Image Segmentation Methods for Object-based Analysis and Classification

Blaschke, Thomas; Burnett, Charles; Pekkarinen, Anssi

doi:10.1007/1-4020-2560-2_12

Cited by 60 publications

(87 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Edges are regarded as boundaries between image objects and they are located where major changes in Digital Number (DN) values occur. This division of the image into a set of homogenous regions is the baseline concept of image segmentation, which has been widely used for object recognition from imagery [42,43]. Image segmentation algorithms can be region growing or edge detectors (among others).…”

Section: Thresholdingmentioning

confidence: 99%

Operational Surface Water Detection and Monitoring Using Radarsat 2

et al. 2016

View full text Add to dashboard Cite

Traditional on-site methods for mapping and monitoring surface water extent are prohibitively expensive at a national scale within Canada. Despite successful cost-sharing programs between the provinces and the federal government, an extensive number of water features within the country remain unmonitored. Particularly difficult to monitor are the potholes in the Canadian Prairie region, most of which are ephemeral in nature and represent a discontinuous flow that influences water pathways, runoff response, flooding and local weather. Radarsat-2 and the Radarsat Constellation Mission (RCM) offer unique capabilities to map the extent of water bodies at a national scale, including unmonitored sites, and leverage the current infrastructure of the Meteorological Service of Canada to monitor water information in remote regions. An analysis of the technical requirements of the Radarsat-2 beam mode, polarization and resolution is presented. A threshold-based procedure to map locations of non-vegetated water bodies after the ice break-up is used and complemented with a texture-based indicator to capture the most homogeneous water areas and automatically delineate their extents. Some strategies to cope with the radiometric artifacts of noise inherent to Synthetic Aperture Radar (SAR) images are also discussed. Our results show that Radarsat-2 Fine mode can capture 88% of the total water area in a fully automated way. This will greatly improve current operational procedures for surface water monitoring information and impact a number of applications including weather forecasting, hydrological modeling, and drought/flood predictions.

show abstract

Section: Thresholdingmentioning

confidence: 99%

Operational Surface Water Detection and Monitoring Using Radarsat 2

et al. 2016

View full text Add to dashboard Cite

show abstract

“…Thereafter, a quantitative LULC change assessment is carried out to develop a land use change matrix, following international guidelines from GOFC-GOLD [10]. The land use change analysis entails localizing the expansion of the major proximate causes of deforestation and quantifying their area-wise impact in the past, for example through an object-based segmentation mapping, where the images are automatically analyzed for spectrally similar objects, divided in a second step into segments and finally [11]. This process should be carried out in two phases, first distinguishing between forest and non-forest and then repeating the object-based classification separately for forest and non-forest areas to produce a detailed land use classification [12].…”

Section: Step1: Data Gathering and Literature Reviewmentioning

confidence: 99%

A Methodological Framework for Assessing Agents, Proximate Drivers and Underlying Causes of Deforestation: Field Test Results from Southern Cameroon

Carodenuto¹,

Merger²,

Essomba

et al. 2015

Forests

View full text Add to dashboard Cite

Abstract:The international debates on REDD+ and the expectations to receive results-based payments through international climate finance have triggered considerable political efforts to address deforestation and forest degradation in many potential beneficiary countries. Whether a country will receive such REDD+ payments is largely contingent on its ability to effectively address the relevant drivers, and to govern the context-dependent agents and forces responsible for forest loss or degradation. Currently, many REDD+ countries are embarking on the necessary analytical steps for their national REDD+ strategies. In this context, a comprehensive understanding of drivers and their underlying causes is a fundamental prerequisite for developing effective policy responses. We developed a methodological framework for assessing the drivers and underlying causes of deforestation and use the Fako Division in Southern Cameroon as a case study to test this approach. The steps described in this paper can be adapted to other geographical contexts, and the results of such assessments can be used to inform policy makers and other stakeholders.

show abstract

“…Land-cover mapping is a common task in this context, where it is attempted to generate a partial or full description of a given area from Earth observation imagery, with an emphasis on element geometric and thematic accuracies. Geographic Object-Based Image Analysis (GEOBIA) has emerged as a viable avenue of approaches, or paradigm, to tackle such remote sensing image analysis tasks [1][2][3][4] due to the common spectral-textural-geometric and thematic correlations of elements of interest in satellite imagery [5][6][7].…”

Section: Introductionmentioning

confidence: 99%

Classifier Directed Data Hybridization for Geographic Sample Supervised Segment Generation

Fourie

Schöepfer

2014

Remote Sensing

View full text Add to dashboard Cite

Quality segment generation is a well-known challenge and research objective within Geographic Object-based Image Analysis (GEOBIA). Although methodological avenues within GEOBIA are diverse, segmentation commonly plays a central role in most approaches, influencing and being influenced by surrounding processes. A general approach using supervised quality measures, specifically user provided reference segments, suggest casting the parameters of a given segmentation algorithm as a multidimensional search problem. In such a sample supervised segment generation approach, spatial metrics observing the user provided reference segments may drive the search process. The search is commonly performed by metaheuristics. A novel sample supervised segment generation approach is presented in this work, where the spectral content of provided reference segments is queried. A one-class classification process using spectral information from inside the provided reference segments is used to generate a probability image, which in turn is employed to direct a hybridization of the original input imagery. Segmentation is performed on such a hybrid image. These processes are adjustable, interdependent and form a part of the search problem. Results are presented detailing the performances of four method variants compared to the generic sample supervised segment generation approach, under various conditions in terms of resultant segment quality, required computing time and search process characteristics. Multiple metrics, metaheuristics and segmentation algorithms are tested with this approach. Using the spectral data contained within user OPEN ACCESS Remote Sens. 2014, 6 11853 provided reference segments to tailor the output generally improves the results in the investigated problem contexts, but at the expense of additional required computing time.

show abstract

Image Segmentation Methods for Object-based Analysis and Classification

Cited by 60 publications

References 0 publications

Operational Surface Water Detection and Monitoring Using Radarsat 2

Operational Surface Water Detection and Monitoring Using Radarsat 2

A Methodological Framework for Assessing Agents, Proximate Drivers and Underlying Causes of Deforestation: Field Test Results from Southern Cameroon

Classifier Directed Data Hybridization for Geographic Sample Supervised Segment Generation

Contact Info

Product

Resources

About