Machine vision in experimental poultry inspection

Abstract

Describes experimental work in using machine vision for inspection and grading in the poultry industry. Outlines the main features of human poultry inspection and grading by size, colour and conditions and describes how neural networks can detect various defects in the poultry [such as bruising, cuts and broken bones]. Explains the development of neural nets, the use of processing elements and HSI histograms for image analysis. Describes texture analysis and the use of two dimensional Gabor filters. Concludes that neural nets appear to be a good tool for detecting poultry defects but further work is needed to develop more specific input.

Citation

Pellerin, C. (1995), "Machine vision in experimental poultry inspection", Sensor Review, Vol. 15 No. 4, pp. 23-24. https://doi.org/10.1108/02602289510102345