# Mapping a Forest with LiDAR

How airborne LiDAR maps a forest: timing laser pulses, multiple returns, the point cloud, finding the ground and the canopy height model

By Jacob Mooney

Topic: Earth & Geography

[Open the interactive Jig](https://ask.jig.com/j/QBEgDtElndlO)

## Transcript

### 1. Mapping a forest with LiDAR

LiDAR stands for Light Detection And Ranging\. A drone or plane fires rapid laser pulses at the forest below and times how long each one takes to come back\. From those times we can build a 3D map of every tree and of the ground hidden beneath them\. Drag anywhere to look around the model\.

LiDAR uses light, not radio waves, so it is not radar, and it measures distance rather than taking a picture\.

### 2. Timing one pulse

Watch one pulse leave the scanner, hit a treetop and bounce back\. Light travels at about 300,000 km per second, so the distance is d = c × t ÷ 2: we halve it because the pulse goes down and back up\.

A pulse from 110 m up returns in about 0\.73 microseconds \(a microsecond is a millionth of a second\)\. The laser is near\-infrared, invisible to us; it is drawn in red here\.

### 3. One pulse, many returns

The laser beam spreads into a small circle, its footprint, so part of it can reach through gaps in the leaves\. Each surface it touches reflects an echo called a return\. The first return comes from the top of the canopy and the last usually comes from the ground\.

Yellow: first return\. Orange: intermediate return\. Blue: last return\. The last return is not always the ground; dense foliage can stop the pulse first\.

### 4. Scanning the forest

The drone flies straight, overlapping lines while a mirror sweeps the pulses from side to side\. GPS gives the drone's position and an inertial measurement unit \(IMU\) gives its tilt, so every return gets an exact x, y and z\.

Real scanners fire hundreds of thousands of pulses per second\. This model shows only about 18,000 of them\.

### 5. The point cloud

Hide the real forest and what remains is the point cloud: a set of points in 3D space, one for every return\. Coloured by height, it already shows the shape of each crown, the gaps between trees and the slope of the hill\.

Compare it with the real forest: switch the trees back on and see how the points trace their crowns\.

### 6. Finding the ground

Software now classifies each point as ground \(brown\) or vegetation \(green\)\. Pulse density is the number of pulses per square metre\. Predict what happens to the ground under the thick canopy when you lower the density, then try it\.

The US national standard \(USGS Quality Level 2\) asks for at least 2 points per square metre\. Dense forest needs more, because few pulses reach the ground\.

### 7. Bare earth: the DTM

Joining the ground points into a smooth surface gives a digital terrain model \(DTM\), the bare earth with every tree removed\. What do you think is hidden under the canopy? Strip the trees away and look closely\.

An old logging road and a stream channel appear\. LiDAR has revealed hidden roads, landslides and even ancient ruins under forests\.

### 8. Canopy height: CHM = DSM − DTM

A digital surface model \(DSM\) follows the highest returns, the top of the canopy\. Subtract the bare\-earth DTM and the hill disappears, leaving the canopy height model \(CHM\): how tall the forest stands everywhere\. Foresters use it to count trees and estimate timber and stored carbon\.

Colours on the CHM run from 0 m \(dark blue\) to the tallest crown, about 32 m \(red\)\. Low pulse density can make trees look shorter than they are\.