| Summary: | 碩士 === 國立臺灣大學 === 園藝學研究所 === 98 === The main purpose of this study is to evaluate the effect of different color qualities and complexity on people’s preferences for natural landscape. In this study, Lab color space is used to quantify color qualities, and Fractal Dimension is applied as the index of landscape complexity. In the beginning, large amount of landscape pictures were collected from the database built earlier. Then researcher used software to analyze each picture, calculating fractal dimension and the color qualities in Lab color space, including Brightness, Hue, Chroma and the variance of color. The final samples contained 210 pictures with varying color qualities and fractal dimension, including 10 Baseline pictures and two groups of 100 pictures each. Second, people’s landscape preferences of each picture were collected through a 10-point-scale questionnaire. Finally, the quantified data was analyzed to vertify the hypotheses. The result shows that people’s landscape preference is negative correlated to L, but positive correlated to a, aS.D., b2, and Chroma in Lab color space. An inverted U-shape curve was found between fractal dimension and landscape preference, 1.333 is the most preferred Db value. The regression model shows that a, b2, Db, Db2 can explain 20.4% of landscape preference. According to the results, people tend to prefer natural landscape which is darker in Brightness, more red, yellow or blue in Hue, more colorful in Saturation and more red-green variances in color, and lower-middle landscape complexity is preferred more. This study proves that Lab color space is an effective tool to quantify color qualities, and Fractal Dimension is a predictive index of landscape complexity. The results can contribute to future studies in the field of landscape preference.
|
|---|
