| Summary: | 碩士 === 國立臺灣科技大學 === 自動化及控制研究所 === 91 === Recently, a CCD camera is usually used in factory automation and community security. However , it is restricted by hardware function that can’t get the real time information at suddenly accident. The system can’t automatically track the suspicious object in real time. Therefore, using an intelligent, automatic monitoring device instead of human is the probable and efficient way.
In this research, the image processing technique and fuzzy control rules are applied to develop an intelligent dynamic tracking system for image with single moving object. Firstly, by using the temporal differencing method, two images are grabbed. The image processing techniques such as spatial filtering, binary of image, image opening, clustering are also used. The moving object is extracted from the captured image and. The fuzzy sliding control method is used to estimate the location of moving object. According to coordinate translation, the motion command is generated driving stepping motor to tracking the moving object on screen. The object to be tracked doesn’t limit on the shape but also the moving state. The average time of the tracking system for a complete process takes about 300~350ms, means 2~3 images per second.
|
|---|
