Tracking without tagging the objects
CAPTA technology is a markerless deep-learning based detection and robust tracking of objects in camera images, for recognition and tracking of workpieces or complete vehicle bodies, support at assembly maintenance tasks and tracking of products during Marketing Demonstrations.
CAPTA Application Software detects objects in camera images with an initially rough simultaneous 6DOF (6 dimensions of freedom) pose estimation followed by an exact edgebased frame-by-frame tracking.
CAPTA opens up unimagined application possibilities, since now any object can be detected and tracked - and are found again, especially when losing position, without the user having to intervene.
All image-based tracking algorithms have the initial task to recognize a trackable object independent from its orientation and start tracking. Once the object is identified and an initial pose is found, many algorithms work very accurate and reliable. Nevertheless, getting the initial pose is the biggest challenge and often has to be found manually.
CAPTA Application Software technology contains a robust algorithm which identifies the basic pose of a trackable object without markers immediately when it enters the camera view.
This algorithm works self-contained, so no need to connect to cloud services or similar, which allows using it inside company networks. It is also very robust against cluttered backgrounds, bad lighting conditions or partly covered objects.
Possible applications are recognition and tracking of workpieces or complete vehicle bodies, support at assembly maintenance tasks and tracking of products during
Marketing Demonstrations – and this even in difficult environments with poor lighting or cluttered background.