Five Causes of Positional Shifts in Drone Surveying

Drone surveying has become widely used in construction, civil engineering, and surveying because it is an efficient method for capturing large areas in a short time. At the same time, many people considering implementation or about to place an order are concerned about "positional shifts." Even if the images look clean, if they don't align with drawings or known control points, if parts of the point cloud or orthophoto are slightly offset, or if datasets acquired on different days don't overlap, the credibility of deliverables drops quickly.

Moreover, positional shifts in drone surveying do not necessarily occur simply because "the aircraft is inaccurate." They are often caused by multiple overlapping factors such as how coordinates are handled, how control points are installed, flight conditions, shooting methods, and processing settings. Therefore, if you address the issue without isolating causes, additional flights or reprocessing may not sufficiently improve the results.

This article first clarifies what is meant by "positional shift" in drone surveying, then explains five common causes encountered on site. It summarizes, from a practical perspective, what ordering parties should check and what implementing parties should注意, why shifts occur, what kinds of sites are prone to them, and how to prevent them.

Table of contents

• What is a "positional shift" in drone surveying?

• Cause 1: Coordinate systems and reference frames are not aligned in the first place

• Cause 2: Improper installation or observation of control points and check points

• Cause 3: The flight plan does not match site conditions

• Cause 4: Image blur, camera characteristics, and aircraft attitude disturbances during capture affect results

• Cause 5: Errors in processing settings, post-processing, or data integration procedures

• Perspectives ordering parties should confirm to reduce positional shifts

• On-site practices implementing parties should enforce to reduce positional shifts

• Summary

What is a "positional shift" in drone surveying?

First, it is important to clarify that a "positional shift" in drone surveying is not a single phenomenon. What is called a positional shift on site can take several different forms.

The most obvious case is when an orthophoto or point cloud, when overlaid on existing drawings, plans, design data, or public control points, appears offset by a few centimeters (a few in) to several tens of centimeters (several tens of in), or sometimes even more. This is a condition in which the entire deliverable is shifted in one direction, and mismatched coordinate systems or reference-point processing issues are suspected.

On the other hand, the overall alignment may be generally correct while only some areas show distortion or discrepancies. For example, the shift may differ between the top and bottom of a slope, it may be difficult to align only around structures, or steps may appear at seams in the point cloud. These are often influenced by shooting conditions, control point placement, or constraint settings used during processing.

Vertical shifts are also important. Even if alignment appears correct in plan view, elevations or as-built checks may show differences of several centimeters (several in) to more than ten centimeters (more than ten in). Since elevation control is as important as planimetric position on construction sites, vertical shifts have a significant practical impact.

A common misunderstanding is to assume that "if the photos are sharp, the positions must be correct." Image clarity and survey-grade accuracy are not the same. Even if images appear well stitched, weak coordinate backing can make the deliverables unsuitable for as-built verification, earthwork volume calculations, or design comparison.

Another misunderstanding is "if the aircraft has RTK, control points are unnecessary." While RTK or PPK-equipped aircraft can improve positional accuracy, they are not sufficient for all sites by themselves. Required accuracy, project scope, terrain, and the surrounding environment still often necessitate ground-based reference checks and verification.

From the ordering party's perspective, it is important not to judge positional accuracy simply by a subjective sense of "it seems to line up." You should confirm which coordinate reference the deliverables are aligned to, what accuracy was targeted, and how verification was performed to determine whether the deliverables are fit for purpose. From the implementing party's perspective, rather than dismissing positional shifts as flight or processing failures, identifying which type of shift has occurred is the starting point for preventing recurrence.

Cause 1: Coordinate systems and reference frames are not aligned in the first place

One of the most common practical causes of positional shifts in drone surveying is that the fundamental coordinate references are not aligned. This is an issue that exists before aircraft and software performance come into play, yet it is easy to overlook.

