Types of Drone Sensors
There is an incredible variety of sensors made for drones and the applications they’re used for. The drone sensors most surveyors are likely to consider fall into four general categories: built-in cameras, small independent cameras, high-end independent cameras and LiDAR. The first three options are different types of standard cameras that are used as photogrammetry tools, while a LiDAR sensor is effectively a laser scanner mounted to a drone. Each of these sensors has it merits and drawbacks. All four are viable to be used in an effective drone survey program with the right Standard Operating Procedures (SOPs) and processing workflow, though they differ in their real-world applications.
Built-in cameras are designed and built specifically for use in the drone they are attached to. They are completely integrated into the drone airframe and cannot be removed nor replaced without significant manual modification. While only a small number of drone manufacturers build first-party cameras, they have proven to be dominant enough as to deserve their own category. DJI is the largest manufacturer of these drones, with the Phantom series being the one most commonly used in survey applications (Note: the DJI X4S camera carried by the DJI Inspire and M200 series are functionally identical to the Phantom 4 Professional or Advanced camera). Built-in cameras offer as large as 20 megapixel sensors with global shutters (as opposed to rolling shutters, which can cause image distortion), and are most often carried on small multirotor airframes.
With the right field SOPs and data processing, built-in cameras can reliably produce survey data at better than 0.1’ accuracy. They are also inexpensive, with the DJI Phantom series costing $1,500 including the sensor. Given that they are built specifically for use in these aircrafts, these sensors are extremely simple and reliable. This means minimal maintenance, calibration, and downtime – and maximum return on investment (ROI).
Since they are built into the aircraft, however, these sensors are less flexible since they can’t be easily swapped out for alternate sensors. Because the image sensors are smaller than some alternatives, they need to be flown fairly low (~100’) to attain the 0.1’ accuracy, meaning their range is somewhat limited to approximately 25 acres/hour.
With their low cost, high reliability, and high accuracy, built-in sensors are usually the best bet for most surveyors focused on topographic and planimetric mapping on projects less than 250 acres.
Small Independent Cameras
Small independent cameras are third-party sensors mounted onto airframes either by the drone manufacturer or by aftermarket modification. They are often ~20 megapixel image sensors with global shutters. These are the most common option for fixed-wing aircraft, which have more weight limitations, as well as some custom-built small multirotors. Due to the mechanics of fixed-wings and the complexity of custom integration on small multirotors, these cameras are often not mounted on a gimbal – a device that allows the camera to move independently of the airframe.
The main benefit of these cameras is their ability to be used on fixed-wing airframes, which have longer ranges than multirotors. If used on a multirotor with a gimbal, they can produce equivalent accuracy to a built-in sensor. Also, since they are not integrated into the airframe, they can be swapped easier than a built-in sensor.
Because they are not built into the airframe, however, they often require more complex work, calibration and maintenance. Specifically, the camera shutter-trigger mechanism can be particularly challenging, resulting in unpredictable data collection failures. Whenever a camera is used without a gimbal, as is this case on nearly all fixed-wings, there are additional data quality issues. Without a gimbal, whenever the drone vibrates, turns or banks to fight wind gusts, images will be blurred, resulting in lower-accuracy data.
The best use case for a small independent camera is on a fixed-wing drone, when large acreages need to be covered at lower accuracy. If a large project only requires spot elevations sufficient for one-foot contours, this type of sensor is an excellent choice.
High-end Independent Cameras
High-end independent cameras are larger cameras developed for uses other than drone mapping (e.g., digital SLR cameras), which must be carried on large multirotor airframes that are designed to carry large sensors. These cameras can often have up to 40-megapixel sensors.
The primary benefit of these sensors is the high image resolution they can produce, which translates to lower (better) ground sampling distance in aerial imagery. This allows them to consistently produce survey data accurate to 0.1’ when used with appropriate field SOPs and data processing. Because of their higher resolution, they can achieve this accuracy at higher flight altitudes than built-in cameras, meaning they can cover slightly more ground and clear tall obstacles without sacrificing accuracy.
The main drawback of these sensors is complexity. They almost always require custom integration, thus being more prone to faults and requiring regular significant maintenance and calibration. The shutter-trigger integration is often particularly fault-prone, and managing autopilot settings to ensure consistently optimal overlap is challenging. Despite the higher resolution of the camera, there is no benefit to accuracy since they must be flown higher to avoid warping and artifacting in data-processing. Despite the larger image sensor, range benefits are minimal due to the heavier camera and airframe. Finally, high-end cameras and the airframes they require are quite expensive, making it harder for a business to get a solid ROI.
Aerotas’ analysis is that a built-in camera is usually preferable to a high-end independent camera, due to getting equivalent accuracy cheaper and more reliably. The best use case for a high-end camera is if very high-resolution orthophotos are required as base-maps on special projects or if an airframe with swappable payloads is required (e.g., to swap for a thermal sensor for roof inspections) – though often it will be cheaper and more reliable to just have separate dedicated drones for other sensors. Regardless, high-end cameras are only recommended for very experienced custom drone technicians.
Several companies are starting to make drone-specific LiDAR sensors. These sensors produce point-cloud data, which requires laser scanning software to manage and reduce it into usable survey deliverables.
The primary benefit of LiDAR sensors over cameras is that they can penetrate some ground cover. On projects that have some sparse tree, bush or grass cover, LiDAR can return some true ground elevations beneath the cover, reducing the amount of supplemental ground data collection needed.
The primary drawbacks of LiDAR come down to complexity and cost. LiDAR integrations are highly technical and complex – so very prone to faults – and require a great deal of time-consuming calibration and maintenance. LiDAR data management is also very complicated. Whereas drafting linework from photogrammetric orthophotos and 3D mesh surface models is somewhat straightforward, LiDAR requires working in point clouds. This requires high-powered computers and a time-consuming process of selectively reducing point clouds down to only the points needed to create the surface. Despite this added complexity, LiDAR sensors are substantially less accurate than cameras – though LiDAR lasers are very precise, their ground-tested accuracy is usually around 0.3’. Finally, LiDAR is very expensive, which makes for a challenging business investment.
Aerotas’ current analysis shows that LiDAR is still maturing. With its high cost to accuracy ratio, it is not a good investment for many survey firms today. However, for firms that frequently work on sites with moderate to sparse ground cover and have a great deal of experience with custom drone technology, LiDAR sensors, and point cloud management, these sensors could make a profitable option.
Choosing the Right Tool for the Job
There is no one right choice of sensor that applies to every company. For most firms focused on small- to medium-sized topographic, ALTA or similar projects, a drone carrying a built-in camera is usually the best option. For firms focused on large projects with lower accuracy requirements, a small independent camera mounted on a fixed-wing aircraft can be a great choice. For firms with substantial drone experience that want to differentiate themselves via very high-resolution imagery, a high-end independent camera can work well. And for firms that regularly work on sites with some ground cover, have large budgets and have ample experience with point cloud management, a LiDAR sensor will work well.
It’s important to understand that the drone is only one part of a successful drone program. Even the best drone will not deliver the survey or business results needed unless it is paired with the right field SOPs and data processing workflow to get final linework.