Orchard harvesting is a very important process，it is tired and boring. If we can use robot to do harvest, life will be better. For robots, there are two challenges to face, one is to identify whether the fruit is mature or not, and the other is to determine the position of the fruit. Using OpenCV's image recognition technology, a machine vision neural network is constructed, the model is trained through the training set of mature fruits and immature fruits image. with dual cameras, we can get position of fruits determined by OpenCV's three-dimensional positioning technology and guide the robot to carry out picking operations. Automatic picking process of the robot will greatly reduce the labor costs and improve the production efficiency of orchards.
Comments