DSM) is used. For the Maussane test site all the ortho-images satisfied the previous requirements. Since a DSM with an appropriate accuracy was not available, the results for the Cosenza test site were not considered significantly representative of the real potential of the WorldView -2 sensor, so are not presented.
In order to evaluate the potentialities of WorldView-2 stereopairs for generating a DSM suitable for ortho-imagery production, tests were carried out using overlapping images from the Maussane Site Test. The results indicated that proper stereopairs could be an effective additional resource in the event of a lack of external good quality DSMs
Location: TE 15, Biologicalsciences Building, IISc
Literature cited 1: None
Literature cited 2: None
ID: 57304
Title: Hyperspectral determination of feed quality constituents in temperate pastures: Effect of processing methods on predictive relationships from partial least squares regression
Author: Susanne Thulin, Michael J Hill, Alex Held, Simon Jones, Peter Woodgate
Editor: F. D. van der Meer
Year: 2012
Publisher: Elsevier, Vol 19, October 2012
Source: Centre for Ecological Sciences
Reference: None
Subject: International Journal of Applied Earth Observation and Geoinformation
Keywords: Pastures, spectrometer, quality, digestibility, cellulose, lignin, protein
Abstract: Development of predictive relationships between hyperspectral reflectance and the chemical constituents of grassland vegetation could support routine remote sensing assessement of food quality in standing pastures. In this study, partial least squares regression (PLSR) and spectral transforms are used to derive predictive models for estimation of crude protein and digestibility (quality), and lignin and cellulose (non-digestible fractions) from field-based spectral libraries and chemical assays acquired from diverse pasture sites in Victoria, Australia between 2000 and 2002.
The best predictive models for feed quality were obtained with continuum removal with spectral bands normalised to the depth of absorption features for digestibility (adjusted R2 = 0.82, root mean square error of prediction (RMSEP) = 3.94), and continuum removal with spectral bands normalised to the area of the absorption features for crude protein (adjusted R2 = 0.62, RMSEP = 3.18) and cellulose (adjusted R2 = 0.73, RMSEP = 2.37). The results for lignin were poorer with the best performing model based on the first derivative of log transformed reflectance (adjusted R2 = 0.44, RMSEP = 1.87). The best models were dominated by first derivative transforms, and by limiting the models to significant variables with "Jack-knifing". X-loading results identified wavelengths near or outside major absorption features as important predictors.
This study showed that digestibility, as a broad measure of feed quality, could be effectively predicted from PLSR derived models of spectral reflectance derived from field spectroscopy. The models for cellulose and crude protein showed potential for qualitative assessment; however the results for lignin were poor. Implementation of spectral prediction models such as these, with hyperspectral sensors having a high signal to noise ratio, could deliver feed quality information to complement spatial biomass and growth data, and improve feed management for extensive grazing systems.
Location: TE 15, Biologicalsciences Building, IISc
Literature cited 1: None
Literature cited 2: None
ID: 57303
Title: An angular vegetation index for imaging spectroscopy data-preliminary results on forest damage detection in the Bavarian National Park, Germany
Author: Fabian Ewald Fassnacht, Hooman Latifi, Barbara Koch
Editor: F. D. van der Meer
Year: 2012
Publisher: Elsevier, Vol 19, October 2012
Source: Centre for Ecological Sciences
Reference: None
Subject: International Journal of Applied Earth Observation and Geoinformation
Keywords: Vegetation indices, Imaging spectroscopy data, genetic algorithm, Forest damage mapping, TAI, Angular index, hyperspectral remote sensing
Abstract: A vegetation index (VI) is presented which is calculated as the inner angles of a triangle. The triangle is spanned between three distinct points (on a spectral curve of imaging spectorscopy data) which are defined for each individual pixel by the wavelength (x - axis) and the reflection value (y-axis). The ideal wavelengths of the three points are dependent on the response variable investigated. The case-study within this paper in which this angular VI is applied, aims at the development of an index to automatically detect spruce bark beetle infections in the Bavarian National Park, Germany. In order to determine the optimum wavelengths to separate damaged from non-damaged trees the three-angle-indices (TAIs) for all possible combinations of available wavelengths of HyMap imaging spectroscopy data in the vis-NIR-region (0.455-0.986?m) were calculated. The resulting 27,417 images served as input predictor variables to a genetic algorithm (GA) which used nearest centroid classifier as fitness functions to detect the most stable predictors and separate the six classes (three damage classes) defined within the study area.
