A Tracking System for Moving Target

Abstract

在本篇論文中我們利用影像追蹤測量動態物體的座標位置，建構出一套立體三度空間影像追蹤測量系統。由於要偵測的物體在移動時具有振動的特性，因此在影像處理部分採用Motion Segment Algorithm與kalman filter運動估測之結合。這個方法可以追蹤形狀是未知的物體，且可以快速在複雜的背景中分辨出目標物。實驗的內容包含兩項，一是目標物追蹤功能測試，測試系統是否在時間內追蹤移動目標物；二是測試系統測量功能，實驗中將實際追蹤測量移動目標物的三維空間座標。整個實驗的目的是測試這套系統的功能，而實驗的結果與預期相符，目標物的移動角速度不能超過硬體上的限制，也就是目標物移動的範圍與速度必須在系統要求內，則系統對任意移動目標做良好的追蹤與預測之功能動作，反之，若是目標物超過了硬體限制，則會造成系統的誤動作。

In this thesis, we have developed a tracking system for moving target. Since the moving object has vibration property, we adopt “motion segment method with the aid of Kalman filter” for motion detection. The advantage of such an approach is that the shape of target object can be easier identified it in a complicated background. There are two main experimental study in this research. One is to test the performance of the tracking algorithm for detecting moving object, and the other is to evaluate the measuring system performance for operating in a larger environment. The experimental results are found to match expected the tracking performance.