Apple Harvest

This project is a collaboration between CMU and University of California-Davis. The main goal for us in this project is to significantly improve the detection of fruit (apples) to thereby improve picking efficiency and picking cycle of a robotic apple harvesters. This capability will expand robotic operation to wider classes of orchards, with trellised and hedged trees with deeper canopies and less stringent pruning and thinning requirements. It is recognized that fruit visibility is key to fruit detection, and we are investigating state-of-the-art deep learning-based object detection in combination with tree foliage agitation via airflow in a targeted and controlled fashion, and that image sequences – rather than individual frames – and multiple dynamic viewpoints from arm cameras to drastically increase fruit visibility and detection.