How to plan drone operations: Our take-aways from mapping large areas
— Research by Sajil KR
Your maps are only as good as your data
The success of a drone operation is dependent on the quality of data gathered. Plan your flights to keep operational bottlenecks, processing capabilities, and economics into account. We conducted large scale operations, reconstructed the data in-house, and also identified assets with our deep learning toolkits. Along the way, we have zeroed-in on key factors that lead to successful data collection.
What to consider while planning your flights?
Type of output: We need to adjust all the other factors of mapping depending on the quality of map desired in 2D or 3D.
Type of drone: Flight area and distance changes with the type of drone. For a fixed wing aircraft the planned area would be more and higher range, while, in case of quadcopters, it would be comparatively small.
Elevation analysis: Elevation profile of the area needs to be taken into account for ensuring accuracy and also the safety of the flight. The difference between the lowest elevation and highest elevation points in the area should come in a certain range to enable us to put the whole map together.
Flying height of the drone: The area covered in a swathe would increase if the drone is flying higher. Changes in elevation can be accommodated if the flying height is accounting for complications arising from the difference in ground sampling distance (should not be more than twice).
Time taken for a single flight: Depending on the battery capacity of your drone, the flight times are different. Your plans need to be fine-tuned accordingly.
Launch location: Launch location should be accessible for the pilots, within regulatory line-of-sight limits, and optimised for the return of the drone in case of emergencies.
Case Study : Deciding your flight plans
For instance, the total area for this deployment is approximately 380 Sqkm with the area of interest encompassing upto 100 discrete villages. In order to process the entire area to create an orthomosaic, it is important to decide the size of the processing chunk based on current infrastructure capabilities. Would not be a problem with cloud but would cost enormously.
Area of Interest:
Type of drone chosen: Fixed wing ( daily throughput would decide how much area we can cover everyday).
Deciding the per unit processing chunk:
As discussed, divisions of the area of interest are decided based on processing capabilities and the total size of the images. From our processing engine configurations and the file sizes, it was concluded that we will internal infrastructure to process 90–100 sqkm in one attempt. Hence, the area is divided into four separate plans with dedicated overlap in the adjacent plans. Approximately, Patch-1 is 85 sqm, Patch-2 is 90 sq Km, Patch-3 is 110 sq Km, and Patch-4 is 96 sq Km.
Factors to be considered in planning ground control points:
- Distance between points: To hold the map effectively, the points should not be extremely close or even very far from each other.
- Accessibility of the location: The location suggested should be safe and practical for a person to reach and lay either permanent or temporary GCP markings.
- Edge and Center points: The points should have an even distribution across the edges and the center of the area of interest.
- Purpose of mapping: The plan would change based on the requirements. Corridor mapping is inherently different from construction tracking in urban areas.
- Elevation profile: If there is a sudden change of the elevation in the area of interest, say, a mountain or a small hill we need to assign points on either side.
- Known GPS points in the AOI: If the area of interest is having any known accurate GPS benchmark, then one GCP should be taken on that location.
- Type of instrument: Type of technology used defines the accuracy of the GCP measurement. The most common options are an RTK or DGPS survey.
- Overlap points: For the data from different flight plans to merge effectively, there should be GCP points in the overlap area between two batches.
Case Study: Planning Ground Control Points
GCPs are planned based on the decided patches within the area of interest. There are two important criteria: The GCPs should be assigned along the edges of the plan and should be evenly distributed across the area. The number of GCPs vary according to the elevation changes in the area of interest. You would need more GCPs wherever there are abrupt changes in elevation.
In this case, we applied a total of 84 GCPs including the ones in overlap areas in enhance accuracy. Patch-1 and 2 have four, Patch 1- and 3 have three and Patch-3 and 4 have four GCPs planned in the overlap area.
The good news is that we at DronaMaps have now done most of the hard work for you. We have an automated decision engine that helps us do robust Pre-flight planning. Do check it out before you embark on your next mission!