The fitness functions intergrated in the GA reached classification accuracies of up to 94.8% when using forward selection models of the most stable genes featuring maximum 50 predictors. Based on those results, three from the original 27,417 predictors were extracted to map forest damages over the full image extend based on a support vector machines (SVMs) classification. We conducted the same experiment with 82 vegetation indices described in literature and achieved slightly lower GA overall accuracies of 86.8% using the nearest centroid classifier. The SVM maps produced with the three best VI predators were comparable to the TAI results.
Location: TE 15, Biologicalsciences Building, IISc
Literature cited 1: None
Literature cited 2: None
ID: 57302
Title: Arid land characterisation with EO-1 hyperion hyperspectral data
Author: R Jafari, M M Lewis
Editor: F. D. van der Meer
Year: 2012
Publisher: Elsevier, Vol 19, October 2012
Source: Centre for Ecological Sciences
Reference: None
Subject: International Journal of Applied Earth Observation and Geoinformation
Keywords: Hyperion, hyperspectral, spectral mixture analysis, arid environment, South Australia
Abstract: The low spectral resolution of multispectral satellite imagery limits its capbility for extracting information in arid environments with sparse vegetation cover. The higher spectral resolution of hyperspectral imagery may improve discrimination of different vegetation types, even with low cover. The aim of this study was to evaluate the potential of Earth Observing 1 (EO-1) Hyperion hyperspectral data to discriminate arid landscape components in the southern rangelands of South Australia. Hyperion imagery was analysed with spectral mixture analysis to discriminate spectrally distinct land cover components. Five distinct end-members were extracted: two associated with vegetation cover and the remaining three associated with different soils and surface gravel and stone. The end-members were characterised with field spectra collected by ASD Fieldspec Pro spectrometer. To confirm the identify of the end-members we also investigated relationships between their abundance and field cover data collected at 54 sample sites using a step-point technique. One vegetation end-member was significantly correlated with Cottonbush (Maireanaaphylla) vegetation cover (R2 = 0.89) that was distributed as patches throughout the study area. The second vegetation end-member mapped green and grey-green perennial shrubs (e.g.Mulga, Acacia aneura) and was significantly correlated with total vegetation cover (R2 = 0.68). The soil and surface gravel and stone were significantly correlated with the field estimates of these physical components.Despite the high spectral resolution of the Hyperion scene, spectral mixture analysis was unable to identify more than five meaningful spectral end-members in this arid environment. This may be the result of low vegetation cover of the region (28%), the lack of spectral contrast in arid vegetation types, and the ground resolution of Hyperion (900 m2) that reduced the ability to identify spectrally pure end-members to represent different land cover components.
Location: TE 15, Biologicalsciences Building, IISc
Literature cited 1: None
Literature cited 2: None
ID: 57301
Title: Cotton fields drive elephant habitat fragmentation in the Mid Zambezi valley, Zimbabwe
Author: Mbulisi Sibanda, Amon Murwira
Editor: F. D. van der Meer
Year: 2012
Publisher: Elsevier, Vol 19, October 2012
Source: Centre for Ecological Sciences
Reference: None
Subject: International Journal of Applied Earth Observation and Geoinformation
Keywords: Elephant, habitat loss cotton fields, GIS, Remote sensing
Abstract: In this study we tested whether cotton fields contribute more than cereal fields to African elephant (Loxodonta africana) habitat loss through its effects on wooland fragmentation in the Mid-Zambezi Valley, Zimbabwe. In order to test this hypothesis, we first mapped cotton and cereal fields using MODIS remotely sensed data. Secondly, we analysed the effect of the area of cotton and cereal fields on woodland fragmentation using regression analysis. We then related the fragmentation indices, particularly edge density with elephant distribution data to test whether elephant distribution was significantly related with woodland fragmentation resulting from cotton fields. Our results showed that cotton fields contributed more to woodland fragmentation than cereal fields. In addition, results showed that the frequency of the African elephant increased where cotton fields were many and small relative to cereal fields. We concluded that cotton fields are the main driver of woodland fragmentation and therefore elephant habitat in the Mid-Zambezi Valley compared with cereal fields.
