Determinação da posição e atitudes de VANT por fotogrametria

This paper describes the concepts of positioning by photogrammetry for Unmanned Aerial Vehicle (UAV), in the frame of project Processing of Images in Real-Time (PITER) developed at the Institute for Advanced Studies (IEAv). This project intends to navigate using georeferenced orbital or aerial images compared to photographs taken during the flight. In order to enable navigation through images, this technique is very important for the autonomous flight of a UAV, since usually these aircraft navigate using the Global Navigation Satellite System (GNSS) associated with an Inertial Navigation System (INS), however, The GNSS system is subject to jamming and spoofing interference, which at any time can be enabled to degrade the accuracy of the system. Due to that, it is impossible to navigate because the INS is a standalone system that during navigation time degrades position accuracy, to accumulate errors in measurements of acceleration and speed of the UAV. Because of that, the INS has its coordinated corrected from time to time by GNSS, in this case the navigation system is dependent on GNSS. However, to perform this replacement it is necessary to calculate the coordinates and attitudes of the perspective center of the camera and its respective accuracies, using control points to appear as well defined in geo-referenced image, the image taken by the onboard camera and performing the correlation between respective points in real time, in order to perform the replacement when the INS coordinate accuracy is smaller than the coordinate obtained by photogrammetry. In this context, the replacement of the GNSS allows the autonomous navigation of UAV.