MULTIPLE FEATURE ANALYSIS FOR MACHINE VISION GRADING OF FLORIDA CITRUS

Both blemish and physical attributes were acquired on commercially graded Florida grapefruit, orange, and tangerine varieties. Using equal numbers of acceptable and rejected fruit, various neural network classification strategies were applied to blemish–related features and blemish plus physical features. The blemish plus physical feature neural net models were the most successful, yielding overall correct classification levels of 98.5% for grapefruit and orange and 98.3% for tangerine. No significant difference was found between the neural net models of standard back–propagation, jump step, or variable transfer functions for the hidden layer.