使用模糊集理論對色彩影像進行天氣狀況判別之研究

本篇論文中 , 我們所架構的系統其主要是由金字塔結構及模糊集理論結合而成 , 其可以應用在影像分析方面 , 且可以模擬近似於人性化思考模式 o 例如 : 人類對天空影像的感覺 , 可以給予一個語言意義表示其狀況 (如 : 白天 , 雨天等) o 此種感覺不易以傳統方法表達 , 且語言意義也難以用數學式表示之 , 所以 , 我們在這篇論文裡以模糊集理論將人類語意轉換成隸屬函數 , 以隸屬函數代表一語意 , 並制定決策規則用以說明景物及物體影像 o 有關影像之隸屬函數是利用影像之統計圖來產生 o 採用模糊集理論的優點是可以簡化演算過程的複雜性 o 另一方面 , 由於金字塔結構是一個相當簡單的階層關係 , 其優點是當影像資料相當大時可以降低其負載量及整個處理過程的演算時間 o 綜合以上之特點 , 我們所架構之系統最主要的目的是讓系統能模擬人類思維方式(金字塔分析模式以及模糊語意判斷) , 藉由模糊推論法則及金字塔結構之層與層的相關性 , 此系統可以對外界天空影像判斷其天氣狀況 , 並獲得一近似人類語意的輸出結果 o

關鍵字

模糊集理論；彩色視覺；推論法則

並列摘要

In this thesis, an application of fuzzy set theory and pyramid structure to the image analysis is presented. It is about to the human sense for image landscape that human can easily assign a linguistic meaning of an image by its scene feeling, however, which is hard to represent with traditional algorithms and mathematical equations. For this reason, we propose an approach using fuzzy set theory and pyramid structure to express the scenery and the decision rules with natural language of human beings via the fuzzy sets. This approach can extract the image global information based on a fuzzy set constructed from the histogram of the image. And, the pyramid structure adopted for analyzing the relationship among the partitions of an image has the advantage that it can reduce the quantity not only for a plenty of image data but also for a long computation time. In this thesis, we show how the fuzzy rules of an image (including gray and color images) is produced and how the linguistic meaning which simulates the human feeling for an image is given. The new approach applied to analyzing the weather feelings from the sky images is presented. Experiments show that the proposed system is feasible.

並列關鍵字

fuzzy set ； color perception ； inference rule

國際替代計量

使用模糊集理論對色彩影像進行天氣狀況判別之研究

主題瀏覽