University of Twente Student Theses


Designing a smart nest box for automated biodiversity monitoring

Bianchi, T.J. (2022) Designing a smart nest box for automated biodiversity monitoring.

[img] PDF
Abstract:Biodiversity monitoring is important as it provides guidelines for decisions on how to manage biological diversity in terms of production and conservation [1]. Monitoring birds can be challenging as they move quickly and nest in enclosed nest boxes. Monitoring nest boxes are nest boxes that have a camera installed which allows the nest box to be monitored. These systems have been used for many years in research and make biodiversity monitoring in this field accessible. The downside of the current monitoring nest boxes is the way the data is collected and analyzed. The systems capture long videos or many pictures, which currently are analyzed manually by experts. The experts go through hours of footage and annotate the data they are interested in, such as bird species or prey which is brought into the nest. This is a tedious process which takes much time. A proposal is made to use Artificial Intelligence (AI) to analyze the data. The AI requires high-quality data from the nest box. By adding a trigger system and high-quality camera in the nest box, the amount of data is significantly reduced and the image quality is greatly improved. This project determines what the optimal smart nest box is and the design process for a prototype is described. The requirements and necessary components are reviewed. A prototype is made, which consists of a Raspberry Pi, Raspberry Pi High Quality Camera, IR beam sensor and a LED strip. The prototype is designed so that it can easily be installed on standardized nest boxes. These nest boxes are designed for great tit and blue tits to nest in and used for research. From the initial experiment, it can be said that the system can sense incoming objects through the nest box entrance with high precision and recall. A 3D printed bird was inserted in the nest box 50 times and sensed all 50 times by the system resulting in a 100% value for both recall and precision. The quality of the recorded images that get taken is not consistent. The 3D-printed bird was manually inserted in the nest box 50 times with a fast entry, which resulted in 64% of the images being sharp and the other pictures being blurry. Overall a good first prototype of a smart nest box is designed in such a way that it is highly configurable. A wide range of sensors can be connected to the Raspberry Pi and programmed using python. Next to this, the camera angle, brightness of the LED strip and behavior of the IR beam sensor can be changed easily for different setups.
Item Type:Essay (Bachelor)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:42 biology, 50 technical science in general, 53 electrotechnology
Programme:Creative Technology BSc (50447)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page