What is LiDAR?

LiDAR stands for light detection and ranging. Standard RGB cameras, such as on most mobile phones or a DSLR, are passive and work by only receiving light. In other words, the lens opens, light comes in and an image is generated. LiDAR sensors are, however, active and transmit light which then bounces back to the sensor. The light transmitted from a LIDAR sensor is more commonly known as laser. The lasers transmitted from the LiDAR sensor are completely harmless. The LiDAR system works by measuring the time it takes for the lasers to bounce back from an object. The time taken is then used by the LiDAR to calculate the distance between the sensor and the object. This is better known as ‘time of flight’ measurements. At LiDAR Scotland we use LiDAR sensors to capture high quality LiDAR survey grade data for our clients. In this article we will explain LidAR, it’s application, benefits and limitations in surveying.

LiDAR survey, LiDAR sensor, LiDAR
Zenmuse L2 – LiDAR sensor

How does LiDAR work?

LiDAR sensors transmit 100,000’s+ laser beams per second. Each laser has the potential to hit an object and bounce back thus the sensor can detect 100,000’s to 1,000,000’s of unique points per second. Light travels at approximately 300,000kmph therefore, the time between transmission and collection of the lasers is very short. Some LiDAR sensors are able to receive more than 1 return per laser. A return is determined by how many objects an individual laser ‘hits’. Imagine a tree canopy. A single laser could potentially clip several leaves or branches after transmission. Sensors such as the L2 that we use can have a maximum of 5 returns whilst transmitting 240,000 lasers per second. If each laser were to hit 5 objects each, that would result in a phenomenal 1,200,000 points being captured per second. Each point will have it’s own location coordinates.

LiDAR survey image, LiDAR, LiDAR sensor
Cross section from a LiDAR survey

Note how the terrain is visible beneath the tree canopy

How does drone mounted LiDAR work?

In order for LiDAR data to be of suitable use, we need to be able to accurately and reliably correlate objects with a known grid system. This helps us understand where each point of an object is in the real world. In the UK, most surveys tie in with Ordnance Survey 1936. This is the most commonly accepted grid system in the UK and lets us determine how far North and East an object is as well as how high it is (Northing, Easting and Elevation). By understanding all 3 coordinates (often referred to as x, y and z) we can then compare the positions between 2 or more objects. This is essential for construction, environmental monitoring, change detection and other industries.

Traditionally, static LiDAR sensors would be set up at a known location. Thus, all laser contacts received by the sensor could be calculated knowing their relative xyz location compared to the sensor. In other words, if the laser bounces off an object exactly 1 metre north of the sensor and at the same height (z), we can calculate it’s position to be +1m North of the known location of the sensor. We could even create our own grid system where the sensor would be at 0, 0, 0. Every object detected would be +/- on all 3 coordinates subject to their relative location to the sensor.

Drone mounted LiDAR can’t use known locations as the drone is airborne and in motion. Thus the position of the sensor is dynamic. In order for the drone mounted LiDAR to be able to accurately measure the position of objects, the drone needs to be able to accurately know it’s own position. Without knowing the position of the drone, it is not possible to accurately identify the location of any object that the LiDAR sensor receives. There are a number of processes that can be followed to to achieve this accuracy. Each requires dedicated hardware and software and each works quite differently to the others. They key methods are GNSS, RTK, PPK and SLAM.

Why use LiDAR?

LiDAR Survey

Real Time Kinematics (RTK) and alternatives

At LiDAR Scotland we utilise the RTK (real time kinematics) workflow. RTK is a process where the drone connects with both the GNSS (global navigation satellite system) and a local NTRIP connection and must maintain this connection throughout the operation.

GNSS is often more commonly referred to as GPS (global positioning system). In reality GPS is a type of GNSS and is based on satellite technology which helps us identify where we are on the globe. GPS is common in our cars and on our mobile phones. It’s great at identifying where we are but it isn’t accurate enough for most technical purposes. GPS/GNSS can generally identify a location within approximately 2000mm horizontally and 3000mm vertically. Imagine a cylinder with a diameter of 2m horizontally and a height of 3 metres vertically. GNSS can accurately confirm that an object sits within an area of space that size. Where in that space does an object sit?? Well, GPS and GNSS cant answer that question.

Add RTK to the mix, and we can start to identify the objects position to +/-40mm horizontally and +/-50mm vertically. RTK works by tying in the accuracy of GNSS with a known fixed location on the earth’s surface. Combing the GNSS precision with the known fixed location, on the earth, helps us remove a whole range of environmental and distance factors that GNSS alone relies on. The fixed point on earth further triangulates our position to within +/-15mm on the z access and +/-10mm on the x and y axis. Hold on a second, we just mentioned 2 different levels of accuracy within this paragraph…. Let us explain more in the next.

