| Summary: | 碩士 === 國立臺灣科技大學 === 電機工程系 === 97 === How to detect and track moving objects in video streams is an important and challenging research problem in visual tracking applications. Among the moving object detection algorithms, the background subtraction method has been widely use due to its less computational load and high detection quality. Nevertheless, May be interfered with by the light and noise or the color of object is similar to the background while shooting, this method can not effective detecting the object, and operation complexity is high. In view of this, we develop a method, in case of not destroying the contour of object, get the contour of the object, connect labeling information and fill the hole of the object, we can get more information and the operation complexity is low. In addition, in the realistic environment, tacking objects are usually non-single object, in the track of many objects, when the objects produced the intersection, we propose a judgement method, calculate the parameter of the object first, Differentiate the front or the back position of the object with the parameter of the object, after get the front or the back position of the object, and then, by judging the boundary coordinate of the object, narrow the search area while tracking object, can reduce operation complexity effectively, detect the object contour and the object position while intersection, can know the result is good by the experiment.
|
|---|
