What are the differences between RTK and IMU? Explaining the combined effects in five points

Table of contents

• What is the difference between RTK and IMU?

• What RTK is good at and what it struggles with

• Strengths and weaknesses of IMUs

• Effect 1 of combining RTK and IMU: It is easier to improve the continuity of positioning

• Effect 2 of combining RTK and IMU: Easier pose estimation of moving objects

• Effect 3 of combining RTK and IMU: It makes it easier to improve the stability of point cloud measurements and as-built verification

• Effect 4 of combining RTK and IMU: Easy to use for construction guidance and machine guidance

• Effect 5 of combining RTK and IMU: Makes it easier to improve work efficiency and repeatability across the entire site

• How to differentiate between standalone use and combined use

• Practical precautions when using point clouds for surveying and construction

• Summary

What is the difference between RTK and IMU?

RTK and IMU are both technologies commonly used in fields that deal with position and motion, but their roles are not the same. If you implement them without a clear understanding of the differences, you are likely to encounter problems such as not achieving the expected level of accuracy, unstable behavior in obstructed environments, and discrepancies in point clouds or construction data. The key point to understand first is that RTK is primarily a system for determining absolute position with high accuracy, while IMU is a system for estimating motion state from changes in attitude, acceleration, and angular velocity.

RTK is a method within satellite positioning systems that uses correction information from a base station to reduce errors and achieve centimeter-level high-precision positioning. A major characteristic is that it makes the coordinates required on site easier to handle in relation to known points and reference points. RTK becomes increasingly important in situations where the correct position needs to be handled with the correct coordinates, such as surveying, as-built verification, marking-out, construction management, and machine guidance.

On the other hand, an IMU uses sensors such as accelerometers and gyroscopes to determine, at a high frequency, which direction the device is facing, how it is moving, and how much it has rotated. Its strength is that it can continue to capture moment-to-moment changes in motion even without signals from satellites. In other words, it is easier to understand if you think of RTK as providing high-precision positioning based on external references, while the IMU provides continuous attitude and motion estimation based on internal sensors.

This difference directly affects how the systems are used on-site. For example, if you need to reliably obtain coordinates point by point in an open area, RTK will be central. Conversely, if vehicles or equipment are constantly moving and you want to closely track their movement, tilt, and changes in direction, the value of an IMU increases. However, in real-world sites, absolute position alone is often not sufficient, nor is attitude estimation alone. Requirements such as wanting to know both position and orientation, continuing to work without stopping even when there is occlusion, and collecting data without gaps while moving often arise simultaneously. Therefore, in practice, rather than treating RTK and IMU as an either-or choice, it is important to clarify which tasks to assign to RTK and which to assign to the IMU, and then combine them.

Especially for point cloud surveying, vehicle-mounted recording, construction machine guidance, and capturing current conditions while walking, if you do not understand this division of roles you cannot explain why results are unstable. RTK provides the positional reference, and the IMU supports the continuity of motion. The basic principle of combined use is to leverage the differences in their characteristics so they complement each other and solve problems that are difficult to address with either one alone.

What RTK Is Good At and What It's Not Good At

The greatest strength of RTK is its ability to handle absolute coordinates with high accuracy. Because points acquired in the field can be placed within a known coordinate system, it becomes easier to link many processes such as survey results, design coordinates, as-built evaluation, drafting, and construction management. The value of RTK lies in its use not merely as a positional reference but as coordinates that can be overlaid with other data.

For example, for tasks such as verifying pile positions aligned with reference points, determining the difference between the current position and the design alignment, and reflecting the locations of existing structures in downstream drawings and 3D models, RTK clarifies the reference for work. Another advantage is that if a different person re-surveys later, it is easier to compare results because they use the same coordinate system. This is very important for improving the reproducibility and explainability of field work.

However, RTK has conditions it struggles with. A typical example is environments where it is difficult to receive satellite signals stably. Under elevated structures, in mountainous areas, in places with dense tree growth, in areas with closely spaced buildings, under bridges, and near tunnel entrances and exits, satellite visibility can be poor and signals are more likely to be affected by multipath reflections. This can make it difficult to maintain a fixed solution and may cause accuracy to become unstable.

Also, although RTK yields high-precision positioning results themselves, relying only on its update cycle makes it poor at smoothly representing sudden movements or fine attitude changes. It can adequately handle measurements at static points or relatively slow movements, but in situations where the equipment is frequently shaken, a vehicle undergoes frequent small accelerations and decelerations, or a sensor’s orientation is constantly changing, a sequence of position points alone makes it difficult to accurately follow the behavior. Even when positions are obtained, the lack of information about heading and tilt can leave corrections required in practice insufficient.

