基於快包法的多目標色彩模型設計

本論文提出種子區域成長法跟快包法為基礎的建立多目標色彩模型的方法。以FIRA (Federation of International Robot-soccer Association) RoboSot規則之中型足球機器人下的全方位視覺系統為發展平台。本方法目的是改善傳統以人工方式建立多目標色彩模型費時、費力的缺點。因不同使用者建立色彩模型造成效果差異，影響了影像系統判斷目標物的穩定性。本論文以種子區域成長法 (Seeded Region Growing, SRG) 來自動選取目標樣本，利用目標顏色與環境中的差異性來判斷目標顏色區域。接下來，將SRG建立的目邊顏色的像素，利用快包法 (Quickhull) 來自動建立目標色彩模型範圍。並且改善一般色彩空間中，用色相、飽和度與亮度之上下界建立之色彩模型容易包含過多不屬於目標色彩樣本的問題，使得色彩模型更貼近目標之顏色，增加色彩分割和目標辨識的成功機率。最後由實驗結果中，得知本論文快包法較傳統人工方式建立之色彩模型更有效率以及更精確。

關鍵字

色彩模型；快包法；種子區域成長法

並列摘要

A system design of multi-object color models based on quickhull is proposed. This study is developed on omnidirectional vision system of middle-size robots with the competition of FIRA (Federation of International Robot-soccer Association) RoboSot. Originally, color models are built manually by user to adjust six thresholds in HSV (Hue, Saturation and Value) color space. The original method is time-consuming, and the color models are unstable by different users. The new system design replace the original method in order to promote the efficiency of building multi-object color models and improve the problem which misjudge the object. This thesis propose a system design that combine two algorithms which are SRG (Seeded Region Growing) and Quickhull to catch the color pattern and build color models and . The SRG algorithm is used to get the color pattern of targets and saved to be the sample of color models. The SRG method can distinguish different color efficiently. The Quickhull algorithm take the color pattern from SRG to build color models and ignore the value parameter in the Quickhull color models because value is unstable for the environment, just using the hue and saturation parameter to bound the region. The experimental result show the Quickhull color model is more efficient and more precise than the original method.

並列關鍵字

Color model ； Quickhull ； Seeded Region Growing

參考文獻

[7] 何丞堯，全方位視覺足球機器人之自我定位系統的設計與實現，淡江大學電機工程學系碩士論文（指導教授：翁慶昌），2009。

[8] 鄧宏志，中型機器人足球系統之即時影像處理，淡江大學電機工程學系碩士論文（指導教授：翁慶昌），2006。

[4] C. B. Barber, D. P. Dobkin, and H. Huhdanpaa, “The Quickhull Algorithm for convex hulls,” ACM Transactions on Mathematical Software (TOMS), vol. 22, no. 4, pp. 469-483, 1996.

[5] R. Adams and L. Bischof, “Seeded Region Growing,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 16, no. 6, pp. 641-647, 1994.

[6] Andrew Mehnert, Paul Jackway, “An improved seeded region growing algorithm,” Pattern Recognition Letters, vol. 18, pp. 1065-1071, 1997.

國際替代計量

基於快包法的多目標色彩模型設計

全文下載

主題瀏覽