Location: TE 15, Biologicalsciences Building, IISc
Literature cited 1: None
Literature cited 2: None
ID: 57300
Title: Identification of cropping activity in central and southern Queensland, Australia, with the aid of MODIS MOD13Q1 imagery
Author: M J Pringle, R J Denham, R Devadas
Editor: F. D. van der Meer
Year: 2012
Publisher: Elsevier, Vol 19, October 2012
Source: Centre for Ecological Sciences
Reference: None
Subject: International Journal of Applied Earth Observation and Geoinformation
Keywords: MODIS, NDVI, Time series, crop, Random forest
Abstract: Cropping activity has an important that extends beyond farming communities, to governments, private industries, and to scientific research. We have developed a remote sensing -based method to detect arable cropping activity in central and southern Queensland, Australia, based on time series analysis of the NDVI layer of MODIS-Terra MOD13Q1 (250-m pixel) imagery. Local auto-regression was used to characterise phenological cycles in the NDVI time series. A random forest was then used to model three broad classes of agricultural vegetation (Grazing, Summer Cropping and Winter Cropping), as a function of phenological metrics and the local variance of the NDVI time series. The latter was found to be the most important distinguishing factor between the three classes. Pixel-by-pixel predictions of the random forest were obtained bi-annually for the study area over a 10-year period. Moderate agreement was seen between the predictions of the random forest and (independent) visual interpretation of Landsat imagery (Cohen ' s) index of agreement, Kc, of 0.59). We then demonstrated how the random forest ' s predictions can be used to define the consistency of cropping activity at the spatial scale of an individual farm property; when compared with (independent) visual interpretation of landsat imagery the agreement was also moderate (Kc=0.68). In comparison with other crop-mapping approaches in the literature, our results have been achieved: (i) without restricting the method to annual NDVI time series ; (ii) without assuming that the time series is regularly spaced and periodic ; (iii) by considering only the ' greening-up ' phase of the phenological cycles.
Location: TE 15, Biologicalsciences Building, IISc
Literature cited 1: None
Literature cited 2: None
ID: 57299
Title: Temporal monitoring of groundcover change using digital cameras
Author: A Zerger, D Gobbett, C Crossman, P Valencia, T Wark, M Davies, R N Handcock, J Stol
Editor: F. D. van der Meer
Year: 2012
Publisher: Elsevier, Vol 19, October 2012
Source: Centre for Ecological Sciences
Reference: None
Subject: International Journal of Applied Earth Observation and Geoinformation
Keywords: Digital camera, Grazing impacts, Kangaroos, Image processing, restoration ecology
Abstract: This paper describes the development and testing of an automated method for detectign change in groundcover vegetation in response to kangaroo grazing usign visible wavelength digital photography. The research is seen as a precursor to the future deployment of autonomous vegetation monitoring systems (environmental sensor networks). The study was conducted over six months with imagery captured every 90 min and post-processed using supervised image processing techniques. Synchronous manual assessments of groundcover change were also conducted to evaluate the effectiveness of the automated procedures. Results show that for particular cover classes such as Live Vegetation and Bare Ground, there is excellent temporal concordance between automated and manual methods. However, litter classes were difficult to consistently differentiate. A limitation of the method is the inability to effectively deal with change in the vertical profile of groundcover. This indicates that the three dimensional structure related to species composition and plant traits play an important role in driving future experimental designs. The paper concudes by providing lessons for conducting future groundcover monitoring experiments.
Location: TE 15, Biologicalsciences Building, IISc
Literature cited 1: None
Literature cited 2: None
ID: 57298
Title: Exploring spatial evolution of economic clusters: A case study of Beijing
Author: Zhenshan Yang, Richard Sliuzas, Jianming Cai, Henk F L Ottens
Editor: F. D. van der Meer
Year: 2012
Publisher: Elsevier, Vol 19, October 2012
Source: Centre for Ecological Sciences
Reference: None
Subject: International Journal of Applied Earth Observation and Geoinformation
Keywords: Economic cluster, spatial cluster, spatial statistics, Moran ' s I, Beijing
Abstract: An identification of economic clusters and analysis their changing spatial patterns is important for understanding urban economic space dynamics. Previous studies, however, suffer from limitations as a consequence of using fixed geographically areas and not combining functional and spatial dynamics. The paper presents an approach, based on local spatial statistics and the case of Beijing to understand the spatial clustering of industries that are functionally interconnected by common or complementary patterns of demand or supply relations. Using register data of business establishments, it identifies economic clusters and analyses their pattern based on postcodes at different time slices during the period 1983-2002. The study shows how the advanced services occupy the urban centre and key sub centres. The Information and Communication Technology (ICT) cluster is mainly concentrated in the north part of the city and circles the urban centre, and the main manufacturing clusters are evoved in teh key subcentres. This type of oucomes improves understanding of urban-economic dynamics, which can support spatial and economic planning.
