In the urban environment and rural area, continuous monitoring at multiple scales (global, regional and local) is crucial in terms of productivity, resilience, and sustainability. Monitoring requires the production of observations that allows environment study, inspection and understanding at different levels of details. However, the information production involves repetitive and straightforward activities, i.e. survey, localisation, object detection, that should be done multiple times in different periods. The procedure could be time-consuming, expensive and risky. Indeed, in many applications, such as Precision Agriculture (PA) or farming, these tasks are accomplished still with the involvement of experts in manual mode or by using static sensor networks (SSNs).

As proposed by the “Rural Development Program 2014-2020”, monitoring requires innovative tools and sustainable solutions. Innovation concerns the acquisition and the rapid analysis of high spatial, temporal, and spectral resolution data. This means boosting the development of service robotics as more precise and integrated acquisition platforms and machine learning techniques for improving automatic information extraction techniques. 

The research, focusing on the geomatics aspects, consists of platform definition, sensors calibration, and definition of quasi real-time machine learning algorithms for near real-real time information extraction. Aircraft inspection for de-icing operations and precision agriculture are used as applications to test technologies and methods.