影像追蹤方法影用在監控系統之研究

A Study of a Monitoring and Control System Using Image Tracking Methods

蔡博智(Po-Chih Tsai)

指導教授 ： 陳冠宇 ； 張永鵬

中原大學/工學院/機械工程研究所/碩士(2002年)

https://doi.org/10.6840/cycu200200375

摘要

本文利用影像處理方法建構以個人電腦為平台的即時影像追蹤系統。目標物判別是將影像經過二值化及中通濾波處理，分離目標物及背景之後，使用影像相減法，移動邊緣檢測法及提出移動目標平移法，求出目標物在影像中之位置，搭配自行設計具有二個旋轉方向的CCD攝影機基座，使攝影機得以即時追蹤移動目標物。本文設計不同實驗以比較三種方法之優缺點。
本文實驗分為二個階段：第一階段為單純背景中的追蹤實驗，目的在找出最佳的馬達控制模型。共設計三種不同的馬達控制模式，分別為固定轉速、變化轉速及一次到位；第二階段為正常背景中的追蹤實驗，目的在模擬真實環境下的影像追蹤。本文提出移動目標平移法配合移動邊緣檢測法，在目標物不變形或是變形量不大的條件下，可以有效改善移動邊緣檢測法無法得到移動目標物最新位置的延遲現象，達成即時影像追蹤的目的。

關鍵字

即時影像追蹤 ； 移動目標平移法 ； 移動邊緣檢測法

並列摘要

In this dissertation, we construct the real-time image tracking system on PC-based platform. To differentiate the target, we convert the image to binary by thresholding and we process it with median-filter. After separating the target object and background, we use image differencing method, moving edge detection method, and moving target shifting method to find the location of target object in images. By collocating the camera base which has 2-direction rotation. It lets the camera tracking moving object immediately. In this dissertation, we design different experiments to compare the merit and defect of three methods.
In this dissertation, the experiments age divided into two stages. The first stage is the image tracking experiments under simple background. The purpose of the experiments is finding the best motor control model. We total design three different model of motor controlling. It’s a fixed rotational speed, a various rotational speed, and a fixed rotational angle. The second stage is the image tracking experiments under normal background. The purpose of the experiments is simulating the image tracking system under the real conditions. In this dissertation, we propose a moving target shifting method to combine the moving edge detection method which can improve the delay decision of using moving edge detection method, under the deformation of the moving target is a little.

並列關鍵字

real-time image tracking system ； moving target ； moving edge detection

參考文獻

Cloud Motion Patterns from Geosynchronous Satellite Image Data,” Pattern Recognition, Clustering, Statistics, Grammars,Learning, Vol.2,pp. 272-292,1970.

[2] Horn, K. P. and B. G. Schunck,“Determining optical flow,” Artificial

Intelligence, Vol.17, pp. 185-203, 1981.

[3] Fuh, C. S. and Maragos P., “Region-Based Optical Flow Estimation,”

Proceedings of IEEE Conference on Computer Vision and Pattern