Autonomous vineyard eco-cyber-physical-system management using computer vision and machine learning for Viticulture

Due to importance of agriculture and crop production regarding to food security, as well as population growth, the use of new technologies is in high priority position. It causes to productivity improvement and mechanization of agriculture alongside decrease of crops injury and labor cost. A great deal of research and experiments have conducted for finding new solutions and practical approaches to mechanization weeds control and herbicide usage reduce. using of Convolutional Neural network and deep learning was used in this study as an attempt to increase the accuracy of grape bunch detection in vineyards. VGG۱۶ standard was utilized in this purpose. The results have shown that learning accuracy for section of grape clusters was ۹۹%. Also, the validity and accuracy of designed machine for grapes' bunch detecting were ۹۸/۴۴%. In addition, compared with other existing algorithms and methods, the performance of suggested method was better than KNN, decision tree and random forest algorithms.

کلیدواژه ها:

Machine vision ، farm and garden management ، convolutional neural network