Ground-nesting birds, fawns and young hares hide in tall grass to protect themselves from predators and other threats. However, this hiding place becomes dangerous, when farmers mow the grass. Thousands of animals are killed when grass is harvested with mowing machines. To deal with this issue, a number of German companies, research centers, and universities have collaborated to develop practical wildlife detection methods, which can be used by farmers to save the lives of the animals and prevent undue delays.
Thermal Imaging Core Platform
One possible method is to use an unmanned aerial vehicle (UAV) equipped with a thermal imaging core to check fields and meadows from the air prior to starting the mowing machines. Roe deer iare known to have the highest fatality rate. A large number of Roe deer fawns are born in May and June months and they prefer to stay in high grass during their initial months of life. However, this is the period when most European farmers begin their mowing operations. When a mower approaches, the fawn’s reaction to this threat is to remain motionless. This instinctual behavior and timing is a deadly combination. It has been estimated that the fatality rate for newborn fawns is 25%, which means one out of four fawns is killed by the mowers.
Martin Israel, researcher at the Remote Sensing Technology Institute of the German Aerospace Center (DLR), informed that apart from being an ethical issue, the harvested grass is often meant for animal fodder. If the remains of a fawn contaminate the fodder, the carcass can be infected with botulism bacteria. There have been a number of cases in the past where livestock has died due to such toxic fodder, which resulted in huge economic loss for the farmers.
Figure 1. The fawns are well camouflaged, making them difficult to detect visually, but they show up very clearly in the thermal image.
Figure 2. This thermal image shows a fawn that was detected with the FLIR Tau thermal imaging core from an altitude of 50 meters. At that altitude a fawn is represented by just 40 of the 327680 pixels.
Israel added that it is difficult to detect fawns hidden in high grass, as the fawns are well camouflaged. However, when a thermal image is used these animals can be detected clearly (Figures 1 and 2). A newborn fawn has a thin coat, which means that the animal’s body does not insulate well and produces plenty of infrared radiation. This makes it easy to detect the animal using thermal imaging.
Bird’s Eye View
According to Israel, a handheld thermal imaging camera does not provide a suitable solution because one needs to be close to the animal to identify it. If one stands further away, the high grass will block the animal, even when using a thermal imaging camera. This makes the entire exercise quite labor intensive when inspecting large fields. After extensive investigations, it was concluded that a top down view would be a perfect vantage point to detect fawns.
As part of the Wildlife Finder project, a collaborative team of the German Aerospace Center (DLR), ISA Industrieelektronik GmbH, agricultural machinery manufacturer Claas Saulgau GmbH, the University ofHohenheim Israel, and the Technical University Munich analyzed the use of an UAV as a thermal imaging core platform. They used Ascending Technologies’ Falcon-8 electrical octocopter and the FLIR Tau thermal imaging core (Figures 3 and 4), which is compact, lightweight, consumes minimum power, and provides high resolution thermal images, making it suitable for this application.
Figure 3. The FLIR Tau thermal imaging core
Figure 4. The best viewing angle for fawn detection is straight down, so the researchers mounted a FLIR Tau 640 thermal imaging core on a UAV.
FLIR Tau Thermal Imaging Core
An uncooled Vanadium Oxide (VOx) focal plane array (FPA) detector in the FLIR Tau thermal imaging core creates sharp thermal images with a resolution of 640 x 512 pixels. The FLIR Tau can view temperature differences down to 50mK (0.05°C). It is compact and measures just 44.5 x 44.5 x 30.0mm in size, and weighs 72g. Its low power consumption means that it has negligible effect on the UAV’s flight time.
Automatic Fawn Detection
The onboard computer utilizes analysis algorithms to ascertain whether the image has the thermal signature of a fawn. To ensure this automatic detection, comparison is made between the thermal signature of the fawn and the thermal signature database. The researchers are presently extending that database to enhance the automatic detection algorithms and thus make the automatic detection more precise.
Whenever the software detects something that looks like a fawn, the location is stored and utilized for a second verification flight. Israel informed that the first flight takes place at an altitude of 50m, but at this altitude the automatic detection of fawns is not quite precise, so the octocopter is sent out for a second flight at an altitude of 30m so that the finds of the first flight can be verified, which further improves the detection accuracy.
A large number of animals, including adult deer, fawns, and rabbits, have been rescued using a single UAV, all in a matter of seven hours of flight time, spread across 15 days. Thus, the thermal imaging wildlife detection system mounted on a UAV offers a promising future, as it is not only fast and user-friendly, but also reliable. The commercial version of this system is just a matter of time, concluded Israe.
This information has been sourced, reviewed and adapted from materials provided by FLIR Cores and Components Group.
For more information on this source, please visit FLIR Cores and Components Group.