Technologies
LIDAR
For a long time our greatest challenge was posed by densely vegetated areas.
Now, using our latest technology’s active procedure, we can measure points and therefore carry out surveys in a fast, precise and comprehensive manner even with heavy vegetation.
What is Light Detection and Ranging (LiDAR)?
LiDAR is fundamentally a distance technology. From an airplane or helicopter, LiDAR systems actively sends light energy to the ground. This pulse hits the ground and returns to the sensor.
Basically, it measures how long it takes for the emitted light to return back to the sensor. In the end, it gets a variable distance to the Earth.
Actually, this is how LiDAR got its name – Light Detection and Ranging.
But let’s dissect LiDAR a little more. For example, what does a LiDAR system generate? What are LiDAR applications in GIS?
What outputs can LiDAR generate?
LiDAR is active remote sensing. This means the LiDAR system sends a pulse of light and it waits for the pulse to return. This is different than passive sensors which collects reflected energy originating from the sun. Active sensors are very accurate because it’s being controlled in the platform.
LiDAR is a sampling tool. What I mean by that is that it has the brute force to send 160,000 pulses per second. In a LiDAR point cloud, it creates millions of points. Usually, the point density is less than one meter with accuracy of about 15 cm vertically and 40 cm horizontally.
A LiDAR unit scans the ground from side to side as the plane flies because this covers a larger area. While some pulses will be directly at nadir, most pulses travel at an angle (off-nadir). When a LiDAR system calculates elevation, it needs to accounts for angle.
Number of Returns
Imagine you’re hiking in a forest. You look up. If you can see light, this means that LiDAR pulses can go through too. Also, this means that LiDAR can also hit the bare Earth or short vegetation. A significant amount of the LIDAR energy can penetrate the forest canopy just like sunlight.
But LiDAR won’t necessarily only hit the bare ground. In a forested area, it can reflect off different parts of the forest until the pulse finally hits the ground:
Using a LiDAR to get bare ground points, you’re not x-raying through vegetation. You’re really peering through the gaps in the leaves. LiDAR collects a massive number of points.
These multiple hits of the branches is the number of returns.
In a forest, the laser pulse goes down. We get reflections from different parts of the forest – 1st, 2nd, 3rd returns until it finally hits the bare ground. If there’s no forest in the way, it will just hit the surface.
Sometimes a pulse of light doesn’t reflect off one thing. As with the case of trees, one light pulse could have multiple returns. LiDAR systems can record information starting from the top of the canopy through the canopy all the way to the ground. This makes LiDAR highly valuable for understanding forest structure and shape of the trees.
Digital Elevation Models
How do you build a Digital Elevation Model from LiDAR?
Digital Elevation Models are bare earth (topology) models of the Earth’s surface. You can derive Digital Elevation Models (or Digital Terrain Models) by using the ground hits from LiDAR. Ground hits are the last return of the LiDAR.
Sometimes the last return may not even make it to the bare ground. But for LiDAR, this is more rare than you think.
Which points are ground hits? There are ways to filter the LiDAR points. Take the ground hits (topology only) meaning the last returns from LiDAR.
Filter last return ground points. Then, interpolate the points. Finally, build your DEM.
With a DEM, you can generate products like slope (rise or fall expressed in degrees or percent), aspect (slope direction) and hillshade (shaded relief considering illumination angle) maps.
Canopy Height Model (CHM)
Light detection and ranging attains very accurate information about the ground surface. We can also get very accurate information about what’s on top of the ground with a Digital Surface Model (DSM).
A Canopy Height Models (Normalized Digital Surface Model (nDSM)) gives you true height of topological features on the ground.
So how do you get true height of features on the Earth?
Take the first return including topology (tree, building). Subtract the last return which are the ground hits (bare Earth).
Light Intensity
The strength of LiDAR returns varies with the composition of the surface object reflecting the return. The reflective percentages are referred to as LiDAR intensity.
But a number of factors affect light intensity. Range, incident angle, beam, receiver and surface composition (especially) influences light intensity. When the pulse is tilted further away, the return energy decreases.
Light intensity is particularly useful in distinguishing features in land use/cover. For example, impervious surfaces stand out in light intensity images. Object-based image classification segmentation can separate these features using light intensity values.
Point Classification
LiDAR data sets may already be classified by the vendor with a point classification. The codes are generated by the reflected laser pulse in a semi-automatic way.
Not all vendors add this LAS classification field. Actually, it is usually agreed in the contract beforehand.
The American Society for Photogrammetry and Remote Sensing (ASPRS) has defined a list of classification codes for LiDAR. For example, classes include ground, vegetation (low, medium and high), building, water, unassigned, etc.
Point classification may fall into more than one category. If this is the case, these points are usually flagged and have secondary classes.