What is the accuracy of Aerial Surveys?
For surveyors and engineers, data that isn’t accurate isn’t just worthless, it can be a major liability. Making sure project data is accurate should be the highest priority for survey projects.
Before we can even talk about measuring and maintaining accuracy, we need to be sure that we agree on what accuracy means. There are a number of different ways to measure accuracy. Some are informal and based off a few check points. Some measure confidence intervals (i.e. 95% of points must be within 0.1 feet of check points) and others are based off contour intervals (i.e. 90% of elevation points must be within ½ contour interval
But the most thorough and reliable set of accuracy standards are the ASPRS Positional Accuracy Standards, which are endorsed by Aerotas and are used throughout our process as well as this guide.
Key components of ASPRS Positional Accuracy Standards
Accuracy is measured by the RMSE (Root Mean Squared Error) of the difference between modeled points and independently measured points.
Accuracy can only be accurately measured using Checkpoints that were not included in the production of the model.
Ground measurements must be made with an independent source of higher accuracy than the aerial mapping system.
Horizontal and Vertical accuracy should be measured independently, and can have different properties depending on vegetation.
Accuracy vs. Resolution
There is often confusion between the concepts of resolution and accuracy in aerial surveying. Put simply, resolution is how much detail is contained in a single photo, usually measured by the size of individual pixels, known as ground sampling distance, or GSD. By contrast, accuracy is the ability to make measurements based off those photos. Accuracy is best measured using statistical methods, such as RMSE, as defined by the ASPRS Positional Accuracy Standards. Accuracy dependent on not just the resolution of the photos, but numerous other factors and errors that exist throughout the surveying process. Because of this, accuracy will always be less than resolution, usually by a factor of about 3.
Absolute vs. Relative Accuracy
There are also two different ways to measure accuracy, absolute and relative.
Absolute accuracy is the amount of error between a model and a known point on the planet earth. Absolute accuracy must be measured against a global coordinate system, such as NAD83 or WGS84.
Relative accuracy on the other hand, effectively measures the accuracy between points within a single project. For example, if you need to measure the area of a concrete pad, it doesn’t matter if that pad is in Texas or Oklahoma so long as the area measurement is correct. This is relative accuracy.
Projects measured in local coordinate systems can have very high relative accuracy, but poor absolute accuracy if they are not georeferenced. Often, projects processed with RTK data also have very high relative accuracy, but only mediocre absolute accuracy if the precise location of the base station is not known, or the baseline distance is too high.
A major risk in drone surveying is believing a project to be absolutely accurate when it is only relatively accurate. This can lead to significant problems when a site is surveyed multiple times.
Factors that impact Accuracy
There are numerous factors that can impact accuracy. Below is a list of some of the biggest factors that impact accuracy. However, every system and every project is different, and there are many things not on this list that can potentially have a negative effect on accuracy.
Flight Altitude – The lower the flight altitude, the better the resolution. If paired with an appropriate number of ground control points as well, this often means better accuracy. But be careful, lower flight altitude, with insufficient ground control points, can actually lead to worse accuracy than a higher flight altitude with the right amount of ground control.
Overlap – For a project to be accurate, it needs to have sufficient overlap. We recommend 75/75 for the vast majority of projects, though there are a few exceptions. Insufficient overlap causes bad accuracy. However, too much overlap can cause problems too. Anything above 85/85 can actually cause matching errors in photogrammetry algorithms and should be avoided where possible.
Lighting – Poor lighting or inconsistent lighting can cause some problems with accuracy. Low lighting can make it difficult to identify features and cause motion blur. While perfect lighting conditions is bright overcast with no shadows, the vast majority of daytime situations, even with lots of shadows, is perfectly OK.
Type of Terrain – Some terrain is easier to process than others. Tree cover and shrubbery can cause errors in accuracy, as can tall grasses. Hardscape (e.g. concrete, asphalt) leads to the highest accuracy, with dirt, gravel, and short cut grasses not too far behind.
Geography / Topography - The geography and topography of a site can have an impact on accuracy as well. Certain features, like dense forests and waterways, are impossible to map and can cause errors depending on where in the site they are placed. Significant changes in elevation, if not adequately managed in flight planning, can either cause for too much overlap or too little, which also causes error.
Quantity of Ground Control Points – A project needs sufficient ground control to have good accuracy. Without enough ground control, accuracy very quickly falls off a cliff. The proper number of ground control points depends on each individual project, the GNSS system used, and the flight altitude. Read our ground control guide for more information.
Distribution of Ground Control Points – Even when there is a sufficient quantity of ground control points, if they are not distributed properly, it can cause error in certain areas of the project. Read our ground control guide for more information.
On-board GNSS system used – A properly managed RTK/PPK antenna on the aircraft can lead to a significant improvement in accuracy, with fewer ground control points than otherwise. However, these systems can be complex and, if not processed properly, can actually hurt accuracy. Read our guide to GNSS systems for more information.
Camera Quality – With all else equal, a better camera can improve accuracy. However, in all of our experience with drone surveying, this is rarely the most important factor to accuracy. If other sources of error are present, then a better camera will NOT improve accuracy at all. At Aerotas, we consistently get 0.1’ RMSE accuracy and better with uncalibrated 20MP cameras, and you should only consider upgrading your image sensor if you are already getting these levels of accuracy and need to take it to the next level.