本研究以國立台灣海洋大學之操船模擬機設備為測試平台，整合Labview環境及軟硬體架構，分別探討以影像處理為基礎技術之視覺傳統式上下疊標導航法及改良式左右疊標導航法以及使用GPS資料之儀器視線導航法（line of sight guidance）共三種導航法，用於自動導引船舶航行至預定航道內之試驗。其中影像處理之導航法包含利用上下疊標導航法及左右疊標導航法這兩種視覺影像導航方式，其乃是利用架設於操船模擬機內之CCD攝影機，結合影像處理技術之色彩空間轉換及型態學閉合，門檻值設定及中值濾波等方法來鎖定兩疊標，並計算兩疊標中心點與本船CCD中心點之角度關係，而後傳送至控制器產生適當之舵角輸出命令，藉此導引船舶至預定航道內。而使用航儀訊號為主之視線導航法則是結合GPS準確之定位資訊用以計算出航向角修正量，並經由船舶航向自航控制器來執行航向變換控制，將船舶導引至預定航行區域。總結，本文應用三種不同之導航法結合內模式控制器以及模糊控制器二種控制器，於國立台灣海洋大學之真時操船模擬機上進行測試，本研究利用模擬機容易調整天候環境等干擾參數之便利性，對模擬之實驗船舶加入風、流等環境干擾條件，用以探討三種不同導航法及所結合之控制器，對於干擾條件下自動導引船舶至預定航行區域之影響，此外亦進行抑制效能及能源耗損之比較分析。由實驗結果顯示，視覺影像導引部份利用上下疊標導航法結合模糊控制器之船舶軌跡收斂行為較優於結合內模式之測試，視覺影像導引結合左右疊標導航法則之試驗船舶軌跡收斂性部分，則呈現近似單點導引效果。而利用航儀GPS定位資訊之儀器視線導航法於不同干擾強度下，期效果則取決於虛擬導航點之設定方式，因而可靠度需另加考量。

This work uses the National Taiwan Ocean University (NTOU) Shiphandling Simulator (SHS) as the testbed to examine the performance of three types of guidance schemes that guide a ship towards a prescribed straight water channel. Specifically, the first two types of guidance methods arc based on image processing technique; namely, the fore-aft leading mark and the left-right leading mark guidance methods, and the third one is a GPS-based Line-of-Sight (LOS) guidance method. The Labview working environment is adopted for the design of the Fuzzy Logic Controller (FLC) and the Internal Model Controller (IMC) adopted in this study. An image processing unit placing inside of bridge of the SHS is used to acquire images of leading marks projected onto the screen. The images of the leading marks are acquired through a sequence of image processing steps: color space transformation, threshold value screening, median filtering and closing operations, then the centers of the leading marks are identified. Finally, the angle between the leading marks and the distance from the leading marks to the center of the CCD camera are fed into controllers to steer the ship towards a prescribed channel. The LOS guidance method uses precise GPS positioning information to compute a reference heading angle to steer the ship through a sequence of course-changing maneuvers, Simulations under the wind and current effects indicate that the fore-all leading mark guidance method with FLC autopilot has better performance than with the IMC autopilot . The left-right leading mark guidance scheme is less satisfactory because of its single leading mark like performance, and the performance of the LOS guidance scheme under different disturbances depend s greatly on the selection of a proper virtual waypoint makes it not reliable in practical use.