The sun is high in the sky and the thermometer is approaching 35 degrees Celsius. Not a breath of wind. The dry air holds a red dust that is stirred up at regular intervals by jeeps and people. Feverish activity is going on behind Fredrik Gustafsson, who stands in the beating sun on the limitless East Africa savannah.
He is professor of sensor technology at Linköping University, and has come to Fredrik Gustafsson, professor of sensor technology at Linköping University. Photo credit Fredrik GustafssonKenya to test newly developed equipment that it is hoped will contribute to preserving a magnificent richness of species in danger of being lost.
“I was deeply affected by the situation on my first visit to Kenya. The threats against the animals stood in stark contrast with the limited resources to combat poaching. The natural world here is invaluable, but the resources set to preserve it are insufficient. This insight led to the idea of using cost-effective technical solutions to help the park rangers in their work”, says Fredrik Gustafsson over a video link from the Ngulia nature reserve, 300 km southwest of Kenya’s capital, Nairobi.
He has been involved in conservation in Ngulia since 2014. This is a rhinoceros reserve of around a hundred square kilometres that is home to approximately 100 black rhinoceros. In the 1970s, around 10,000 black rhinos were living in the same area. The dramatic reduction is mainly the result of poaching, hunting for the valuable rhinoceros horn.
Billion dollar industry
Ground rhinoceros horn is used in traditional Chinese medicine to “cure” diseases and “give energy”. The demand for it increased dramatically in 2007 when a politician in Vietnam claimed that it had cured him of cancer. There are five different species of rhino and two are found in Africa – white and black rhinoceros. Photo credit Björn PerssonThere is, however, no evidence that this is the case, in particular when you remember that rhinoceros horn mainly consists of keratin – the material of fingernails and hair.
However, despite the unproved claims of medical effects, the demand for rhinoceros horn is the force behind a black market worth billions of dollars, with huge sums to be earned. A single rhinoceros horn can be worth up to SEK 500,000, making it more expensive per kilo than both gold and cocaine.
Poaching operations to get hold of the highly sought-after material are brutal. Many animals bleed to death on the savannah after having been sedated and the horn removed with an axe or motor saw: it is often cut right at the very root to maximise profit. In some cases, the illegal trade in rhinoceros horn finances terrorist activity, which makes the work of the park rangers particularly risk-filled. Many poachers are prepared to kill to get hold of the precious material.
One factor that benefits ruthless poachers is the enormous areas over which the animals move. The Tsavo West national park – of which the Ngulia rhinoceros preserve is a part – is the same size as the Swedish province of Skåne. Only around 100 park rangers work here with the task of monitoring the well-being of the animals 24 hours a day. It involves a lot of manual work and long working hours.
“The park rangers must be able to monitor the animals’ well-being on the savannah. At the moment, a great deal of work is needed from far too few rangers to watch over the animals. And the job is full of risk. If we can automate parts of the surveillance, the rangers can work more effectively and more safely”, says Fredrik Gustafsson.
The project A mobile app is used by the park rangers to report anything unexpected. Photo credit Fredrik Gustafssonled by Fredrik Gustafsson has developed a mobile app that has been used for the past five years by the park rangers in Ngulia. They use it to report anything unexpected such as unknown fireplaces, footprints and other signs of people, and anything else that affects the conservation. The app is connected to a database in which information from all park rangers is collected. This makes it possible for the central command centre to follow everything that happens in the park in real time on a digital web-based application, the Ngulia Dashboard.
The system has been developed taking the work situation of the park rangers into consideration. Paul (not his real name) has worked as park ranger in Ngulia for nearly 7 years, and can confirms that surveillance has become much less demanding with the aid of the technology.
“We can use the app to see where we patrolled earlier in the month, and can thus always choose new roads to sweep across our area much more efficiently than before. And if we see a poacher, we can report this immediately, and be sure that our commander receives the information and can send reinforcements”, says Paul, who cannot use his real name or appear in a photograph for reasons of security.
Machine learning and AI
One piece of equipment the researchers are testing in Ngulia is an intelligent game camera. These cameras are deployed on the savannah at locations often visited by animals, such as waterholes. The cameras are activated by motion, and connected wirelessly to a server. Software in the camera performs an initial analysis of the image. If there is the possibility that the image contains people or any one of “the big five”, it is sent to the server for more advanced software analysis, before being presented in Dashboard as a report.
Sara Olsson and Amanda Tydén are authors of the software, having developed a recognition algorithm for their degree project for the Media Technology and Engineering master’s programme at Linköping University.
Sara Olsson, former student at the Media Technology and Engineering master’s programme at Linköping University. Photo credit Emilia Larsson“We used machine learning to enable the algorithm to identify the different animals and distinguish them from humans. We collected many images of animals such as rhinos, elephants and large cats to use as training material. The current version of our model can distinguish between seven animal species and humans, and can follow the motion of an individual in a sequence of images to determine observations of a unique animal”, says Sara Olsson.
The software has been tested on rhinoceros in the Kolmården Wildlife Park outside Norrköping, Sweden, but Fredrik Gustafsson points out that the conditions here are, despite everything, quite different. The technology has progressed well, he says, but many problems remain before it is completely reliable in the field.
“Above all, the technology must be robust. Further, it must be easy to install and we are currently facing several practical challenges with this aspect. And then there’s an incredible amount of local bureaucracy to deal with. Everything we do must be approved at different levels of the Kenya Wildlife Service, which is the Kenyan government agency with responsibility for all the national parks. It’s a real challenge”, says Fredrik Gustafsson.
In the next step, sensors developed by Fredrik Gustafsson and his research group are to be connected to Ngulia Dashboard and used to monitor the rhinos. The sensor is a smooth cylinder, approximately one centimetre in diameter and three centimetres in length. The intention is to insert the sensor into the animal’s horn to be able to track its activity and well-being.
This idea is now being tested on two of the rhinoceros at the Kolmården Wildlife Park to see how it functions. The cows that graze in the Tinnerö oak forest outside Linköping will also be fitted with a version of the sensor such that visitors to the nature reserve can follow the animals in real time and obtain more information about the various individuals. The experiment in Tinnerö is part of a collaboration with the Agtech2030 Vinnova initiative. The idea is that analysis of animal behaviour can improve the well-being of both Swedish grazing animals and animals around the world threatened with extinction.
The dust has settled somewhat, and it will soon be time for lunch in the beating sunshine. Before ending the call, Fredrik Gustafsson says that there has been considerable interest from the Kenyan government agencies, and many other Kenyan parks intend to adopt the model used in Ngulia.
“The concept has been successful, and many other parks in Kenya are interested in the same type of system. This will, however, require large-scale production of the equipment, which in turn requires financing. This is why we have created the ‘Smart savannahs’ research project, with the objective of obtaining funding such that we can scale up the system to all other parks in Kenya that have rhinoceros”, says Fredrik Gustafsson.
The article har been published in LiU Magazine nr 2/2021.
Photo credit Björn Persson