Furthermore, relying solely on RTK has the problem that work tends to stop the moment the signal temporarily deteriorates. When quality temporarily drops due to effects such as a change from a fixed solution to a float solution, an interruption of correction information, or a deterioration in satellite geometry, special care is needed in handling the data from that period. While re-observation can address this in surveying, in work that records continuously while moving, that momentary loss can later lead to major rework.

In short, RTK is strong at handling correct positions in the correct coordinate system, but it is weak in obstructed environments, temporary reception degradation, following continuous motion, and detecting attitude changes. Therefore, rather than treating RTK alone as a universal solution, it should be operated on the premise of which conditions will cause its accuracy to deteriorate.

What IMUs Are Good At and What They're Not Good At

The strength of an IMU lies in its ability to track a device’s motion and changes in orientation at high frequency. Even without external position information, changes in acceleration and angular velocity allow it to capture in detail which direction something is moving, how much it is rotating, and how it is tilting. This property is highly useful for understanding the behavior of sensors, vehicles, and handheld devices while they are in motion.

For example, when recording on foot or measuring while a vehicle is in motion, the equipment is constantly subject to movement. If such vibrations and changes in attitude are ignored, the orientation and alignment of the acquired data tend to become unstable. Using an IMU makes it easier to capture instantaneous changes in tilt and rotation, which in turn makes it easier to maintain consistency among point clouds, images, and motion trajectories. Another advantage is that, even with short occlusions or momentary signal dropouts, it is easier to bridge the motion using information about the movement immediately beforehand.

However, an IMU is not good at maintaining an accurate absolute position on its own over long periods. Estimation that accumulates movement using only internal sensors tends to gradually build up errors as time passes. This is the so-called drift problem. Even if the estimate looks reasonable at the start, over long distances or long durations the estimated position can drift away from the true position. While an IMU is strong for orientation and relative motion, maintaining long-term accuracy in coordinates requires an external reference.

Also, an IMU, without external references, cannot be directly tied to the coordinate system required on site. If you want to produce deliverables aligned with design coordinates, public coordinate systems, or site reference points, an IMU alone tends to be insufficient as a final deliverable. Even if it can track changes in position, the result often becomes ambiguous as to which reference system those positions are relative to. To treat the data as survey results or construction-management data, a process to tie the data to reference coordinates at some point is indispensable.

Furthermore, IMU results are influenced by initial conditions, sensor quality, calibration state, and mounting posture. Under conditions such as the sensor being mounted misaligned, excessive vibration, large temperature changes, or movement patterns that differ from those assumed, the reliability of the estimates can decrease. Therefore, an IMU does not automatically become stable simply by being present; its installation conditions and computational methods must also be managed.

In other words, an IMU is strong at capturing instantaneous motion and attitude and at maintaining the continuity of successive movement, but it is a technology that struggles to guarantee the correctness of absolute coordinates on its own over extended periods. The difference — RTK being strong with external references and the IMU being strong in internal continuous estimation — is most clearly reflected here.

Effect 1 of combining RTK and IMU: Makes it easier to improve positioning continuity

The primary effect of combining RTK and an IMU is that it makes it easier to enhance the continuity of positioning. On site, satellite visibility is not always good. Passing beside buildings, skimming under elevated structures, going under trees, or moving near heavy machinery or structures—situations where satellite reception temporarily degrades—occur frequently. If you operate with RTK alone, accuracy can drop or the solution can become unstable at those moments, making continuous operations prone to interruption.

If an IMU is available here, even short segments where satellite signals become unstable can be more easily filled in using the immediate prior movements and attitude changes. Of course, an IMU alone cannot completely maintain absolute position over long periods, but it is effective for short-term gap filling and smooth connections. In practical work, whether this short-term filling is possible greatly affects the amount of data loss and the number of re-measurements.

For example, when acquiring current conditions on foot, even a few seconds of occlusion can cause jumps in the trajectory, increasing the effort required to reconcile data in post-processing. In vehicle-based surveys, if the position solution is disturbed at the moment of passing under bridges or near street trees, linear deliverables can appear wavy. By combining RTK and an IMU, you can maintain the coordinate reference during RTK-reliable segments while the IMU supports continuity in segments where the signal momentarily weakens.

