Establishing ground control points
UAVs are fitted with Global Positioning System (GPS) devices which record the geographical location of the camera at each position that a picture is taken. This information is intended to be used to align and georeference the pictures into a seamless image. In most cases, GPSs fitted on UAVs are termed as “consumer grade”, although some few have higher precision GPSs (Real Time Kinematic) installed. This means that the positions obtained with these devices are not as accurate as those obtained from differential/kinematic GPS units (e.g. eBee SenseFly RTK). Therefore, the use of data from the UAV GPS alone for georeferencing the UAV pictures may lead to a lower geometric accuracy and negatively affect the quality of products that will subsequently be derived.
One way of improving the geometric accuracy of your pictures is to establish and determine the coordinates of ground control points (GCPs). GCPs are built/constructed on the ground of the mission area prior to flight. The construction can be done with concrete or existing rocks/structures can be painted directly (Figure 5.4). Figure 5.4 demonstrate how the STARS team in Mali constructed GCPs. The cross may have dimensions of about 1.5 x 1.5 m and a line width of 20 cm. GCPs are often painted white over a dark background. Coloring them this way ensures that they are easily recognizable in the resulting images.
Ideally, GCPs are supposed to be evenly distributed over the entire flight area to ensure a good geometric correction (Figure 5.3). It is a good practice to construct as many GCPs within your area of interest as possible.
Figure 5.4. Left: construction of a ground control point site. Right: painting of an existing structure to serve as a GCP (source: STARS ISABELA team)
The geographical coordinates of GCPs must be determined to a high accuracy. The accuracy of the GCPs must be determined in line with the spatial resolution of the intended imagery. In general, GCP accuracy should be higher than 1/3 of the ground spatial resolution of the intended image. This will often require the use of a differential/kinematic GPS. A kinematic GPS (unlike handheld or consumer grade GPS) must have two receivers; one set up on a point whose coordinates are accurately known (reference station), and the other on a point whose accuracy has to be determined (i.e. the GCPs; rover station). Monuments/pillars with known coordinates often belong to, and are maintained by, national mapping agencies. You would have to contact such agencies to obtain coordinates of the monuments to enable you process the GPS information. Details about the operation of kinematic GPSs can be found here.
Figure 5.5. Example of a handheld GPS
Figure 5.6a. Example of a kinematic GPS receiver (reference). (Source: STARS ISABELA team)
Figure 5.6b. Example of a kinematic GPS receiver (rover). (Source: STARS ISABELA team)
As we would see later, the GCPs can be used together with the geotags (locations from the consumer grade GPSs onboard UAVs) in UAV image processing to achieve optimal georeferencing and accuracy of derived products (i.e. ortho mosaics).