What LiDAR gives you that a camera cannot

A camera measures color; a LiDAR measures distance. That difference produces three practical advantages. It works in darkness and in harsh light, because it supplies its own illumination. It penetrates vegetation — some pulses reach the ground between the leaves, producing a terrain model under a forest canopy that no photograph can reconstruct. And it measures directly rather than inferring geometry from overlapping images. In exchange it is heavier, more expensive, and returns no color. Broader payload and airframe context lives in our drone development hub.

LiDAR vs photogrammetry

Both produce a 3D model, and the choice between them is practical rather than ideological. Photogrammetry is cheap, gives faithful texture and color, and excels over open ground and textured surfaces. LiDAR wins under vegetation, on thin objects such as conductors, cables, and railings, in difficult lighting, and wherever a client needs a direct measurement instead of a reconstruction. Many professional systems carry both and fuse the results: geometry from the scanner, color from the camera. Payload selection generally follows the same logic as choosing any other sensor package — see thermal camera drone payloads for the parallel case.

What actually determines accuracy

The common mistake is to equate system accuracy with sensor accuracy. In reality every point in the cloud is the sum of three measurements, and each contributes error:

  • The range measurement from the scanner itself.
  • The position of the aircraft at the instant of that measurement.
  • The attitude — the orientation of the sensor at that same instant.

The second and third usually dominate. That is why a serious mapping system includes a survey-grade GNSS receiver with RTK correction from a base station or post-processed kinematic workflow, and a high-quality inertial measurement unit rigidly coupled to the scanner. It is also why boresight calibration — the precisely known transform between the IMU and the scanner — is part of commissioning, and why ground control points still earn their keep on jobs where the deliverable has legal or contractual weight. The navigation sensor background is in selecting a drone flight controller.

Time synchronization is not a detail

If the scanner's timestamps and the navigation solution's timestamps are offset by a few milliseconds, the cloud smears and no amount of post-processing recovers it. Hardware-level synchronization, usually driven off the GNSS pulse-per-second signal, is a small engineering item that determines the quality of the entire product.

Processing the point cloud

The raw cloud is not a deliverable. The chain runs roughly: strip spurious returns from dust, rain, and multipath; classify points into ground, vegetation, structures, and linear features; build a digital terrain model; then extract what the customer actually pays for — contours, cross sections, stockpile volumes, clearance reports, building models. That last step is the commercial one. A company selling deliverables sits in a much stronger business position than one selling scans, because the deliverable is what the client can hand to an engineer or a regulator.

Applications that justify the price

  • Volumetrics. Stockpiles at a quarry, cut and fill on a construction site — places where accuracy converts directly into dollars on an invoice.
  • Utility corridors. Measuring conductor clearance to vegetation and ground. This is the application where LiDAR's advantage on thin objects is decisive, and it usually runs as a long linear mission, which raises the operational questions in BVLOS drone operations.
  • Forestry and topography. Terrain models beneath canopy, plus canopy height and biomass estimates. Adjacent to the sensing work described in agricultural drones and precision farming.
  • Archaeology and land planning. Revealing structures and earthworks hidden under vegetation.
  • Navigation. A lightweight scanner used as a perception sensor rather than a survey instrument, for obstacle avoidance and flight in GPS-denied spaces.

Integration considerations

A scanning payload is heavy and power-hungry, so it drives the weight budget and cuts endurance — which is why survey aircraft in this class are usually larger multirotors or hybrid platforms rather than compact quads. The tradeoff between configurations is laid out in fixed-wing VTOL vs multirotor. Three more items decide whether the system performs:

  • Rigid mounting. Any flex between sensor and IMU becomes error directly. This argues against soft isolation mounts and for a stiff, well-characterized bracket, with vibration handled at the airframe level instead.
  • Data volume. A single flight produces gigabytes. Plan fast onboard media, field offload, and backup before the crew leaves the site — losing a dataset means reflying the job.
  • Operations and cost. US commercial flights fall under FAA Part 107, with corridor and long-range work needing additional authorization. Overall program budgeting for a custom aircraft is discussed in what drone development costs.

Integrating a scanning payload, or building an aircraft around one? Projects House can size the platform, design the mount and synchronization chain, and work through the accuracy budget with you. Get in touch through the contact form.