Study on the radioactive pollution in urban features using knowledge-based fuzzy classification of VHR satellite imagery

Safety is of primary principles of living in human communities. Preparation and provision of necessary considerations for encountering hazards are main targets of the crisis management. Nuclear risks are one of hazards threatening in the human life. Since radioactive contaminants sustain for years after the incident, investigation into nuclear hazards and its damage on living environment and urban features is so vital. This study essentially aims at evaluating the risk of radioactive contaminants to urban land uses. Due to high resolution satellite images, remote sensing technology has been considered as an advanced technology to generate information covering urban areas. Information on land cover is one of the most important tools of management during crisis. Land cover maps can be prepared through techniques for high resolution satellite image processing and extracting urban features. In this study, the fuzzy object-oriented method is applied to classify such phenomena. In the proposed method, a fuzzy rule-based strategy and hierarchical model are employed to overcome noise between classes. Fuzzy rule-based classification method is used as well as optimization and improving features of multi-scale analysis. Considering blocks of WorldView2 sensor, 91% of object detection is implemented with an average accuracy. When classification image of urban features is produced, the risk of radioactive contaminants to each recognized object is determined based on EDEM model.