On-line measurement of the physical properties of each cane consignment at the factory

Grower deliveries to a mill comprise billets of the cane stalk, leaves, tops and roots, with everything else classified as extraneous matter (EM) that doesn't contribute to the quantity of sugar produced and can negatively impact the milling process and output. The physical properties of the cane supply are currently not measured (except intermittently at one mill) and thus parameters such as EM content and billet dimensions are not routinely monitored, therefore are unmanaged. By measuring the properties of each load, milling costs associated with high EM (increased sucrose losses in bagasse, mud and molasses, chokes in milling, reduced liquor purity) and reduced factory throughput, which leads to extended season lengths, could be avoided. Growers could also use the information to better suit varieties to different regions and conditions, along with improved harvesting outcomes. This project will investigate the use of computer vision and machine learning at mills to measure various parameters in the harvested cane loads brought from farms to the sugar mill.