This happens because multiple coordinate references often coexist on site. Public coordinates, arbitrary/project coordinates, local coordinates, drawing-based coordinates, coordinates from past deliverables, and GNSS-observed coordinates can all differ in their bases; even though they all represent "location information," they will not line up cleanly if their references differ. Examples include using different plane rectangular coordinate system zones, confusing geoid heights and ellipsoid heights, or applying on-site local shifts—these causes are subtle but can have a large impact.

For example, if the client’s existing drawings were created in arbitrary coordinates but the contractor processes and delivers results in public coordinates, the entire deliverable will appear shifted. Also, if a site has been operated "in the same coordinates since long ago" but the origin of those coordinates is undocumented, new contractors may mistakenly treat them as public coordinates. In such cases, the deliverables may be internally consistent but become unusable the moment they are overlaid with other data.

This cause is especially likely when existing drawings or past survey data are used in combination. Projects such as land development, roadwork, river works, slopes, and maintenance, where multi-year or multi-contractor deliverables coexist, require particular caution. Renovation or addition work tends to have more complex coordinate origins than new sites.

To prevent this, the top priority before flying is to clarify "what to align to." Implementing parties must confirm the coordinate system to be used, the zone number, the vertical reference, the origin of known points, and the reference of existing drawings, and share these not only verbally but as documentation among stakeholders. Ordering parties should not simply request "please do it the same as last time"; they need to specify what reference was used previously.

A practical verification on site is to check consistency in advance at easily comparable locations such as known control points or clear corners of structures. Discovering inconsistencies after flying causes major rework, so it is more efficient to verify ground coordinate consistency early.

A common misconception is that "public coordinates must always be correct." Even when public coordinates are used, if the existing on-site data are in arbitrary coordinates, the deliverables will not match in practice. Conversely, arbitrary coordinates can be workable if all stakeholders understand the reference and the intended use of the deliverables. What matters is not which reference is "correct," but which reference the deliverables must be aligned to.

Ordering parties should check whether the coordinate basis of the deliverable matches the specification and agreements, whether overlaying with existing results is expected, and whether the reference is documented in metadata or reports. Implementing parties should use double-check procedures for coordinate input into software, conversion parameters, and vertical reference handling. The first step to preventing positional shifts is not expensive equipment but the basic action of aligning references.

Cause 2: Improper installation or observation of control points and check points

Control points and check points are critical for stabilizing accuracy in drone surveying. However, on sites where positional shift complaints arise, ground-based reference setup is often insufficient.

Control points are reference points that tie aerial photos to real-world coordinates. Check points are used to verify how closely processing results match reality. Even when using RTK-equipped aircraft or PPK processing, ground verification remains valuable, especially when deliverables will be treated as survey data. Proper placement and observation according to site conditions are key to ensuring accuracy.

Positional shifts occur when there are too few control points, biased placement, poor visibility, unstable installation locations, or observation errors. For example, if control points are concentrated on one side of the survey area, the opposite side will be weakly constrained, leading to overall twist or localized distortion. In sites with large elevation changes, placing points only on flat areas can produce errors on slopes or at elevation differences.

Control points that are not clearly readable in photos are another common problem. Markers may be too small, blend into the ground, be hard to see in shadow, hidden by vegetation, or appear poorly on slopes; such difficult image interpretation undermines processing stability. Even if observed coordinates are correct, they are meaningless if the points cannot be accurately picked on the photos.

This cause tends to occur in rushed projects, confined sites, undulating terrain, and areas with mixed grassland and bare ground. Sites that combine slopes, cut-and-fill, embankments, terraces, and structures require thinking beyond planar placement and considering elevation differences. Conversely, flat, obstacle-free sites may achieve acceptable results with fewer points, but that does not justify omitting check points.

To prevent this, it is important to separate the roles of control points and check points. Using all points for processing makes the result look good but leaves no independent verification. Implementing parties should plan constraint points and verification points separately and place them not only around the perimeter but also within the survey area as needed. Ordering parties should check not just the number of points but where they were placed and what residuals were found at the check points.