Location: TE 15, Biology Sciences Building, IISc
Literature cited 1: None
Literature cited 2: None
ID: 57297
Title: Adaptive thematic object extraction from remote sensing image based on spectral matching
Author: Cheng Qiao, Jiancheng Luo, Zhanfeng Shen, Zhiwen Zhu, Dongping Ming
Editor: F. D. van der Meer
Year: 2012
Publisher: Elsevier, Vol 19, October 2012
Source: Centre for Ecological Sciences
Reference: None
Subject: International Journal of Applied Earth Observation and Geoinformation
Keywords: Thematic object extraction, spectral matching, adaptive, remote sensing
Abstract: Thematic object extraction is of great significance to remote sensing applications. Its procedure is always complicated while the result is not so precise, especially for object with various subtypes. An adaptive extraction method based on spectral matching, considering both spectral and spatial information, is proposed to extract thematic object completely and accurately from remote sensing image. This method considers the essential spectral representation of thematic object through endmember selection, and then achieves complete extraction via "whole-local" scale spectral matching. Expeirments on ETM+ images to extract water and bare land are employed, and the results demonstrate the effectiveness and universality of this method through comparison with maximum likelihood classification and support vector machine (SVM) classification.
Location: TE 15, Biology Sciences Building, IISc
Literature cited 1: None
Literature cited 2: None
ID: 57296
Title: Integrated use of spatial and sematic relationship for extracting road networks from floating car data
Author: Jun Li, Qiming Qin, Chao Xie, Yue Zhao
Editor: F. D. van der Meer
Year: 2012
Publisher: Elsevier, Vol 19, October 2012
Source: Centre for Ecological Sciences
Reference: None
Subject: International Journal of Applied Earth Observation and Geoinformation
Keywords: Floating car data, road network extraction, road map undating, sematic relationship
Abstract: The update frequency of digital road maps influences the quality of road-dependent services. However digital road maps surveyed by probe vehicles or extracted from remotely sensed images still have a long updating circle and their cost remain high. With GPS technology and wireless communication technology maturing and their cost decreasing, floating car technology has been used in traffic monitoring and management, and the dynamic positioning data from floating cars become a new data source for updating road maps. In the paper, we aim to update digital road maps using the floating car data from China ' s National Commercial Vehicle Monitoring Platform, and present an incremental road network extraction method suitable for the platform ' s GPS data whose sampling frequency is low and which cover a large area. Based on both spatial and sematic relationships between a trajectory point and its associated road segment, the method classifies each trajectory point, and then merges every trajectory point into the candidate road network through the adding or modifying process according to its type. The road network is gradually updated until all trajectories have been processed. Finally, this method is applied in the updating process of major roads in North China and the experimental results reveal that it can accurately derive geometric information of roads under various scenes. This paper provides a highly-efficient, low-cost approach to update digital road maps.
Location: TE 15, Biology Sciences Building, IISc
Literature cited 1: None
Literature cited 2: None
ID: 57295
Title: Spatial variability of terrestrial laser scanning based leaf area index
Author: Guang Zheng, L Monika Moskal
Editor: F. D. van der Meer
Year: 2012
Publisher: Elsevier, Vol 19, October 2012
Source: Centre for Ecological Sciences
Reference: None
Subject: International Journal of Applied Earth Observation and Geoinformation
Keywords: Spatial variation, terrestrial laser scanning (TLS), heterogeneous forest, effective leaf area index (LAIe), occlusion
Abstract: Forest stand point clouds generated from multiple scan locations using terrestrial laser scanning (TLS) have diverse range of spatial distribution patterns. These in turn have an effect on the direct leaf area index (LAI) estimation from the point cloud. However, the most effective placement of the scanning equipment in homogeneous vs heterogeneous stands has not been investigated. In this research, TLS was used to sample an evenly planted Douglas-fir (Pseudotsuga menziesii) seedling forest stand and a mature heterogeneous forest stand dominated by Douglas-fir (P menziesii) and Western hemlock (Tsuga heterophylla). A new method, circular point cloud slicing, was developed to explore the spatial variationof point density for both azimuthal angular and radial directions. The results show that alone, a central location 3600 scan data, does not capture all of the stand characteristics and less than 50% of variation of the estimation of effective leaf area index (LSIe) of a mature heterogeneous stand. Thus, reducing occlusion, by incorporating additional lateral side view scans, is necessary to comprehesively represent the canopy structure, and structural variation of the heterogeneous forest stand. It was also shown, based on thd assumption that the comprehensive scan combination can fully represent the forest stand, and that LAIe estimated from the comprehensive multi-direction mosaiced dataset are higher by twofold compared to the result from central scan only.
