Detection of Animal Occurrence Using an Unmanned System

• Main Authors: Lešetický J., Matějka P., Olmr M.
• Format: Article
• Language: English
• Published: Sciendo 2019-09-01
• Series: Scientia Agriculturae Bohemica
• Subjects: unmanned aerial vehicle; harvest; animal detection; animal monitoring; image analysis
• Online Access: https://doi.org/10.2478/sab-2019-0028

In recent decades, there has been an increase in the work speed and breadth of agricultural technology used to mow grasses. This modernization has resulted in a decline in wildlife. There are several conventional ways to prevent these losses. The most well-known and simplest technique is to search for wild animals using dogs and a phalanx. The dogs are trained to systematically search the area and drive the animals out. Efficiency is increased when visiting a site regularly, thus disturbing the animals, which are then consequently less likely to fawn. The effectiveness of the swarm line depends on the number of participants involved. The recommended spacing is set at 1–3 m. An effective modern means seems to be the use of an unmanned system and thermal cameras. This article presents a proof of concept of a detection system that is capable of detecting the object searched for in grassy vegetation with more than 96% success, regardless of the flight level. The study contributes to automated detection based on the basic principles of threshold.