Scientists from the University of Wrocław will be saving lives
Professor Tomasz Niedzielski and his team from the Department of Geoinformatics and Cartography at the University of Wrocław have developed a system to locate people lost in open areas. In just two years, they plan to have it working to save lives in search and rescue missions.
The cutting-edge combination of geographical knowledge, programming skills and drones have given rise to a new algorithm, the task of which can be compared to looking for mushrooms from the sky. This algorithm can detect people from aerial photos, in which a human is only a clump of pixels, indistinguishable from animals and bushes.
“The direct users of our system are going to be mountain and sea rescue units, as well as the police and border guard. But it is a social project, from which we are all going to benefit,” said Prof. Niedzielski.
How does it work? The first step comprises determining the potential area in which the missing person is expected to be found. Then, a drone is sent over the area to make hundreds of aerial photographs and carefully observe it.
The second step is to send the pictures for analysis, using the new algorithm which has been specially designed to detect people. This is the biggest challenge for the project, because on one hand the technology has to be adjusted for precise detection of people, and on the other it has to properly automate the operation and create the conditions for processing very large sets of data in real time, in order to provide the rescue team with the location of the missing person as soon as possible.
Niedzielski’s team is working with the Jura Group of the Mountain Volunteer Search and Rescue (Górskie Ochotnicze Pogotowe Ratunkowe (GOPR)). GOPR has 150 full-time rescuers, who together with volunteers can secure an area from the Karkonosze mountains all the way to the Bieszczady mountains. This constitutes 15% of the country’s land area (per the Polish Statistical Yearbook and GOPR’s official website). It takes several hours just to organise a rescue team and gear to search such a large area. What can happen to a missing person during this time? Freezing temperatures, wind, darkness, sicknesses and injuries are just some of the conditions they must face.
It is not the first project in which Niedzielski has used drones. The professor’s first encounter with them was in 2007, during workshops for young scientists in Montpellier. He decided then that he would want to engage in high-resolution ground observation from low altitudes, if the technology allowed him to. After five years he achieved this goal, by which time he was already Head of the Cartography Department (now the Department of Geoinformatics and Cartography).
Professor Niedzielski used drones, amongst other things, to research the condition of various bodies of water and draw maps of snow cover. Later, he became interested in the problem of detecting people from a drone’s perspective. He flew up a drone and recorded images of an area of land in which the research team was standing, and then studied the resulting pictures. Could a person be identified from within them? Would it be possible to create a program to do that automatically? To find out, Niedzielski applied to the Iuventus Plus programme. He received a science grant from the programme and in 2015 began his research into detecting missing people in uninhabited areas.
The research team included the creators of a human-detecting algorithm, experts on geoinformation, as well as technicians and technical crew.
The first stage of research was conducted on a small scale. It included dozens of hours in the seclusion of a computer laboratory, developing the algorithm. This may look like the regular work of an IT company, but the main idea was not only to code the algorithm, but include geographical knowledge within it.
The second stage included large-scale field experiments. Their goal was to test the solution under various environmental conditions – at different hours and in different seasons, on flat and varying terrain, and in searching for both adults and children. The tests showed the good repeatability and high efficiency of the algorithm, and the results were published in widely-recognised International science magazines, including Geomatics, Natural Hazards and Risk, the Journal of Field Robotics and Pure and Applied Geophysics, and were very well received by the scientific community.
Usually, everything would end at this stage. The basic research was successfully conducted by scientists, the research question answered and the results published. No wonder the creators of the algorithm began looking for new challenges and new sources of funds to further develop the system.
It was then that experts in the field gave a clear signal that the team should continue the project. This was because they saw the opportunity the new system represented for saving people’s lives, the rescue services’ desire to actually use the system, and the fact that the purchase of a drone for a couple of thousand zloty would soon be within reach of even the smallest rescue units. The new system was very well received by, amongst others, the GOPR Jura Team, the Anna Pasek Foundation and FlyTech, the leading Polish manufacturer of professional drones for geodetic surveying, farming and energetics.
“The opinion of practitioners was particularly important to me, as a scientist. I could predict the enthusiastic reaction of the scientific reviewers, but the opinion of the professionals in the rescue services, new technologies and business was still a mystery to me. It turned out that the scientists and the practitioners had completely different perspectives on our project. The first group was mainly interested in the research methodology used, while the latter were mainly considering its possible uses in the field conditions and its marketing potential,” said Prof. Niedzielski.
