Statistical Data Mining Algorithms for Optimising Analysis of Spectroscopic Data from On-line NIR Mill Systems: Improving System Calibrations for Quality Measures (Old ID 22492)

Sugar Research Australia
Role

Principal Investigator

Description

NIR spectroscopy is a rapid, non-invasive method for determining cane quality attributes (e.g. CCS, brix, fibre and biomass). NIR methods for sugarcane are advanced and work well for 90% of cases. When NIR methods fail, industry must resort to expensive laboratory analysis. This project will involve using novel statistical data mining techniques applied to large NIR databases to investigate improved calibrations for cane quality measures. Specifically this project is interested cases that were previously difficult to calibrate. Improved calibrations will benefit industry through reduced costs associated with extensive laboratory analysis for samples unsuited for standard industry calibrations.

Date

01 Jan 2016 - 30 Jun 2019

Project Type

CONTRACT_RESEARCH

Keywords

NIR;Spectroscopy;Statistics;Sugarcane;Data Mining;Calibration

Funding Body

Sugar Research Australia

Amount

143000

Project Team

Ronald White;Yvette Everingham