From flight to pipeline
A data team sees the last step: a dataset lands in the bucket. This is everything that happens before it, and what each step protects.
1. The specification
Everything starts with four numbers and a format: the area, the ground resolution in centimetres per pixel, the overlap between images, and the capture window. Then the deliverables, such as raw imagery, orthomosaic, elevation model or point cloud, the reference system your maps use, and where the data should land. Getting this right in the scoping call is what makes the rest predictable.
2. The flight plan and the permits
The resolution sets the flying height. The overlap sets the spacing of the flight lines. The area and the aircraft's speed set the hours. From that we build the flight lines, file the plan and obtain the permits with a certified local operator. The plan also fixes the time of day, because the sun angle changes shadows and, with them, what a model learns.
3. The capture, watched in real time
During the flight the operator's pilot flies the lines while our team runs the sensors and the mobile operations base on the ground. The base has its own connectivity, so coverage is checked as it happens, even in remote areas with no infrastructure. A gap or a bad line is caught and re-flown while the crew and the aircraft are still on site, not discovered a week later in the office.
4. Field quality control
Before anyone leaves the site we check that every image exists, that exposure and sharpness are within tolerance, that the positioning logs are complete, and that the overlap actually achieved matches the plan. This is the step that makes a 100% delivery rate possible: problems are cheap to fix while the aircraft is still there.
5. Processing
Photogrammetry turns overlapping images into geometry. Image orientation recovers where each photo was taken, ground control points tie the block to the reference system where the project requires survey-grade accuracy, and from that block we generate the orthomosaic, the elevation and terrain models, the 3D model or the point cloud. If your product needs the raw imagery instead, we deliver it organised, georeferenced and indexed.
6. Validation
We check the products against the specification: coverage, resolution, positional accuracy, seams and artefacts. What fails is reprocessed or re-flown. What passes is documented, so the dataset arrives with a record of what was checked.
7. Delivery and traceability
Datasets are delivered in the format and reference system you use, such as GeoTIFF or cloud-optimised GeoTIFF for rasters, LAS or LAZ for point clouds, and tiles where a map application needs them, directly to your cloud, servers or systems. Each mission carries one identifier that ties the flight, the technical record and the invoice together, so provenance is documented from the flight line to the row in your database.
What this means for a data team
You never handle aviation logistics, and you never receive a surprise. The specification you approved is the specification that lands, with the evidence of how it was checked. That is the difference between buying imagery and buying an operation.
Want the dataset to land in your pipeline, not in an inbox?
Tell us your format, reference system and destination. We scope the mission in one call.
