CROCO : Semi-automatic Image Analysis System for Crown Condition Assessment in Forest Health Monitoring

抄録

A semi-automatic image analysis system, CROCO, was developed for assessing tree crown condition objectively and at low cost in forest health monitoring. The system is composed of a Macintosh computer, a digital camera, the commercially available software Adobe Photoshop and the public domain software NIH Image. Most of the procedures are automated by the batch processing tool of Adobe Photoshop and a set of macro programs of NIH Image, which enable us to analyze large numbers of images rapidly and at low cost. After pre-processing, the macros automatically generate a silhouette from the color image based on the between-class variance method and calculate the index DSO as a measure of crown transparency based on two fractal dimensions. The precision of the system was examined in terms of weather conditions, camera angle (CA) and overlap rate of the target crown with other trees (OR). The results suggested that CROCO can provide a consistent measure of DSO irrespective of either cloudy or sunny conditions, if photographing conditions satisfy the criteria of CA less than about 45 degrees and OR less than about 50% of crown width. CROCO may be used as a control assessment to detect and correct observer bias.