Yes, with RTK, we can get the lower/better of the 2 accuracies when using rovers, positioned statically, to create a known or fixed location. We have to bear in mind that drones are airborne and mobile thus are subject to additional factors that can and will impact accuracy. With drone mounted RTK, we therefore can achieve accuracies to within +/-40mm horizontally and +/-50mm vertically.

The above mentioned NTRIP is normally a, paid for, subscription which as accessed via a mobile sim. The operator, in the field, uses the sim to connect to the NTRIP (networked transport of RTCM via internet protocol) and receive correction data from the fixed location (often referred to as a base station). This correction data is then used to correct the position from the sensor which has already been loosely determined via GNSS. And voila, we have our accuracy (although sounding simple, there is a bit more to it when operating in the field).

PPK is similar to RTK. Rather than real time, PPK relies on post processing the data. Hence, PPK stands for Post Processed Kinematics. In short, the data is processed in the office using additional software and specific data recorded during the flight. This process helps improve the accuracy. Whilst some operators prefer, PPK, at LiDAR Scotland we choose RTK. If the RTK does encounter errors, we can always fall back on PPK after the works are complete.

SLAM – simultaneous localisation and mapping. SLAM works by the sensor system using the correlation between points to accurately draw a 3D map of the surrounding environment. The sensor knows where it is by monitoring its position relative to the positions of it’s surroundings. How does the sensor know where it has moved to? Well, because it has tracked how the position of it’s surrounding environment has shifted in response to the movement of the sensor. This method of LiDAR is probably the most similar to how we, as people, know where we are in a room. When we walk around a room, some objects get closer and other objects get further away. From our eyes tracking the object(s) positions, we know we have moved and where we have moved to. SLAM is similar to this and works effectively indoors or in GPS denied environments. SLAM still needs to be told what it’s starting coordinate is before it can align an object detected by it’s LiDAR to a know grid system such as Ordnance Survey 1936. SLAM also suffers from drift. Drift is a by product which increases during the duration of the scan. Scan too long, and the data accuracy diminishes. To improve accuracy, SLAM can also utilise RTK.

LiDAR Scotland prefers RTK as the majority of our work is outdoors and RTK is, in our opinion, the most effective workflow to help deliver accurate and reliable outputs/deliverables.

What happens if RTK can’t connect to a mobile network or wifi?

RTK doesn’t always require an NTRIP connection (remember NTRIP requires internet access to connect to a known base station). The alternative is to set up a base station at a known location. As long as the base station and the drone can communicate with each other, you can use a base station to correct your GNSS location and the drone operates as a rover. There is one challenge with this and that is you need to have a surveyed known location in advance or an alternative method of creating a known location. At LiDAR Scotland we have several solutions to overcome this challenge.

Drone mounted LiDAR is the most effective tool available to collect high volume and high accuracy topographical or structural data. Unlike photogrammetry (the science of generating 3D environments from 2D RGB images), LiDAR only has to see a target once to record it’s xyz coordinates. As a by product, LiDAR can detect objects that are mostly hidden behind other objects. For example, ground below trees. The drone flying over a forest only requires small gaps in the foliage for the lasers to penetrate the canopy and record ground points. Modern LiDAR sensors, such as our L2, can capture upto 5 targets per laser. What this means is that each laser can hit 5 objects and record the xyz of each. LiDAR has been proven to be effective when the canopy hides as much as 95% of the ground.

LiDAR is also effective at detecting complex structures including fences, cables, pylons, and more. Drone mounted LiDAR is now commonly used for overhead utility inspections as well as planning installation routes for both overhead and underground utilities.

LiDAR is easier to process than photogrammetry. The data sets are generally smaller and thus require less processing power. LiDAR requires LiDAR survey specific software to process. It then requires post processing to convert the output (point cloud) into usable data that can be utilised for it’s intended purpose.

RGB image of forest

LiDAR survey image of floor of forest above

Summary

Drone mounted LiDAR is the most effective tool at delivering high precision survey grade holistic data. Outsight report that LiDAR is changing construction operations and safety. Drone mounted LiDAR is now accepted as the one of the leading tools to capture high detail and high accuracy data. Topographic, BIM and digital twin projects can now be facilitated with fast turnarounds and affordable deliverables. Learn more about how LiDAR Scotland (a division of Drone Scotland Limited) can help support your projects….. contact us at info@lidarscotland.com or 0141 302 4685

Leave a Reply

Your email address will not be published. Required fields are marked *