作者: Stefanos Vrochidis , Victor Epitropou , Anastasios Bassoukos , Sascha Voth , Kostas Karatzas
DOI: 10.1007/978-3-642-33412-2_37
关键词: Air quality index 、 Computer science 、 Environmental data 、 Line (text file) 、 Optical character recognition 、 Data mining 、 Task (project management) 、 Image map 、 Variety (cybernetics) 、 Image processing
摘要: Analysis of environmental information is considered utmost importance for humans, since conditions are strongly related to health issues and a variety everyday activities. Despite the fact that there already many free on-line services providing information, several cases, in which presentation format complicates extraction processing such data. A very characteristic example air quality forecasts, usually encoded image maps heterogeneous formats, while initial (numerical) pollutant concentrations, calculated predicted by relevant model, remain unavailable. This work addresses task semi-automatic based on template configuration tool, methodologies data reconstruction from images, as well Optical Character Recognition (OCR) techniques. The framework tested with number forecast heatmaps demonstrating satisfactory results.