In practice, increasing the number of control points does not always improve accuracy. Poorly added points can include misobserved or hard-to-read points and destabilize processing. What matters is placing an appropriate number of clear, stable points in necessary locations and observing them correctly. Low-quality control points cannot be compensated for by quantity alone.

Typical misconceptions include "RTK eliminates the need for control points" and "control points at the four corners are sufficient." Required conditions vary by the shape and purpose of the survey area. Long narrow roads, slopes, rivers, and areas around structures may not be sufficiently constrained by corner points alone. Even when RTK improves shot positions, it does not automatically remove errors caused by terrain or processing conditions.

For ordering parties, a report including maps of control/check point locations, observation methods, and residual summaries is a quick quality check. Implementing parties should prepare a placement plan suited to the site before flying and choose points based on terrain, obstacles, and accuracy requirements rather than simply area size. Drone surveying looks like an aerial operation, but the key to preventing positional shifts lies on the ground.

Cause 3: The flight plan does not match site conditions

In drone surveying, how you fly matters more for quality than simply being able to fly. Positional shifts often occur because the flight plan does not suit the target terrain or objectives.

This is problematic because photogrammetry estimates shape and position from overlapping images. If required overlap is not ensured, flight altitude is too high, ground speed is too fast, altitude settings are too constant relative to terrain changes, or necessary shooting directions are lacking, processing lacks essential information. The result can be localized shifts or distortion, point-cloud gaps, and unstable orthophotos.

For example, a flight plan designed for flat development ground applied unchanged to slopes or highly undulating terrain will cause ground resolution to vary across elevation differences, making it difficult to meet necessary image conditions. For long narrow sites like roads or rivers, simple grid flights may provide weak lateral constraints, causing wave-like global shifts. In areas with many structures, nadir-only images lack side information and shape around buildings or retaining walls becomes unstable.

This cause is common when the site is not adequately inspected and template flight settings are used. It is more likely when multiple sites are handled under tight schedules or when the pilot does not fully understand the survey objectives. Even if the flight appears nominal, required quality conditions may not be met.

To prevent this, first clarify "what the deliverable is for" and plan the flight accordingly. Whether you need orthophotos, point clouds, earthwork volumes, or as-built verification changes the required capture density and flight method. Prioritizing covering large areas quickly can lead to irrecoverable problems downstream.

In practice, consider elevation changes, not just planimetric layout, when deciding flight altitude. Flying at a constant altitude means being higher over valleys and lower over ridges, which destabilizes ground pixel size and image conditions. Use terrain-following, split the survey into zones, or perform additional flights as needed. For narrow sites add cross-flight lines; for slopes include oblique imaging; around structures capture more angles—adapt to the target.

A common misconception is "increasing overlap solves everything." While insufficient overlap is an issue, merely increasing image count is not the solution. Adding many blurred or poor-condition images only increases processing time and may destabilize results. The important point is to collect images that are effective for processing with the necessary directions and homogeneous conditions.

Ordering parties should ask in the estimate or proposal what meter-class accuracy is assumed, what flight plan will be used, and whether there is consideration for slopes and structures to reduce future trouble. Implementing parties should not skip on-site checks before flying and should plan based on terrain, obstacles, sunlight conditions, and task objectives. Although drone surveying gives an impression of automated flight, planning tailored to each site is where human experience makes the biggest difference in preventing positional shifts.

Cause 4: Image blur, camera characteristics, and aircraft attitude disturbances during capture affect results

While attention often focuses on coordinates and processing settings when talking about positional shifts, if the quality of captured images is unstable, no amount of careful processing can fully compensate. Image blur at capture, shutter conditions, camera distortion, rolling shutter effects, and attitude disturbances from wind are causes often overlooked.

Photogrammetry depends on correctly matching image features. If shutter speed is insufficient and blur occurs, feature positions become ambiguous. If wind causes the aircraft to vibrate, the same ground object can appear differently across images, producing variability in processing. Rolling-shutter cameras can introduce slight distortions during movement or high-speed flight, which destabilize coordinate estimation.

