How do I assess data quality?

Display image footprints and metadata, and set attitude filters for feedback on mission planning and map quality.

The quality of your data is dependent on the quality of your mission planning. It is imperative that you read and understand the content in the Operation and in the Mission Planning sections of the SlantRange Integration Guide, regarding sensor mounting and aircraft flight planning considerations. Several tools in SlantView will help you visualize your flights and recognize factors that affect data quality.

Displaying image footprints

Aircraft heading, roll, pitch, yaw, and speed over ground all affect the alignment and overlap of images relative to one another. Click DisplayImages’ footprints to show image outlines overlaid on the field in the Map Window. For best results, fly in straight flight legs with consistent velocity and orientation. To produce high quality maps, images should have at least 30% along-track overlap and 20% sidelap, as shown in Figure 1.

Figure 1: Image along track overlap and sidelap

To get a sense of how aircraft orientation affects along-track overlap, consider a multirotor at the beginning of a new flight leg. After the multirotor turns 180 degrees at the edge of a field, it will accelerate to the velocity set in the mission planning software for its next pass. This acceleration causes the multirotor to pitch forward, angling the field senor slightly behind the aircraft. As it reaches its max velocity, the acceleration goes to zero, and the field sensor returns to nadir orientation (as established by the field sensor’s angled mounting bracket). Along-track image overlap will vary according to multirotor acceleration throughout a given flight; we recommend flying slightly beyond the edges of the field to allow the multirotor to change direction and accelerate while it is not imaging content of interest. SlantView filters out images captured during turns using the Set Filters parameters in the Settings—Processing menu based on thresholds for heading, roll, pitch, and yaw.

To get a sense of how aircraft orientation affects image sidelap, consider an aircraft in the presence of a crosswind. A multirotor’s roll angle will change to maintain the given heading, tilting the field sensor to the port or starboard side. A fixed wing aircraft’s yaw angle will change in the presence of a crosswind, and the aircraft will “crab” to maintain a given heading, resulting in image footprints that are slighlty rotated relative to the heading of the given flight leg. Image rotation can compromise image overlap (as seen in Figure 3). For best results, fly perpendicular to the crosswind. On a multirotor, the roll angle will be mirrored from one flight leg to the next, and even though the field sensor is leaning to the port or starboard side, image overlap from one pass to the next will be consistent. When flying parallel to the wind, the pitch angle is dramatically higher when flying into the wind versus with the wind, comprising consistent along-track image overlap, potentially reducing data quality.

Keep the relationships between weather conditions, aircraft orientation, and resulting image location in mind when planning missions, and diagnosing map artifacts. Lack of overlap can result in poor image registration and holes in your maps.

For reference, two maps’ image footprints collected from high quality mission plans are shown below. Notice the consistent along-track image overlap and image sidelap, combined with nearly perfect parallel flight legs in Figure 2. Better flights produce better data. However, you do not need perfect weather conditions or perfect flight planning to produce high quality data. Notice the changes in aircraft heading present in the flight legs of Figure 3; flown with slightly higher overlap to compensate for crosswind, SlantView is still able to produce very high quality maps. The calibrated SlantRange sensor combined with SlantView’s processing and filtering techniques have extracted high quality results around the world in all kinds of weather, in all stages of growth.

Figure 2: Very high quality flight, consistent image overlap

Figure 3: High quality flight, maintaining overlap in the presence of crosswind

Displaying image metadata

Click Display—Image view—Show metadata to view the aircraft roll, pitch, yaw, heading, altitude, and speed over ground statistics shown in Figure 4. After enabling the metadata view, while active in the Map Window, drag the cursor over an image to see its metadata displayed in the top left of the Image Window. Combined with viewing Image Footprints, image metadata will help you identify sources of error.

Figure 4: Image metadata

Identify a few flight legs and move the mouse from one image to the next in order with the image numbers shown in the Title Bar (boxed in red in Figure 4). Assess the variation in roll, pitch, yaw, heading, altitude, and speed over ground across the flight leg. Good flight legs will show minimal variation in these values. Large differences of about 5-10º in a given flight leg, could be the result of gusting winds, aircraft instability, etc. If there are gaps in flight legs that look like missing images, it is possible that SlantView excluded images from the map based on the images' metadata as established by the Set Filters thresholds in the Settings menu.

Setting metadata filters

In the Settings menu, edit the Set filters parameters to display images within a given percentage of steady-state altitude, pitch, roll, yaw, or heading. The number of images used in a given map based on the filter settings is shown boxed in red in Figure 5. To remove more or fewer images from the map, the user can increase or decrease the width of these filters on aircraft attitude. Images can be removed from the Workspace by narrowing the filters and Recalculating. However, to increase the width of the filter, you must Reprocess a Dataset from scratch with the wider tolerances indicated in the Settings—Set filtersmenu before pressing OK to Process the Dataset.

Note
If the default SlantView filters throw out too many images in a given Dataset, improper mission planning or poor flight conditions are almost certainly limiting the quality of your data.

Figure 5: Aircraft attitude filtering options

Note
Do not attempt to manually remove images from a Dataset folder when viewed in the file explorer. SlantView uses the metadata collected from all images taken from takeoff through landing (even images excluded from map generation) to establish ground elevation, flight altitude, flight leg direction, and a multitude of other critical measurements. Removing images can compromise the integrity of the flight metadata. SlantView will automatically exclude unnecessary images from your maps.