Lyon, Tristan Dean (2024) Evaluation of a low-cost machine vision detection method for Varroa Destructor mites on honeybees. [USQ Project]
![[img]](https://sear.unisq.edu.au/style/images/fileicons/text.png) |
Text (Project – redacted)
Lyon_T_Dissertation_redacted.pdf Download (7MB) |
Abstract
Varroa Destructor mites are an incredibly destructive parasitic mite that feeds upon honeybees, and can cause the collapse of beehives in high infestation levels(Plant Health Australia nd.-a). These mites are a relatively new pest to Australia, having only arrived in NSW in 2022 (Plant Health Australia nd.-a) ending the country’s status as one of the last countries free from the pest. The introduction of this pest has the potential for enormous economic impacts of up to $70 million per year (Australian Government Department of Agriculture 2024) in addition to environmental impacts from the loss of pollination provided by feral bees.
While there are many studies around the world, including some work in Australia (Wheeler 2021a) for the detection of Varroa mites, most utilise expensive industrial cameras or rely on uploading images for remote processing. These methods present problems for widespread deployment in Australia due to the cost and lack of data reception in remote locations. This research aims to investigate a low-cost, machine vision system for detecting Varroa mites in hives that can work remotely in Australian weather.
For this research a system was built utilising a camera that is sensitive to both visible and infra-red light based on previous research by Bjerge et al. (2019a). The camera was connected to a Raspberry Pi 4 for local image capture and processing, with the system powered by a battery charged by a solar panel for remote operation. This system would capture images of bees entering the hive in an area of NSW with Varroa mites, and these images would be used to develop the machine vision algorithm for detecting the mites.
From the resulting images it was discovered that there were problems with the developed system, including camera focus issues and excessive heat in the processor. The result of this was that the collected images were not of sufficient quality for an algorithm to be developed for detection of mites during the timeframe of the project. Prior research indicated that good discrimination between bees and mites should be expected with the illumination wavelengths in use, however that was not seen. This is most likely because of the cheaper single sensor camera being used in this research. Changes to the wavelengths of illumination being used may be necessary to achieve the expected discrimination, if it is at all possible with the camera used.
|
Statistics for this ePrint Item |
| Item Type: | USQ Project |
|---|---|
| Item Status: | Live Archive |
| Faculty/School / Institute/Centre: | Current – Faculty of Health, Engineering and Sciences - School of Engineering (1 Jan 2022 -) |
| Supervisors: | Low, Tobias |
| Qualification: | Bachelor of Engineering (Honours) (Mechatronics) |
| Date Deposited: | 16 Mar 2026 05:39 |
| Last Modified: | 16 Mar 2026 05:39 |
| Uncontrolled Keywords: | Varroa Destructor mites; honey bees; machine vision system |
| URI: | https://sear.unisq.edu.au/id/eprint/53136 |
Actions (login required)
 |
Archive Repository Staff Only |
