| The automatic separation of defective eggs from qualified ones is a fundamental issue at the poultry industry for both economical and health reasons. At a poultry farm, after the eggs are laid, they must be cleaned, sorted by quality and size and packaged into cartons for marketing. Eggs with cracks and leaker are a major issue as they degrade all the mechanical parts while processing on the graders and packers, but at the small ones which are present in most poultry farms. So it is very important to separate the highly defective eggs from qualified ones prior to its packing and shipment to the scarce and high-cost graders/packers which deal with poultry distribution over large regions.Flash candling is a quality sorting operation which involves searching for several defects: cracks, blood spots, dirt spots, and undergrades, any one of this downgrade the egg quality. In a conversional flash candling operation, eggs are illuminated against a dark background. Cracks appear as bright lines, blood spots as red regions, dirt as dark spots, and yolks as gray shadows. Up to 150000 eggs per hour can pass before the eyes of a candler. The inspectors are subject to eye fatigue and a demand for continuous close attention. As a result 10 to 20% of the cracked eggs are usually missed. Computer vision system offers the potential to minimize misclassifications and improve consistency in candling eggs.Some research has been performed in eggs when the pursued defect had a unique dominant color. Thus to differentiate blood spots, stain and cracks some studies have been undertaken by training neural nets with color histograms with a good performance on one defects. The average accuracy of the neural networks, which was trained all defects eggs was less than the accuracy obtained when the data were restricted to a single... |
