Wouter Bac, Wageningen University and Research Centre - Business Unit Greenhouse Horticulture, The Netherlands: Robust Pixel-Based Classification of Obstacles for Robotic Harvesting of Sweet-Pepper
This seminar focuses on detection of sweet-pepper plant parts to construct an obstacle map to plan collision-free motion for a harvesting manipulator. Objectives of the work were to segment vegetation from the background; to segment non-vegetation objects; to construct a classifier robust to variation among scenes; and to classify vegetation primarily into soft (top of a leaf, bottom of leaf and petiole) and hard obstacles (stem and fruit) and secondarily into five plant parts: stem, top of a leaf, bottom of a leaf, fruit and petiole. A multi-spectral system with artificial lighting was developed to mitigate disturbances caused by natural lighting conditions. A new performance measure was introduced for CART pruning and feature selection. Use of the performance measure rendered the classifier more robust to variation among scenes because standard deviation among scenes reduced 59% for hard obstacles and 43% for soft obstacles compared with balanced accuracy.
The research performed is part of the CROPS project (www.crops-robots.eu) funded by the European Commission. The talk is built around an article, with the same title, that will appear in Computers and Electronics in Agriculture.
Wouter Bac is a PhD candidate at Wageningen University and Research Centre; Business Unit Greenhouse Horticulture. His supervisors are Prof. Eldert van Henten and Dr. Jochen Hemming. He received his BSc and MSc degree in Agricultural and Bioresource Engineering from Wageningen University and spent one year abroad for an internship dealing with greenhouse climate control, at Israel’s Technion Institute of Technology, and for thesis research dealing with localization of indoor robots, at the University of Illinois at Urbana-Champaign.