Enhancing data collection procedures for non-destructive chicken egg fertility determination using NIR hyperspectral imaging

The hatchery industries are faced with huge economic losses in millions of dollars, resulting from incubating nonfertile eggs that will never become chickens. There is therefore an urgent need for non-destructive techniques to predict the fertility chicken eggs early enough (especially prior to incubation). The project seeks to solve the identified problem via optimizing modelling parameters and performances of new and existing egg models using state-of-the-art hyperspectral imaging technology in conjunction with pattern recognition and multivariate analyses. This study will eventually come out with deployable chicken egg fertility classification models using adequate amount of light source illumination, optimised egg orientation position and appropriate targeted modelling features. Modelling results will eventually be ready for deployment in an industrial online multispectral imaging classification system for chicken eggs.

Adeyemi Olutoyin Adegbenjo
Faculty Supervisor: 
Shiv Prasher
Partner University: