Abstract
This paper studies the texture of the tree crown making the forest scene in the middle distance. Our procedures are these: (1) 48 subjects are employed in an experiment to classify the forest scenes by their own impressions when they see them, and they are divided into five groups. (2) The typical photographs taken by each group are analyzed by using GMHL in terms of the indexes of the dispersion (DP), the coefficient of skewness (CS), and the coefficient of excess (CE). (3) Finally the relevance of the results of (1) and (2) are examined. From above studies, we understand that DP and CS are effective as the indexes for quantitative classification and the results obtained from the subjects tend to be characteristically placed on the frame of coordinates made by indexes. Consequently, the effectiveness of texture as the axes of analysis / arrangement for classifying forest scenes are positively shown form a quantitative viewpoint.