The Clock Is the Design Constraint

In wilderness search and rescue, survival probability for a lost hiker drops sharply after the first twelve to twenty-four hours, and in cold water or avalanche burial the useful window is measured in minutes. Every engineering decision in a SAR aircraft traces back to that clock: how fast the system gets airborne, how much ground it covers per battery, and how quickly the operator gets a usable answer rather than a folder of images to review later.

That is a different design brief from a mapping or inspection drone. A survey aircraft optimizes for accuracy and repeatability. A SAR aircraft optimizes for time to first detection under conditions that are, by definition, bad: night, wind, rain, smoke, terrain that blocks radio, and an operator who has been awake for nineteen hours.

Detection Is the Whole Product

Thermal imaging is the primary sensor, and its real-world limits are routinely oversold. A 640x512 uncooled microbolometer with a 25 mm lens gives a human-sized target roughly 2 to 3 pixels of vertical extent at 400 ft (120 m) altitude, which is enough to notice as an anomaly but not enough to classify. Detection range collapses further in three common conditions: wet vegetation that has equalized with body temperature, mid-afternoon sun that heats rocks and bare soil above 98°F, and dense canopy that simply blocks line of sight. The practical answer is lower and slower flight with tighter transect spacing, which costs endurance. Sensor selection and integration tradeoffs are worked through in choosing and integrating a thermal payload.

Visible-light cameras still matter, because bright synthetic fabric is often the strongest signal available in daylight and because a rescue team wants an RGB frame before it commits people to a slope. Serious systems fly both, either as a dual-sensor gimbal or as two aircraft on complementary patterns.

The second detection channel worth building is RF. Many missing people carry a phone, and a lightweight receiver that logs cellular or Wi-Fi probe signatures with GPS position can narrow a search area faster than any camera. This raises legitimate legal and privacy questions in the United States and needs counsel before it ships, but the physics work and agencies ask for it.

Onboard detection assistance is now expected. Running a person-detection model on the aircraft or at the ground station and flagging candidate frames in real time changes the operational model from post-flight review to live alerting. Reviewing 4,000 thermal frames after landing is how a find gets missed at 3 a.m.

Not Just Looking, Also Helping

Once a subject is located, the aircraft can do more than orbit. Practical payload deliveries in the 1 to 5 lb (0.5 to 2.3 kg) class include a radio, a water bottle, a thermal blanket, a strobe, or an inflatable flotation device dropped upstream of a swimmer. Each one needs a release mechanism that will not misfire in flight, will not freeze, and can be armed with gloves on. A two-stage electromechanical release with a mechanical safety pin is the usual answer.

A loudspeaker and a searchlight turn the aircraft into a two-way node. Being able to tell a subject to stay put, and to give ground teams a light to walk toward, is often worth more than another sensor. Both are power-hungry, and their draw has to be in the energy budget from the start rather than bolted on. Pack sizing, sag under pulsed loads, and cold-weather capacity loss are covered in drone battery systems. Expect 20 to 40 percent capacity loss at 14°F (-10°C) unless the pack is insulated or heated.

Requirements the Mission Forces on You

  • Weather. Real deployments happen in the storm, not after it. Target IP44 as a floor and IP54 if the budget allows, with conformal-coated boards and drainage paths rather than sealed cavities that trap condensation.
  • Wind. A useful aircraft holds position in 25 mph (40 km/h) sustained wind with gusts to 35. That drives disk loading, motor headroom, and control authority, not just marketing copy.
  • Endurance and coverage. Quote the number that matters, which is acres per battery at detection altitude with the payload installed, not hover time on a bare airframe. Where endurance dominates, a fixed-wing VTOL covers far more ground per charge, and the airframe tradeoff is laid out in fixed-wing VTOL vs multirotor.
  • Navigation without GPS. Canyons, dense forest, and collapsed structures degrade satellite positioning exactly where searches happen. Visual-inertial odometry and terrain-following are covered in GPS-denied navigation.
  • Ready in under five minutes. No tools, no calibration dance, no laptop. Props that stay on, a single battery latch, and a startup sequence that completes while the team is still pulling on harnesses.
  • Beyond visual line of sight. Meaningful search areas exceed what an operator can see. The waiver path and the airworthiness evidence it requires are described in BVLOS operations and FAA waivers.

Working Alongside Everyone Else

A SAR drone never operates alone. Manned helicopters may be working the same grid, other agencies may be flying their own aircraft, and incident command needs a shared picture. That means an output format their software already reads, position and imagery pushed to the incident map rather than saved to an SD card, and a strict altitude deconfliction protocol written into the operating procedure. When the aircraft has to stay up continuously over a fixed site such as a collapse or a river crossing, a tethered system removes the battery swap entirely.

Who Actually Buys

Volunteer SAR teams have the need and almost no budget, typically under $15,000 per unit and often funded by grants or donations. County sheriffs, state emergency management agencies, fire departments, and coast guard units have real procurement authority, and they buy programs rather than aircraft: training, spares, service contracts, and documented reliability. Build for the funded buyer while keeping a stripped configuration the volunteer teams can afford, and expect a sales cycle of nine to eighteen months with a pilot deployment in the middle of it. More context on the technology stack sits on the drone development hub.

Build the Aircraft for the Bad Night

Projects House develops mission-specific UAS platforms end to end, from payload selection and airframe sizing through firmware, ground software, and certification testing. Tell us the mission profile and coverage target through our contact form and we will scope the aircraft against it.