Aerial robotics
Inspection drones and the evidence needed to identify a defect
Relate image sampling, viewpoint and sensor choice to the inspection question, while preserving the limits of confined-space systems.
An inspection drone places a sensor where an inspector needs evidence. The requirement might be a visible surface feature, a temperature pattern or a measurement from a contact instrument. The aircraft's ability to reach the location should be assessed separately from the sensor's ability to answer the question.
Read a confined-space system in its documented configuration
Flyability describes Elios 3 with a protective cage, camera payload and LiDAR-assisted mapping and stabilization. Its documentation also describes optional payloads and software used to locate inspection observations. The aircraft, sensor and processing configuration should remain together in a report. [1]
The product page includes a specific ingress-protection qualification. It says the aircraft was designed toward IP44 requirements but is not certified IP44. Do not shorten that statement to IP44-certified. A protective cage likewise does not establish suitability for every confined or hazardous environment. [1]
Pixels across a feature are a useful first calculation
Consider an idealized image covering a flat, front-facing surface 6 m wide using 4,000 horizontal pixels. The nominal sampling is 1.5 mm per pixel. A 3 mm-wide feature spans only two pixels in that simple geometry.
This calculation does not establish reliable detection. Blur, focus, lighting, contrast and viewing angle can reduce the information available. A projected pixel count should be followed by a task-specific assessment using representative features and the approved capture arrangement.
Increasing the number of recorded pixels also does not prove an improvement when the image is out of focus or the target is obscured. Keep the original files and the acquisition conditions, rather than relying only on compressed screenshots.
Scroll sideways for all columns.
| Question | Evidence required | Unsupported shortcut to avoid |
|---|---|---|
| Is a surface feature visible? | Image quality at the relevant size and viewpoint | Treating camera resolution as a detection guarantee |
| Is a thermal pattern meaningful? | Thermal record and interpretation conditions | Reading an ordinary color image as a temperature measurement |
| What is the material thickness? | A suitable documented measurement method | Inferring thickness from a clear video alone |
| Where is the observation? | Traceable position or asset reference | A close-up without a location record |
Map the surfaces that were not observed
For an illustrative tank inspection, divide the required surfaces into identified regions before the flight. Mark each as inspected with acceptable evidence, observed without adequate detail, or not observed. This is more informative than marking the whole tank complete after a drone entered it.
Retain the reason for each gap. An obscured surface, an inaccessible angle and a missing position reference require different follow-up work. The inspection team can then plan another method without repeating the entire survey blindly.
A useful finding includes the original image or measurement, its asset location, acquisition time and any interpretation limits. The flight route should support that record, not replace it.
Plan retrieval before an inaccessible flight
The operating team must assess the approved flight envelope, the environment and the method of retrieving a stopped aircraft. A confined location can make a minor failure difficult to recover from. Do not enter a tank or other hazardous space to retrieve equipment outside the site's approved procedure.
Use the manufacturer's stated behavior for the exact control and communication configuration. A mapping function does not establish autonomous escape from every obstruction or every lost connection. This guide provides an evidence-review method and does not validate a confined-space operation.
Sources and scope
Sources were consulted for this revision. Manufacturer descriptions are identified in the text. Worked examples are illustrative calculations, not measurements from a TechniaHQRobot test.
- Flyability Elios 3 configuration and qualifications
Manufacturer description of the platform, mapping and payloads, including its explicit IP44 certification qualification. No independent defect-detection rate is claimed.