Image insertion into a video stream using a combined physical sensors and pattern recognition

摘要

(57) the event one or more cameras of the [summary] raw video insertion system (110) has a (113) sensor that detects the tilt of the camera zoom, focus, and bread. Sensor data from each camera is fed to the raw video insertion system to give an indication of the approximate insertion occurring in the video scene. Tally data and sensor changes taking the search mode of the pattern recognition live video insertion systems usually inherently. Final determination of the correct insertion position is determined by analyzing the structural features and / or accurate video images. This analysis compares the position of the structural features and / or within the video frame in the corresponding position of the image of the previous reference image or a common and its surrounding insertion position.