3 cm smouldering char
At 50 mm EFL, F/5 is only around the advisory gate; F/3.2 gives stronger margin. This is the cleanest first survey-design driver.
Mission decision packet ยท 14 June 2026
Choose the camera role before the next prescription.
The latest radiometry trade shows the current 200 mm prescription is a compact spotter, not the one-hour area-survey camera. AAO should now choose the operating role, target class, fleet convention, and detector branch that the next optical design must satisfy.
Decision 1
Decide whether EmberScope is an area-survey camera or a cue-following spotter.
| Option | What it means | Evidence from the current model | Design consequence |
|---|---|---|---|
| Area survey | Optimize around short EFL and wider field so the camera covers ground fast enough for revisit planning. | At 120 m altitude, 25 mm EFL gives about 8.2 cm GSD, 52 m swath, and about 47 min single-drone revisit for 100 ha. A 50 mm EFL gives about 4.1 cm GSD, 26 m swath, and about 95 min single-drone revisit. | The next optics target moves toward roughly 25-50 mm EFL, fast LWIR optics, strong off-axis correction, and stray-light control. |
| Spotter | Keep the long EFL so the camera inspects small targets at fine plate scale after a coarse cue. | The committed 200 mm EFL gives about 1 cm GSD and 7 m swath; the 3 cm smouldering-char target overfills a pixel, so F/5 has margin in the model. | The existing prescription remains directionally useful, but it cannot be presented as the one-hour area-survey solution. |
Decision 2
Choose the faintest target class that must drive the design.
1 cm ember-class hot target
The 50 mm survey case needs faster optics than F/5 for comfortable margin. This is plausible as a stress case for the first prototype.
1 cm weak warm edge
The model reaches advisory margin only near F/1.6 at the finer 50 mm survey plate scale, and drops out at coverage-greedy plate scale.
Recommended default
Optimize for 3 cm smouldering-char detection first, then report the 1 cm ember and weak-warm-edge margins as stress cases.
Decision 3
Close the fleet and revisit convention before treating FOV as fixed.
| Question | Current working value | Why it matters |
|---|---|---|
| Is one-hour revisit per drone, per small team, or per fleet? | The requirements update names one-hour revisit but does not close fleet count. | EFL and FOV change materially if a 3-drone or 10-drone fleet is allowed to meet the revisit target. |
| Does AUD 100,000 cover all payloads, spares, calibration kit, and integration? | Current assumption says yes, excluding drones. | This controls whether detector choice and custom optics can scale beyond one prototype. |
| Which altitude and speed should be the operating reference? | The trade used 120 m and 12 m/s for comparability with prior survey models. | Altitude and speed directly set GSD, swath, dwell, atmospheric path, and route spacing. |
| How many hectares need one-hour revisit? | The 15 June sizing grid now translates target area into payload count and max unit cost. | A 500 ha hourly block needs either three payloads at 120 m or two payloads with altitude authorisation; 1000 ha pushes below the current custom cost floor. |
| Can a COTS or camera-only payload replace custom optics for find-only survey? | The 16 June COTS payload screen shows some fast 640-class paths clear the first-order 3 cm target physics gate at 120 m. | The open gate moves to exact product evidence: raw/pre-AGC access, calibration state, mass, power, platform lock-in, false positives, and quoted unit cost. The 20 June quote packet converts those gates into a vendor evidence request, the 21 June quote comparison register defines the scoring needed before a COTS response can close the branch, and the 22 June evidence ledger tracks which answers and sample bundles are still missing. |
Decision 4
Pick the detector branch tightly enough to set the optical target.
The current models use a 640 x 512, 17 um-pitch uncooled LWIR detector reference and keep Boson+ 640 radiometric as the preferred custom-optics branch. That is enough for first-order plate-scale and F-number work, but not for a fabrication package.
The next pass should confirm the exact detector core, NETD and specified F-number, radiometric or raw/pre-AGC access, frame rate, interface, power, thermal path, and whether any alternate detector changes the fleet-cost or F-number trade.
Recommended answer set
Use this as the next optimization target if AAO wants a survey-camera prescription.
| Field | Recommended answer for next optimization |
|---|---|
| Operating role | Area survey camera, not a narrow-field spotter. |
| Target class | 3 cm smouldering-char detection as the design driver; 1 cm ember and weak warm edge as stress cases. |
| Reference altitude and speed | 120 m altitude and 12 m/s until customer operations specify otherwise. |
| Plate-scale family | 25-50 mm EFL survey family, with the final value chosen by target-class and fleet-count trade. |
| F-number family | F/3.2 or faster for the first survey prescription; report F/2 and F/1.6 variants if feasible. |
| Detector reference | 640 x 512, 17 um uncooled LWIR radiometric branch until a specific core is selected. |
| Fleet convention | State explicitly whether one, three, or ten drones are allowed to satisfy the one-hour revisit target. |
Effect on open work
The brief narrows the next question without pretending the requirement is closed.
RFS-011
Remains open because this is a decision packet, not stakeholder sign-off.
RFS-019
Remains blocked until survey-vs-spotter and target-class answers are chosen.
RFS-028
Should not spend more effort on mirror-count seeds unless the search includes the selected plate scale, F-number, and obstruction-free propagation gates.
Source evidence
See the 9 June expert requirements update and the 14 June F-number / aperture trade for the model basis.