Georeferencing
Georeferencing (image registration) can be defined as the registration of an image (usually, all data are converted into a matrix format and then shown as an image) coordinates to a specific geographic coordinate system. As a result, each pixel of the image will correspond to a point (a small area) on the ground and can be shown in geographic software. If the data collection device is equipped with RTK-GPS, the output would precisely match the real-world coordinates. Lidar and SAR scanners usually work in this way. However, most of the devices use GPS receivers with an accuracy of more than a meter. As a result, the recorded data might be off by up to several meters. In this case, Ground Control Points (GCP, marked points on the ground with known coordinates) are used to align images, and the precision of alignment depends on the number (0.2-0.4 GCPs per acre) and distribution (edges of the study area plus inside the area) of the GCPs [1] [2]. In case absolute georeferencing is not required and several images need to be registered on top of each other, some prominent mutual features can be used as GCPs to alight the images.
Georeferencing can be done either for a single image or for mosaics that are created using multiple stitched images. Some studies cover the whole study site in a single image avoiding commercial mosaicing software. In this approach, special attention must be paid to the effects of wide-angle imaging in calibration and GSD differences. When the GCPs are determined, georeferencing can be done manually or by a customized script.
References
[1] P. Martínez-Carricondo, F. Agüera-Vega, F. Carvajal-Ramírez, F.-J. Mesas-Carrascosa, A. García-Ferrer, and F.-J. Pérez-Porras, “Assessment of UAV-photogrammetric mapping accuracy based on variation of ground control points,” International journal of applied earth observation and geoinformation, vol. 72, pp. 1–10, 2018.
[2] M. X. Tagle Casapia, “Study of radiometric variations in Unmanned Aerial Vehicle remote sensing imagery for vegetation mapping,” Lund University GEM thesis series, 2017.