A research team at Monash University’s Faculty of IT is using Raspberry Pi to keep an eye on honeybees pollinating crops at a strawberry farm in Victoria, Australia. Their aim is to develop an affordable, automated pollination monitoring system.
Climate change, pesticide use, and other human behaviours have affected the efficiency of food production. Pollinators play a key role in global food production and ecosystem management, and farms worldwide use insects such as honeybees to pollinate crops such as strawberries, raspberries, and almonds. Monitoring pollinators and managing them effectively boosts crop yields, supporting the viability of farms.
Why use Raspberry Pi?
Current methods used for pollination monitoring are time-consuming and limited in data collection capacity. The team wanted to develop a low-powered, affordable camera system to collect insect movement data across different farm locations. Harsh farm conditions meant the camera setups needed to be highly robust against weather, heat, and chemicals. The system also needed to be easily upgradable and customisable, and compatible with multiple sensors. Raspberry Pi hardware ticked all of these boxes.
How the pollination monitors work
Nine Raspberry Pi-based camera units were mounted on tripods above strawberry vegetation, where they record continuous video of insects over five-hour periods every day. The recordings allow the researchers to track insect movements and monitor their contribution to strawberry pollination.
Each unit comprises a Raspberry Pi Camera Module 2 and a Raspberry Pi 4 Model B. Heatsinks were attached to the Raspberry Pi boards to absorb the heat generated when recording continuous video in a humid outdoor environment. A specially designed case from our approved resellers The Pi Hut, designed to house a Raspberry Pi board along with a Camera Module, keeps everything safe. 20,000 mAh power banks power the system. Videos collected using the system were processed by Polytrack 2.0 outdoor insect tracking software to analyse pollination.
In future research, the team is planning on upgrading the camera system to record data on temperature, humidity, and sunlight using different sensor units available for Raspberry Pi. This will help better understand variations in insect pollination behaviour due to changes in weather and climate. They are also researching using Raspberry Pi to process videos at the edge to make their systems more efficient. Our recently released Raspberry Pi Camera Module 3 Wide could also broaden their horizons by capturing a wider field of view from each pollination monitoring unit.
Meet the team and read more
Dr Malika Nisal Ratnayake conducted this research alongside Chathurika Amarathunga and Asaduz Zaman, supervised by Associate Professors Adrian G Dyer and Alan Dorin, at Sunny Ridge strawberry farm in Victoria, Australia. It was funded by AgriFutures Australia and an Arc Discovery Projects grant.
Read the team’s full research paper: Spatial Monitoring and Insect Behavioural Analysis Using Computer Vision for Precision Pollination. They also published a dataset with over 2300 insect tracks, test videos, and images of four insect types. The code run by the Raspberry Pi and the insect tracking software are available on GitHub.