Riconoscimento neuronale di prodotti alimentari tramite X-Ray Imaging
Alessandro Celona - Università degli Studi di Messina - [2004-05]
A New Parametric Scheme for X-Ray Imaging Food Classification: quality assurance has a great relevance in the industrial food production chain. One of the most widespread techniques, used for non-standard product assessment, is based on X-Ray imaging. This work introduces a new multi-spectral technique for identification of food product content. A false colour image is generated from three independent acquisitions at different X-Ray energies and the correspondent spectrum shape in the HSI color-space is classified by a neural network. The new method is presented together with preliminary results for three types of packaged foods, showing a good discrimination power of the chosen parametrization.
Questa tesi è correlata alla categoria