However, the continuation of the project is beyond the usual scope of the team’s work, and means that the professor and the other creators, Mirosława Jurecka and Dr. Bartłomiej Miziński, have had to take on new roles. Both are students of Professor Niedzielski and are very active, young people. Even though they have already published very well-received scientific papers, they do not stop here and make every effort to influence the real world. Niedzielski said it reminds him of something his own professor said when he was still a young scientist – that at first what counts is how much you publish and where you publish it, but ultimately what matters is what you do.
“I’m sure I won’t stop my scientific work and publishing, but I want it to impact the lives of people outside the university. I would like to show them that what I do can be useful to them,” said Prof. Niedzielski.
At the moment the system is at TRL level 4. The Technology Readiness Scale is a 9-level scale describing the maturity of innovative technology. In order for the new algorithm to be usable in actual searches, the research needs to be continued and the project further tailored to the working conditions of rescue teams. Professor Niedzielski predicts that this could take the next two years.
The team therefore decided to establish the SARUAV Limited Company, under a patronage contract with the University of Wrocław, and commercialise their system. The Epic Alfa Limited Company investment fund then became interested in the Company. Its own main investor is PGE Ventures, from the PGE Capital Group, the largest Polish company in the energy sector. The fund invested in SARUAV within the BRIdge Alfa programme, organised by the National Centre for Research and Development. On the University’s part, the commercialisation was supported by the Innovation and Knowledge Transfer Centre, and the Office of Legal Advisers.
Establishment of the company required specifying details about their intellectual property, that is the know-how of the system’s creators. This was necessary in order for the license conditions to be determined. “At first I didn’t expect that establishing a company would require such vast legal analyses. It cost us a lot of time and effort, but regardless of that I think that the decision to commercialise the technology was the right one to make. If I had to make it again, I would definitely do the same thing”, said Prof. Niedzielski.
SARUAV officially launched at the end of 2018 under the name SARUAV, Search and Rescue Unmanned Aerial Vehicle, and was the first spin-off technology company operating under the patronage of the University of Wrocław. Its creators hope that the way they paved will be followed by other scientists interested in commercialising their results.
The purpose of the R&D works conducted by the Company is to create a demo version of the SARUAV system, which can be used by a client with any drone equipped with a camera. As designed, the system will inform the rescue team about potential locations of missing persons on open terrain, thus increasing the probability of finding them and reducing search time. Time is of the essence in determining the survival chances of lost and missing persons. The SARUAV team is aware that their new system can’t solve all of the problems of rescue teams, but it will definitely serve them well in their work.
AUTHOR: Aleksandra Draus
Tomasz Niedzielski, Chairman of the Board at the SARUAV Limited Company, mathematician and geographer. Received his PhD in Geodesy and Cartography, and also holds a PhD in Earth Science in Geography. He is also a doctor habilitatus of Earth Science in Geography and a professor of Earth Science. Tomasz Niedzielski is an expert in modelling and predicting hydrology processes. He is currently working on processing systems for data gathered by drones, particularly for the purposes of automated detection of missing people. Apart from the University of Wrocław, he has worked at the University of Aberdeen in Great Britain, the Polish Space Research Centre, and the Wrocław Centre for Networking and Supercomputing. He has managed multiple research projects, financed from both Polish and foreign sources. Prof. Niedzielski is the author and co-author of several dozen publications in prestigious international scientific magazines, a European patent application and over 200 conference and seminar presentations.
Mirosława Jurecka, Member of the Board at SARUAV Limited Company, geographer specialising in physical geography, GIS and cartography. Currently working on her PhD dissertation on the problems of using geoinformatic solutions in searching for missing people. She specialises in mathematical modelling of the behaviour of missing people in open terrain, based on the circle and mobility models. She also works on analysis of aerial photos taken by drones, particularly on methods of processing images for the purposes of detecting people. She has several years of professional experience in geoinformatics. Mirosława was a scholar of the Foundation for Polish science and an executor of one of the programs for the Ministry of Science and Higher Education, and is the co-author of several publications in prestigious international scientific magazines.
Bartłomiej Miziński, Member of the Board at SARUAV Limited Company, mathematician and informatician. He has a scientific degree in Earth Science in Geography, obtained for his PhD dissertation on designing automatic modelling systems in hydroinformatics. His interests include methods of modelling and prediction of time series, stochastic systems, statistics and image processing techniques. He worked as a researcher at the University of Wrocław and at the Wrocław Centre for Networking and Supercomputing. He has executed four scientific projects and is a scholar of the Foundation for Polish Science. Bartłomiej also received an award from the Polish Mathematical Society. He is the co-author of several internationally-recognised publications in prestigious magazines, and of a European patent application.