It is essential that Australian growers have access to pulse varieties that combine a suite of desirable traits including, increased grain yield, improved agronomic characteristics and optimal quality traits. These traits ensure the viability of the industry and maximise returns to growers on their on-farm investment. Quality traits are specific for each export market and determine grain prices. For pulses, the traits are predominantly defined by physical seed characteristics such as seed size, colour and lack of defects, but recently seed composition has also become important as pulses are used as a source of plant-based proteins. The application of sensor technologies, such as multispectral sensing, is pertinent for measuring these traits as a range of seed characteristics can be derived from seed images. It is important that plant breeding programs have access to a suite of high throughput technologies to identify desirable germplasm which will result in commercial cultivars which are competitive in the export market.

This project will implement multispectral image-analysis to develop a suite of algorithms to determine physical seed traits and develop models to quantify seed damage, due to disease and abiotic effects which impact on marketable quality. The data will be used to evaluate lentil, field pea, chickpea and faba beans. Additionally, this project aims to develop near infrared spectroscopy (NIR) and NIR-hyperspectral applications as there is increasing demand to source plant-based protein from pulses. While these algorithms are initially developed for plant breeding and research purposes, it is anticipated that the technology may be adopted by the broader grains industry at the point of receival to quantitate grain quality.