Location: TE 15, Biology Sciences Building, IISc
Literature cited 1: None
Literature cited 2: None
ID: 57294
Title: Spatial and thematic assessment of object-based forest stand delineation using an OFA-matrix
Author: A Hernando, D Tiede, F Albrecht, S Lang
Editor: F. D. van der Meer
Year: 2012
Publisher: Elsevier, Vol 19, October 2012
Source: Centre for Ecological Sciences
Reference: None
Subject: International Journal of Applied Earth Observation and Geoinformation
Keywords: Forest management, stands delineation, object fate analysis, thematic accuracy, OFA-matrix
Abstract: The delineation and classification of forest stands is a crucial aspect of forest management. Object-based image analysis (OBIA) can be used to produce detailed maps of forest stands from either orthophotos or very high resolution satellite imagery. However, measures are then required for evaluating and quanitfying both the spatial and thematic accuracy of the OBIA output. In this apper we present an approach for delineating forest stands and a new Object Fate Analysis (OFA) matrix for accuracy assessment. A two-level object -based orthophoto analysis was first carried out to delineate stands on the Dehesa Boyal public land in central Spain (Avila Province). Two structural features were first created for use in class modelling enabling good differentiation between stands: a relational tree cover cluster feature, and an arithmetic ratio shadow/tree feature. We then extended the OFA comparison approach with an OFA-matrix to enable concurrent validation of thematic and spatial accuracies. Its diagonal shows the proportion of spatial and thematic coincidence between a reference data and the corresponding classification. New parameters for Spatial Thematic Loyalty (STL), Spatial Thematic Loaylty Overall (STLOVERALL) and Maximal Interfering Object (MIO) are introduced to summarise the OFA- matrix accuracy assessment. A stands map generated by OBIA (classification data) was compared with a map of the same area produced from photo interpretation and field data (reference data). In our example the OFA-matrix results indicate good spatial and thematic accuracies (>65%) for all stand classes except for the shrub stands (31.8%) and a good STLOVERALL(69.8%). The OFA-matrix has therefore been shown to be a valid tool for OBIA accuracy assessment.
Location: TE 15, Biology Sciences Building, IISc
Literature cited 1: None
Literature cited 2: None
ID: 57293
Title: Change detection approaches for flood extent mapping: How to select the most adequate reference image from online archives?
Author: R Hostache, P Matgen, W Wagner
Editor: F. D. van der Meer
Year: 2012
Publisher: Elsevier, Vol 19, October 2012
Source: Centre for Ecological Sciences
Reference: None
Subject: International Journal of Applied Earth Observation and Geoinformation
Keywords: SAR image, Change detection, Reference image, floods
Abstract: Synthetic Aperture Radar images are routinely used for delineating flooded areas. Processing algorithms are often based on change detection techniques that enable a comparison of backscattering signals between the flood image and a reference image. However, as of today, there is little guidance on how to rapidly and reliably extract the most adequate reference image from an online archive. Our study proposes a method that allows the automatic and objective identification of the best reference image with respect to a given flood image. The proposed method consists of two processing steps. First, a subset of archived candidate images acquired on the same track, with the same polarization and in the same period of year as the flood image is created. Next, site-specific time series of regional backscattering values are established and the effects of flooding on the backscattering behaviour are statistically evaluated. We propose two complementary anomaly indexes and their combination in a single index as a means to identify the most adequate reference image for flooding-related change detection application. The reliability of the proposed method is demonstrated in three representative case studies targeting the flood prone areas of the Severn River (United Kingdom), the Red River (United States) and the Meghna River (Bangladesh).
Location: TE 15, Biology Sciences Building, IISc
Literature cited 1: None
Literature cited 2: None