| Research Abstract |
In the European forest health monitoring, crown transparency is visually assessed, but problems exist in observer variability. The objectives of this research is to develop the image analysis system for crown condition assessments, called CROCO, to examine the precision of CROCO in terms of digital camera types and weather conditions and in comparison with the visual assessments.
In July and August 2001, we photographed 80 trees of four species (Norway spruce, fir, beech and oak) around Birmensdorf in Switzerland, using 4 different digital cameras, under different sky conditions (with-sun, against-sun and cloudy). The camera A tended to underestimate crown transparency (CT) for all species under with-sun condition as compared to the other cameras. The camera B tended to overestimate CT for beech, and this is because many leaves were highly reflective to be close to white. The variability of the camera A and C with about one million dot resolution was larger than that of the camera B and D with three million resolution. So we recommend using a digital camera with resolution of more than three million dots for CROCO.
We examined the relationships between the CROCO estimates and the visual estimates by the representatives from 12 European countries. The data were obtained at the International Calibration Course in Slovak 1999, including 15 trees for each Norway spruce, fir, Scots pine, beech and oak. The results showed that the CROCO had less variability as compared to the visual assessments.
|
