A plane flies over the Gaspé Peninsula at 300 km/h. Below it, an instrument fires a laser beam at the ground — then another, then a hundred thousand more every second. Each flash sets off, strikes a branch, a roof, a rock, and comes back. By measuring that round trip, the device works out a distance. Repeat the operation billions of times and you obtain the three-dimensional portrait of a territory, tree by tree.
This is the principle of LiDAR — Light Detection and Ranging. Unlike an aerial photograph, which records colours, LiDAR records distances. And that difference changes everything: it doesn't just see the surface, it measures it.
The principle: timing light
Light travels at nearly 300,000 km/s. A LiDAR sensor emits a pulse, then measures the delay before the echo returns — a few millionths of a second (at 300 m altitude, the round trip takes about 2 microseconds), timed to the nanosecond. Since the speed of light is known, that delay translates directly into distance: distance = (speed × time) / 2. The division by two accounts for the round trip.
On its own, a distance isn't enough. To know where the point struck lies, the instrument must know its own exact position and orientation at each shot. That is the job of two quiet companions: a GNSS receiver (precision GPS) that gives the plane's position, and an inertial measurement unit (IMU) that tracks its every pitch, roll and yaw. By combining the measured distance, the plane's position and the beam's orientation, each echo becomes a point with known coordinates.
Millions of points: the cloud
A survey does not produce an image, but a point cloud — a mass of (x, y, z) coordinates floating in space. On Québec surveys, density commonly reaches several points per square metre. Zoom into a point cloud and you can make out power lines, roof ridges, the trunks of isolated trees.
Each point also carries an intensity: the fraction of light returned. Asphalt pavement, water and a sheet-metal roof do not reflect in the same way. This information later helps distinguish surfaces.
The trick that creates all the value: multiple returns
Here is what sets LiDAR apart from almost everything else. A single laser pulse is not infinitely fine: as it passes through a tree, part of its light bounces off the upper leaves, another off lower branches, and a last part reaches the ground before travelling back up. The instrument records several echoes for a single shot — first return, intermediate returns, last return.
The first return traces the top of the forest. The last return slips all the way down to the ground. No aerial photo can do that.
This ability to “see beneath the trees” is why LiDAR upended relief mapping. Where conventional imagery is stopped by the forest canopy, the laser finds the gaps. It is also what makes unexpected uses possible, from measuring the forest to discovering archaeological sites hidden under the canopy.
The ground of Québec, to the metre
Québec's LiDAR acquisition program, led by the Ministry of Natural Resources and Forests, progressively covers the inhabited and developed territory. The vertical accuracy of the surveys varies with the terrain — on the order of a few centimetres in open ground to a few tens of centimetres under forest cover — and the derived terrain models are delivered at a resolution of 1 metre — each one-metre cell carries an elevation.
At that fineness, you can make out a drainage ditch, a one-metre bank, the old course of a logging road. It is precisely this data — open and distributed on Données Québec — that Québec 3D turns into navigable relief.
See for yourself
Once cleaned and placed on a grid, the point cloud becomes a terrain model. Pick a hill, a valley or a quarry on the map and watch the relief rise into volume.
Explore an area in 3D →The takeaway
LiDAR is not a sophisticated camera: it is a rangefinder that repeats a distance measurement billions of times, keeping an exact record of where and how each measurement was taken. Its strength rests on an almost mundane idea — timing light — and on one decisive detail: a single shot can bring back several echoes, including the one from the ground beneath the trees.
One question remains: how do you go from that raw cloud, where roofs, foliage and terrain are mixed together, to a clean map of the bare ground? That is the whole craft of the geomatics pipeline.