This cause is common in open sites exposed to wind, coastal areas, mountainous terrain, along slopes, at high elevations, or places with mixed strong sunlight and shadow. Flying may be possible in windy conditions, but survey quality may still be compromised. Small aircraft in particular show attitude changes in the image quality, and forcing flights in poor conditions can yield visually complete but unstable results.

Shooting time of day also matters. Long shadows in early morning or late afternoon can make parts of the ground dark and destabilize image matching. Strong midday sunlight can cause blown highlights or glare. Surfaces such as water, glass, metal, or monotone pavement that lack distinct features further magnify the impact of image quality.

To prevent this, first do not judge flight feasibility solely by "can it be flown safely." You must assess whether "survey quality can be ensured." Wind speed within flightable limits may still be unfavorable for required precision. Implementing parties should evaluate not only wind speed but also wind direction, gusts, and terrain-induced turbulence.

Shutter settings are also important. Leaving exposure to automatic settings can produce insufficient shutter speeds and many lightly blurred images under some conditions. Individual blurred images may be hard to spot, but they affect overall processing. After capture, implement procedures to inspect multiple representative images on site for blur, exposure, and shadow conditions, and decide on additional capture if needed.

A common misconception is "if the flight succeeded, the data must be usable." Flight viability and survey-quality viability are not the same. Another misconception is "processing software will correct it," but while software corrections help, they have limits when source images are too poor.

Ordering parties should check shooting-day weather conditions, whether additional flights will be performed, the criteria for re-shooting, and quality-check methods. It is important to have a provider that can postpone capture when conditions are unfavorable. Implementing parties must make final decisions based on on-site wind, light, and shadow conditions and perform on-site image checks so they can take additional action if needed. Positional shifts may appear to originate in the processing room, but their seeds are often sown on the flight site.

Cause 5: Errors in processing settings, post-processing, or data integration procedures

Positional shifts in drone surveying can occur in processing even when on-site capture appears problem-free. As aircraft and capture performance have improved in recent years, differences in processing settings and post-processing have become more visible as sources of quality variation.

This happens because deliverables are produced by merging photos in software, assigning coordinates, and converting to point clouds, DSMs, DTMs, and orthophotos as needed. In this process, errors in reading control points, coordinate input mistakes, unit mix-ups, inclusion of unnecessary images, handling of camera calibration, insufficient outlier removal, and inconsistent merging conditions for multiple flights can all cause shifts.

For example, a single-digit typo in a control point coordinate, swapping X and Y, or having only the vertical reference in a different basis can severely degrade overall accuracy. When integrating data captured on multiple days, inconsistent flight conditions or reference processing often produce steps or mismatches at boundaries. Inappropriate ground classification after point-cloud generation can lead to misinterpretation of ground elevation and be mistaken for positional shifts.

This cause is common in projects with tight deadlines and high processing volumes, organizations where operators follow varied procedures, or where capture and processing responsibilities are separated and communication is weak. If the flight team does not adequately communicate site conditions and the processor is unaware of hard-to-see control points or site characteristics, unnatural results may be overlooked.

To prevent this, do not treat processing as mere software work. You must confirm which points were used as constraints, which were used for verification, where residuals occurred, whether there are localized distortions, and whether boundaries of multiple flights show unnatural differences. Rather than relying solely on average error, assess residual distributions in conjunction with site topography.

Also, checking only orthophotos may miss issues. Planimetric alignment can appear correct while vertical errors exist. Conversely, steps in point clouds may be due to ground classification or differences in target objects rather than positional shifts. Define inspection points for each type of deliverable to make quality control manageable.

A common misconception is "the software will automatically produce high-precision results." Modern software is powerful, but automatic processing is stable only when preconditions are satisfied. If those preconditions are violated, the software may produce seemingly plausible deliverables that are not survey-grade. Another misconception is "a small mean error means everything is fine." Even with a good mean, localized biased shifts can disrupt site operations.

