Програмний комплекс розпізнавання людини на відеозображеннях з безпілотних літальних апаратів для пошуково-рятувальних операцій
The effectiveness of search and rescue operations, regardless of the nature of the emergency (earthquake, avalanche, loss of orientation of the tourist on the ground, the consequences of hostilities, etc.) is largely determined by the time elapsing until the casualty is identified.Existing methods of inspection of territories, based mainly on manual combing of the terrain and visual analysis of video materials, have limited effectiveness in conditions of difficult terrain, reduced visibility or a large search area, which determines the relevance of the development of automated means of detecting people.
The purpose of the project is to develop a software package for human recognition on video and photo images obtained from an unmanned aerial vehicle, based on modern computer vision algorithms.
The software package provides processing of video stream and images obtained during aerial reconnaissance of a given territory, and automated recognition of signs of human presence in conditions of partial visibility, natural and man-made obstacles.The results of the recognition, including the detection coordinates and relevant personnel, are transmitted to the rescue units in real time by means of telecommunications to make operational decisions.
The proposed approach allows to significantly reduce the time required to identify the casualty, reduce the load on the personnel of rescue services and increase the safety of search operations due to the remote nature of the primary reconnaissance.
The developed complex has an interdisciplinary nature of application and can be used within various types of rescue operations, in particular: elimination of the consequences of earthquakes and man-made disasters, rescue operations in ski resorts in conditions of avalanche danger, search for people lost in the open area (in a forest, mountain or water zone), as well as search for victims in the affected areas as a result of hostilities.
Students and graduate students of the department take part in the implementation of the project.