查詢結果分析
來源資料
頁籤選單縮合
題名 | 使用雙眼立體機器視覺定位樹果 |
---|---|
作者 | 鍾瑞永; 李芳繁; | 書刊名 | 農業工程學報 |
卷期 | 39:4 1993.12[民82.12] |
頁次 | 頁74-89 |
分類號 | 435.3 |
關鍵詞 | 立體視覺; 影像處理; 樹果定位; Stereo vision; Image processing; Tree fruit positioning; |
語文 | 中文(Chinese) |
中文摘要 | 本文主要目的在於建立樹果自動化採收系統之視覺系統基本模式,研究內容包括機器視覺3D定位系統之攝影機參數校正理論與取像、影像分割、特徵區域定義及定位值析出之方法,以期利用視覺系統定位樹果3D位置。 機器視覺定位系統使用通用立體攝影機幾何模型同時擷取左右兩張影像,以提高取像靈活性;但同時卻增加左右兩影像的比對困難度,因此使用輔助雷射光點增強目標果物之特徵。取像完畢後,以目標果物的色相及彩度值為標準分割原始影像,判別影像中是否含有目標物,並依雷射光點之紅色與色相值為標準搜尋左右兩影像之對應點,代入3D定位計算式中,求取目標果物三度空間位置。 |
英文摘要 | The main purpose of this paper is to establish a basic model of machine vision system for an automatic citrus fruit picking device. A machine vision 3-D positioning system was used in this study. The general stereo camera geometric model was utilized, for greater flexibility, to grab images simultaneously by two cameras located at different places, but it also increased the degree of matching difficulty. To solve this problem, an assistant laser light point was employed to add feature to the target fruit. The hue and saturation values of the fruit were used to segment the image and then to judge if there was any target fruit in the image. The red and hue values of the laser light point were used to search the corresponding points which were used to calculate the 3-D position of the target fruit. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。