What Are Point Clouds?
For years now, SurvTech Solutions has been creating and processing point cloud data sets in a wide variety of environments, with multiple applications in mind. But just what are point clouds, how do they work, and what is SurvTech doing with them?
Point clouds are a dense compilation of many (sometimes millions of) points, representing the x, y, and z of geospatial data, where each point represents a specific, 3-dimenional location, in relation to the other points in the cloud. These clouds can be captured from multiple types of data collection technologies and give users the ability to create a wide range of application specific products. Next, we will explore a couple of the technologies that SurvTech utilizes on a regular basis.
One of the most common platforms for data collection is called ‘LiDAR’, which stands for Light Detection And Ranging, and is a play on the word RADAR. SurvTech uses LiDAR in both aerial and ground-based applications. For Aerial acquisition, we use unmanned and fixed-wing (airplanes) aerial vehicles to collect topographic data sets ranging from 100’s of acres to 1000’s of square miles; from narrow utility corridors to wide area approaches; all while pulsing out up to 1 million laser pulses per second. In addition to their accuracy and consistency, one major strong point of most LiDAR systems is the ability to receive multiple returns per pulse. This is an advantage in vegetated or obstructed areas, as it will increase the chances of collecting valid ground locations in such areas. Stationary, ground based, projects are conducted with LiDAR, as this technique is referred to as terrestrial LiDAR or 3D scanning. With the LiDAR unit being static, the LiDAR returns are extremely accurate, in the 1-millimeter range. Typically, terrestrial LiDAR is utilized to measure very intricate 3D structures, like buildings, or industrial plants, where the structure is comprised of structural steel (beams and columns), piping, conduit, and industrial equipment. Basically, if you look at any 3D structure and think to yourself, this would be impossible to measure, then terrestrial LiDAR is the method to be used. It is used extensively in the oil and gas, power, and mining industries. Similarly, recent advances in scanning technology have produced the non-GPS dependent Indoor Mobile scanning systems, such as the NavVis M6, utilizing simultaneous localization and mapping (SLAM) algorithm to achieve relative accuracy (how well the data fits itself in overlapping scan areas) in the centimeter range, which allows the user to collect point cloud data indoors or in other non-suitable to GPS locations. Using recent advancements in data processing and dissemination and when paired with 360° imagery, we are now able to produce a “digital twin” of scanned locations. The digital twin can then be accessed and utilized anywhere around the globe, so that planning, maintenance, and design decisions can be made remotely. Another methodology of scanning, would include Mobile LiDAR, to feature a scanning system (LiDAR, GPS, IMU) deployed on trucks, cars, ATV’s, hydrographic survey vessels (boats), or even backpacks, to collect large amounts of data from the ground or water level.
Another source of point cloud data is from multibeam sonar (sound) technology, rather than light. Multibeam data is the basis upon which our hydrographic team collects and analyzes 3D data from below the water’s surface. In the same way that lasers are used to pinpoint geospatial locations, sonar uses sound attenuation to measure distance to objects in water by measuring the time it takes for an echo to return; therefore, a system that collects sonar data is known as an echo sounder. The multibeam echo sounder incorporates reading from multiple “beams” simultaneous to collect data in a swath under the scanning vessel. The deeper the water, the wider the swath, as the sonar fans out. The multibeam sonar is also equipped with an inertial measurement unit (IMU), just like the vehicles that collect mobile LiDAR data. The IMU tracks the motion of the boat (heave, pitch, and roll) to properly position the sonar head thousands of times per second.
As we have hinted at already, both LiDAR and sonar must be georeferenced to accurately depict the geospatial dimensions of the scanned surface. For this, we incorporate GPS coordinates that have real time corrections, such as Real-Time Kinematic (RTK) or post processed corrections (PPK). These methods are accomplished by having a survey grade GPS “rover” receiving corrections from an RTK base station (a second GPS receiver setup on a known location) setup onsite that is transmitted by radio or the base station collecting the data to a file to be post processed in the office. Another form of RTK correction is performed by using a cellular network to connect to a base station that many miles away but is broadcasting corrections via the internet. TheNetworked Transport of RTCM via Internet Protocol (NTRIP) enables access to multiple permanent reference stations over the internet and can achieve 1cm accuracy.
The resulting point clouds (datasets) collected from these various technologies and platforms can be combined into one seamless 3D model. By georeferencing all the data collected, the resulting point clouds can be seamlessly combined. The seamless point clouds can then be used for numerous final deliverables, such as topographic and bathymetric maps, digital elevation models (DEMs), digital terrain models (DTMs), 3D CAD and Building Information Models (BIM), point cloud and imagery viewers, and even animation and virtual reality. Point clouds are a major leap for our society transitioning many services from the real world to the virtual world.