Cheryl McCarthy, University of Southern Queensland, National Centre for Engineering in Agriculture, Australia: Practical Application of Machine Vision in Australian Agricultural Research at NCEA
Machine vision has the potential to automate agricultural tasks that are typically performed by human visual inspection. The practical machine vision system for agriculture consists of an imaging sensor, processor, and robotic output, in a mechanical structure tailored to perform an automation task. Image analysis is simplified where possible by lighting, waveband filtering, and multiple sensors. NCEA has demonstrated a versatile range of applications, where the machine vision system can be a portable handheld device, towed on a tractor, or fixed to the side of a shed. Example applications of low-cost machine vision systems by NCEA that will be showcased in the webinar are:
- In-crop machine vision-based weed identification, which has potential to reduce herbicide usage by precisely identifying and spraying weeds from amongst the crop. Low-cost machine vision sensors have the ability to achieve discrimination between weed and crop using shape, colour, texture and height information.
- Automated surveillance of bait boxes at Australian ports, which are likely entry points for disease-carrying bees from cargo boats. A smartphone has been used as a low-cost machine vision sensor inside the bait box and the system has potential use in a network of remotely-monitored bait boxes at ports around Australia.
- Fodder quality assessment, cattle condition scoring, and animal species recognition.
Dr. Cheryl McCarthy is a mechatronic research engineer with the University of Southern Queensland's National Centre for Engineering in Agriculture. Cheryl's research involves developing machine vision and sensing systems to automate tasks in agriculture that are traditionally manual and labour intensive. Currently her projects include precision sensing of weeds for the cotton and pyrethrum industries and remote beehive monitoring for improved bee pest surveillance. Cheryl's Ph.D. work focused on the development of a machine vision sensing system for measuring cotton plant internode length, for potential use as a plant stress indicator for precision irrigation. Cheryl's interests include the application of mechatronic engineering to agriculture, imaging spectrometry, and aerial photography.