The research activity of Numip area is devoted to image processing and analysis techniques, focused at the numerical level, towards the symbolic level. A major aim is to prove the added value of the information that can be extracted from real imaging data, thanks to filtering, enhancement, segmentation and data-fusion advanced methods. Earth Observation by satellite data in the monitoring of flooded areas, image and volume analysis of Tomographic and Ultrasound acquisitions are only two specific application examples. Robustness is faced by reducing the use of parameters (thanks to adaptive and fuzzy approaches); numerical quality metrics are proposed for non-linear filtering performance evaluations. General purpose image processing and segmentation methods have allowed the development of the Content-based image retrieval (CBIR) system “Kingfisher”, of the multi-temporal/multi-parametric 3D image enhancement and display method, and of the fire/smoke detection system.