Résumé: Recently, the plantation of sugarcane is done in mechanized manner with billets that are short segments of sugarcane harvested. The automated harvesting procedure can destroy billets and reduces quality of billets. Deep convolution neural network is applied to diagnose sugarcane billet damage (DCNN). A novel technique, namely Aquila Sailfish Optimizer (ASO) algorithm is devised for weight update of neurons in training the DCNN and enhances the efficiency of DCNN. The ASO is obtained by incorporating Aquila Optimizer (AO) and Sailfish Optimizer (SFO). The classification of sugarcane billet damage is done by Chronological Aquila Sailfish Optimizer (CASO) algorithm trained Deep Quantum Neural Network (DQNN), which is trained with obtained by incorporating Chronological concept in ASO. Here, the sugarcane billet damage will be classified into five types, including Crushed (blue), Cracked (red), No buds (yellow), Two buds (green) and Single damaged bud (orange). The developed CASO-based DQNN presented highest precision of 91%, recall of 93.3%, F-measure of 92.1%, and accuracy of 91.5%.

Langue: Anglais