Continuous overlap pursuit method to detect moving objects in natural scenes

摘要

To detect moving objects in natural scenes by image data, the authors have developed an object detection algorithm called the continuous overlap pursuit (COP) method. The COP method, which uses continuous plural image data and pursues difference images, can detect moving objects even with much background noise such as swinging trees, white noise, and changes of brightness and other factors in the image data. The COP method is constructed by basic image processing, so it is applied easily to image processing hardware for real-time recognition.