Abstract
The wearable video system supports decision making through analyzing video images. Because personnel who wear the system or the object monitored is always moving, the technology of motion detection and target contour extraction used in the wearable video system are studied. First, the frequently-used algorithms for motion detection are studied, including frame difference method and background subtraction method, and an algorithm composing of those methods for motion detection is proposed. The principle using traditional snake model to extract contour is analyzed. The result of motion detection is used to auto initialize the contour of target band, and B-spline Snake model is used to extract the contour of moving object. The test achieved good results.