Design and development of fruit picking robot based on deep learning and binocular ranging.
This project combines convolutional neural networks, binocular ranging and embedded systems. The ultimate goal is to quickly and accurately identify fruits locally and control robotic arm picking. First, the U-Net network is used to train a model that can distinguish fruits and backgrounds. The model is mounted on the robotic arm through an embedded system. When the binocular camera identifies the fruit to be picked, the distance between the robotic arm and the fruit is measured through the ranging algorithm, and the robotic arm is controlled to pick.