This article presents some possibilities for obtaining sensory data from the environment (such as meteorological data, pollution level, night vision, etc.), using unmanned aerial vehicles. Attention is paid to some specific requirements to UAVs used as flying platforms for sensory data acquisition. The process of creating a prototype of a system for collecting and transmitting data in real time from the site of a crisis event using UAVs is analysed. The authors propose to use a specialised neural network tuned to identify half-hidden (half-buried by disaster) people when analysing images received from UAV-borne sensors.