Influence of mushroom strains and population density on the performance of a robotic harvester

The effects of mushroom strain, distribution, size, growing angle and flush number on the performance of an experimental robotic harvesting rig are described. The robotic system has three main elements: a machine vision system incorporating mushroom location and picking order selection algorithms, a computer-controlled Cartesian robot and a specialized mushroom picking end-effector. The picking strategy is based on picking freely available mushrooms or mushrooms on the outside of a clump first. Two series of experiments were conducted; the first using mushrooms grown in deep polythene bags and the second using mushrooms grown in shallow plastic trays. The experimental robotic harvester was capable of successfully picking about 70% of the mushrooms from a growing container. Mushroom strain had a significant effect on the picking success of the harvesting rig and there was a higher success rate with larger mushrooms. The success rate was highest (76%) with mushroom strain Somycel U1. The differences observed between strains were mainly due to differences in the number of mushrooms produced in the cropped area rather than to differences between individual mushrooms from different strains. High densities of mushrooms (which occurred particularly in the second flush of a crop) resulted in poorer picking performance, but the picking performance was not significantly affected by mushrooms touching each other or growing on the periphery of groups. (C) 1997 Silsoe Research Institute.