Special Issue "Object-Based Image Analysis"
A special issue of Remote Sensing (ISSN 2072-4292).
Deadline for manuscript submissions: closed (30 July 2011)
Dr. Stefan Lang
Assoc-Prof, Division Head, Department of Geoinformatics - Z_GIS, University of Salzburg, Schillerstr. 30, 5020 Salzburg, Austria, Europe
Website | E-Mail
Phone: +43 662 8044-5262
Interests: Earth observation (EO); big Earth data; multi-scale spatial analysis; spatial data management; object-based image analysis (OBIA); knowledge representation; computer vision; multi-dimensional modelling; validation
Object-based image analysis (OBIA) has emerged over the last years from integrating geospatial concepts and advanced image analysis techniques. Spatial properties like size and form, neighborhood and context, scale and hierarchy, are utilized for better exploit imagery and other image-like continuous data. In parallel, advances in sensor technology and new processing methods (e.g. grid computing) has strongly supported the maturing of OBIA making it an established approach for image understanding.
OBIA conceptually utilizes two interrelated methodological pillars for effectively handling the complexity of recent highest resolution imagery. These are: (1) segmentation / regionalization for image object delineation and providing scaled and hierarchical representations; (2) advanced classifiers reflecting the interaction between spectral and spatial properties by either using learning algorithms or making these explicit in transferable rule sets. The term image analysis (instead of image classification) is used to stress that the process of OBIA is iterative and cyclic.
Today, driven by international programmes and initiatives like GEO or GMES, the provision of imagery should no longer a bottle-neck per se (leaving aside specific requirements such as real time provision, particular detail or atmospheric conditions). The bigger challenge, instead, is the information extraction and the provision of added-value products with constant quality and reliability, high transferability and re-usability of algorithm and rule sets, and clear validation concept.
This special issue asks for papers that serve a growing scientific community interested in OBIA. Particular topics include: critical assessments of commercial software applications; new developments with potential of operational use; methodological and conceptual achievements for object-based change detection; and object validation and addressing complex (composite) classes.
Dr. Stefan Lang
- advanced classifiers
- scaled representation