The speed of the sUAS is determined by the aircraft and sensor limitations. Multirotor aircraft are more adaptable to speeds which typically range from 5-9m/s. At a minimum, they can hover in place and increase in speed up to the limitation of the specific system. Fixed-wing aircraft are less adaptable, as they always require forward movement and maximum and minimum speeds are dependent on each specific system. Sensor limitations are related to the desired overlap as stated in the previous tab.

Altitude is primarily adjusted to achieve a desired spatial resolution and field of view for a certain sensor. The lower the altitude, the higher the spatial resolution (or point density for LIDAR sensors) and smaller field of view. Altitude is also adjusted for safety and legal reasons. A sUAS must be safely operated so it does not contact structures on the ground (trees, cell phone towers, terrain). Most sUAS are operated under 400ft (120m) AGL for regulatory reasons.

Proper overlap and side-lap are critical to successful and efficient sUAS remote sensing operations. Overlap is affected by sUAS forward flight speed and the scan rate or image capture interval of the sensor. Side-lap is adjusted by the closeness of parallel flight lines. Specific to structure from motion missions, overlap and side-lap should be 70%-90% each, though many have varying opinions. As a rule of thumb, you need more overlap and side-lap for homogeneous areas of interest, low pixel resolution sensors, and high distortion lenses.

Without a preexisting lens calibration file, a structure from motion software will generate its own calibration file using the overlapping images and appropriate control such as direct georeferencing or ground control points. The disadvantage to this method is the calibration can vary from dataset-to-dataset for the same sensor. This can cause problems for a temporal analysis of an area of interest.

An imaging sensor can be optically calibrated in a laboratory setting by a manufacturer, seller, or user. This is typically done by taking multiple pictures of a target, importing those photos to a structure from motion or other lens-calibration software, and generating a distortion file for that lens. The benefit of this methods is the distortion correction for the lens will be consistent for all data products generated from it.  

This body of knowledge extends our current experience. If you would like to add to this section, please contact the administrator.

This body of knowledge extends our current experience. If you would like to add to this section, please contact the administrator.

This body of knowledge extends our current experience. If you would like to add to this section, please contact the administrator.

GCP information can be pulled from source imagery like Google Earth and the National Agriculture Imagery Program (NAIP) for control points that have not moved or changed from the base imagery date to your current data set’s date. This type of ground control is referred to as a PIP, or Photo Identifiable Point. Good examples of PIPs are corners of buildings, sidewalks, and roadways. Bad examples of PIPs are trees because they change shape (growth & damage), color (leaf color), and composition (leaf-on or leaf-off).

Surveyed GCPs are any point within a dataset that has its location captured in-person by some sort of GPS equipment. This can be a PIP or purpose-made target such as a checkerboard panel. There is a range of equipment and associate accuracies for capturing surveyed GCPs. At the least accurate, a phone can be used to capture the location of a GCP. Next, hand-held GPS systems can achieve accuracies around 3 meters. Finally, survey-grade GPS can be used to achieve accuracies in the centimeters.