Convolutional Neural Networks and Transfer Learning for Quality Inspection of Different Sugarcane Varieties
Moisés Alencastre-Miranda, Richard M. Johnson, Hermano Igo Krebs
Abstract
In this article, we employed computer vision and deep learning techniques to select and plant healthy billets, which increased plant population and the yield per hectare of sugarcane planting. We employed well-known convolutional neural network architectures to process large image data sets and transfer learning techniques to expand the results to different sugarcane varieties. It would be very time consuming to collect and label large data sets for each sugarcane variety, for which quality inspection is needed, prior to planting. We used a two-step transfer learning process to extend the trained architecture to new varieties. We compared results obtained during transfer learning using AlexNet, VGG-16, GoogLeNet, and ResNet101 architectures to classical computer vision methods. Our goal was to determine the best approach to detect damaged and good billets in the shortest processing time. Best results in both time and accuracy were obtained with AlexNet. For AlexNet, we compared the permutations of three sugarcane varieties in order to find the best model to identify the healthy sugarcane billets. We then reduced the number of images employed to retrain the model to determine the tradeoff between time and performance. Ultimately, one needs only a few dozen billets of the new variety to retrain the network. Our approach led to meaningful increments in the yield per hectare ranging from 33 to 80% depending on sugarcane variety.