Ordering parties should request not only deliverables but also documentation that clarifies the quality control approach: how control and check points were handled, residual summaries, the coordinate system used, and processing condition overviews. Implementing parties should standardize processing procedures, use double checks for input values, organize verification items per deliverable, and ensure information sharing with field personnel to reduce human errors.

Perspectives ordering parties should confirm to reduce positional shifts

Positional shifts in drone surveying are not only the implementer’s responsibility. How the ordering party checks and specifies requirements greatly influences the likelihood of problems. Especially for first-time orders, comparing solely on price and delivery date can expose you to positional-shift or accuracy issues later.

First, clarify the intended use of the deliverables. Whether they are merely for construction progress records, for earthwork calculations or as-built verification, or for overlaying with existing drawings and design data affects the required accuracy level. If the purpose is ambiguous, the implementer cannot judge how much accuracy to pursue.

Next, sharing coordinate references is critical. If existing drawings exist, communicate their coordinate reference, the information on known on-site points, and whether consistency with past deliverables is required. Ambiguity here often leads to the typical problem: "the deliverables look fine but don't overlay with existing materials."

Also, during estimates or proposals ask how control and check points will be treated, how much on-site verification will be performed, and how accuracy will be confirmed. These questions reveal whether the provider is offering mere capture services or a survey-quality-conscious approach. Simply asking these questions can change the substance of proposals.

When checking deliveries, do not judge by the orthophoto’s appearance alone. Perform checks according to the intended use: verify against known points, overlay with existing drawings, test for localized distortions, and check vertical accuracy. If reports are too sparse to show the basis for accuracy, be cautious about how you use the deliverables.

On-site practices implementing parties should enforce to reduce positional shifts

For implementers, preventing positional shifts depends less on secret tricks and more on rigorous adherence to basics. Even with high-performance aircraft, deliverables will be unstable if reference confirmation, ground points, flight planning, shooting conditions, or processing checks are ambiguous.

First, understand the site objectives before flying. If the task is more than aerial imagery—if it is a survey deliverable—you must know what is required. Capture design varies for roads, slopes, development sites, and around structures.

Second, do not neglect preparation of ground references. Control and check points take effort but skipping them often causes large losses downstream. Consider visibility, stability, and placement balance according to site conditions.

During the flight, judge conditions not only by safety but by survey-quality criteria. Rather than flying because "a bit of wind is okay," consider whether the wind allows the required accuracy. Perform on-site image checks after capture and make decisions about additional on-site measures if problems are found.

In processing, do not leave everything to the software; interpret residuals and unnatural features. Pursuing numbers alone without understanding the site makes it easy to miss anomalies. If flight and processing responsibilities are split, set up a system to share site conditions and observations between teams.

Summary

There is no single cause of positional shifts in drone surveying. Multiple factors relate to shifts: mismatched coordinate systems and references, inadequate control and check points, flight plans that do not suit the site, image blur and camera/attitude issues at capture, and errors in processing settings and post-processing. It is also troublesome that shifts commonly result from multiple causes overlapping.

Therefore, preventing positional shifts requires more than thinking "a high-performance drone guarantees safety." You must manage the entire workflow: defining what to align to, securing ground references, flight planning suited to the site, judging capture conditions, and careful verification in processing.

For ordering parties, it is important not to judge by price or appearance alone but to confirm that the required accuracy for the intended use is guaranteed. For implementers, operational capability that includes coordinate management and quality assurance—not just flying skills—is essential. Drone surveying is a very effective method, but when used as survey data, understanding and addressing positional shifts is unavoidable.

In practice, combining the drone’s efficiency for wide-area capture with high-precision ground checks is effective. For example, confirming reference points, performing supplementary ground surveying, or partial rechecks on the ground can enhance overall deliverable reliability. In such cases, using solutions that combine ground high-precision positioning—such as the iPhone-mounted high-precision GNSS positioning device LRTK—with drone surveying is practical. Rather than completing everything with drone surveying alone, assigning roles to aerial and ground methods to secure accuracy makes it easier to operate with fewer positional shifts.