A Nondestructive Image Analysis Technique for Estimating Whole-tree Leaf Area

in HortTechnology
View More View Less
  • 1 Graduate Research Assistant, Urban Horticulture Institute, Cornell Univ., Ithaca, NY 14853.
  • | 2 Associate Professor/Program Leader, Urban Horticulture Institute, Cornell Univ., Ithaca, NY 14853.

The development of a rapid, accurate, yet nondestructive technique for expressing whole-tree leaf area would be extremely useful in studying various growth phenomena in trees. The objective of this research was to evaluate the accuracy of an image analysis process adapted for estimating the leaf surface area of four broad-leafed tree species (Amelanchier L. `Robin Hill Pink', Tilia americana L. `Redmond', Sophora japonica L. `Regent', Fraxinus americana L. `Autumn Purple' and Fraxinus pennsylvanica Marsh.). Video images of photographs taken of each tree canopy were quantified by an image analyzer into unitless surface area values or silhouette areas. The relationship between estimated leaf area as calculated from silhouette area and actual leaf area of these trees as determined by a leaf area meter was highly correlated. Use of this technique would enable a researcher, simply from serial photographs of the canopy, to retroactively estimate leaf or canopy area at crucial interim periods.