The Arcona developers team keeps on working on the proprietary AR Viewer’s advancement. In order to improve the usability of the application, our team is constantly testing and implementing new algorithms of 3D objects positioning on the screens of mobile devices.
“Tracking the movement patterns of the user device is very important for precise and stable positioning”, says Alexander Emelyanov, Ph.D., R&D Team lead of Arcona. “This tracking is based on analysis of numerous images from a built-in photo-video camera. In robotics, matching of several dots with highly contrasting characteristics is sufficient for that, but when it comes to the AR systems it’s much more complicated to determine the trajectories. In the hands of the user, the motions of the device are not that smooth and can hardly be compared to those of cameras the robots are equipped with. We strive to match the images very different from each other, this process requires to be based on more than some separate dots”.
The Arcona engineers have developed an alternative algorithm of matching the images based on the analysis of the relative position of the straight lines. Unlike dots lines have a broader set of features, enhancing the accuracy and effectiveness of camera trajectories determination.
Besides that, our algorithm ‘looks’ not only at the image brightness close to the line but also at its color. This increases the reliability of the images matching.
The new algorithm has successfully passed testing, we’ve had really good results. The stability and accuracy of tracking the trajectory of the user camera are much higher than similar indicators of standard algorithms. In addition, we were able to obtain more reliable matches from fewer matched lines. This saves system resources and increases performance, which means that the application will be able to work well on less powerful mobile devices.