AUTOMATED MICROPROPAGATED SUGARCANE SHOOT SEPARATION BY MACHINE VISION

Z. Wang, Paul Heinz Heinemann, P. N. Walker, Charles W. Heuser

Transactions of the ASAE · 1999 · 12 citations · 0 references

Concepts

Abstract

A prototype vision-guided separation mechanism for Stage 2 micropropagated sugarcane shoots was developedand tested. Micropropagated sugarcane plantlets grown between parallel plates showed two-dimensional structure whichgreatly simplified shoot identification using machine vision. Two identification methods were developed to locate the shootpositions in the shoot clump image. The shoot locations identified by the vision system were used to guide a pair of steppermotor-driven x-y tables to separate and transplant shoots. Machine vision algorithms and computer control algorithmswere developed, integrated, and tested in the prototype sugarcane shoot separation system. The best combination ofidentification algorithm and separation method resulted in 85% of total shoots successfully separated. A set of shootsseparated by the vision-guided system and a second set manually separated were compared in continued culture. A total of76% of shoots survived that were separated by the automated system versus 83% for manual separation.