MACHINE VISION TECHNIQUES FOR MEASURING THE CANOPY OF TOMATO SEEDLING

The canopy area of tomato seedlings is an important indicator of their quality. A machine vision system was designed for automated measureme nt. In a well-controlled lighting environment, it is feasible to use c onstant threshold values as a first step in partitioning images of tom ato seedlings into the meaningful parts: canopy, root-growth medium, a nd a background. The selection of threshold values, however, should be based on an analysis of a training set of similar tomato seedling plu gs. An image processing procedure was developed and evaluated to segme nt seedling canopy from background and root-growth medium parts of the seedling images. The segmentation process adopted an adaptive thresho lding technique, the Otsu method, and extended it for an efficient mul timode operation. The multimode segmentation operation accurately and quickly identified the canopy portion of tomato seedlings. (C) 1996 Si lsoe Research Institute.