Robotic detection of leaf angle and grasp point with low-cost sensors

Junzhe Wu, Wangqi Ma, Girish Chowdhary

Keywords: Neural network, Sensor Fusion, Leaf angle, Grasp point

TL;DR: A field robotic system equipped with low-cost sensors is utilized to automatically detect leaf angle and grasp point.

Abstract: Leaf angle is a significant architectural parameter of plant canopy due to its influence on solar light absorption and photosynthetic efficiency, and hence also on the plant growth and productivity. Traditional way for leaf angle measurement is manual, and researchers need to go to the field and use a protractor to collect leaf angle data. However, manual leaf measurement is labor intensive, not ergonomic, and often not consistent between different people. In this paper, TerraSentia, a field robotic system equipped with low-cost sensors is utilized to automatically detect leaf angle and grasp point. The proposed method can produce a consistent leaf rolling angle estimate quantitatively and qualitatively on multiple corn leaves, especially on leaves with multiple different angles.