This improvement in continuity is not just about making things look neat. It also affects the stability of data quality when linking information to subsequent tasks such as drafting, point cloud integration, as-built verification, and organizing construction records. Results in which the trajectory jumps, point spacing becomes irregular, or orientation changes abruptly are difficult to analyze or explain, and therefore hard to use as deliverables. Combining RTK and IMU reduces such unnatural discontinuities and makes it easier to consolidate the data into a form usable on site.

However, what should be noted here is that during signal blockage, no matter what you do, the original RTK accuracy will not necessarily be maintained. An IMU is fundamentally strong for short-term compensation and maintaining continuity, but it is not something that can completely overcome prolonged blockage environments. Therefore, on site you need to anticipate in advance how much blockage can be easily absorbed by the compensation and which sections require stricter quality checks. Do not overtrust it; it is important to incorporate quality management assuming the use of compensation.

Effect 2 of Combining RTK and IMU — Easier Orientation Estimation for Moving Objects

The second benefit is that it becomes easier to determine the orientation (attitude) of a moving object. RTK alone can determine position, but it may not be sufficient to tell which direction a device or vehicle is facing, how much it is tilted, or how it is rotating. In particular, for vehicles with mounted sensors, handheld devices, and construction machinery, attitude information as well as position directly affect the quality of the results.

For example, in point cloud measurement, even if only the sensor's position is known, if it is unclear which direction it was scanning, the orientation and overlap of the acquired point clouds are likely to become misaligned. Even in processes that combine images, if the camera's pose during capture is not stable, it can negatively affect the accuracy of the merging. The same applies to construction machinery and guidance devices: if you do not know not only the current position but also the direction of travel and the equipment's tilt, the quality of operational decision-making deteriorates.

Adding an IMU makes it easier to capture fine motions caused by lateral sway, vertical sway, turning, and acceleration/deceleration, and to treat position and orientation as a single entity. This is highly significant for on-site measurement tasks performed while in motion. If you are only measuring stationary points, orientation information is of limited importance, but when acquiring data while continuously moving, errors caused by unknown orientation tend to become larger.

Also, being able to determine orientation makes it easier to review device handling and measurement methods. For example, if data quality is poor in only a particular section, it becomes easier to determine whether that is due to the satellite environment, large sway while walking, or sharp turns of a vehicle. The value of using them together is not only that you can assess the accuracy of the results, but also that it becomes easier to analyze why those results occurred.

In practice, people tend to focus only on positional errors, but disturbances in attitude often affect positions and the way point clouds appear. In particular, for as-built records, pavement and slope acquisition, 3D modeling around equipment, and the operation of sensors mounted on moving platforms, attitude stability determines the quality of the results. Combining RTK and IMU enables more practical quality control that does not rely solely on position.

Effect 3 of combining RTK and IMU: It tends to improve the stability of point cloud measurements and as-built verification

The third effect is that it makes it easier to improve the stability of point cloud measurements and as-built verification. Point clouds and three-dimensional data are not just about acquiring a large number of points; if the coordinate system, the position, and the orientation from which the data were acquired are not properly aligned, misalignments and distortions will become apparent when performing registration or comparisons later. The combined use of RTK and IMU is very effective in enhancing this consistency.

Having RTK makes it easier to provide a reference of absolute coordinates to acquired data. This in turn makes it easier to overlay data collected across multiple days and results obtained with other instruments. In as-built verification, coordinate alignment is indispensable for evaluating the difference from the design. If coordinates remain ambiguous, discussing the discrepancies becomes less meaningful. RTK plays the role of creating that foundation.

On the other hand, there are cases where handling subtle attitude changes and the continuity of motion during point cloud acquisition with RTK alone is difficult. In particular, in pedestrian- or vehicle-mounted surveys, because the sensor is constantly in motion, even slight sway or rotation is reflected in the results. An IMU can help by capturing those motion changes during acquisition, making it easier to suppress point cloud distortions and unnatural seams.

This effect also impacts the efficiency of post-processing. If the reference coordinates are stable and the movement trajectory and pose changes are captured reasonably, it becomes easier to align point clouds and check for noise. Conversely, if position is captured but pose is unstable, or pose can be estimated but absolute coordinates are weak, you will need to force alignment during post-processing. As a result, not only does processing time increase, but results also tend to vary between operators.

The same applies to as-built verification. For embankments and excavations, paving, elevation control, and inspections around structures, it is necessary to compare the shape at a given time with the design or previous measurements. If the position or orientation during measurement is unstable, it becomes difficult to tell whether observed changes are actual differences or measurement errors. Combining RTK and IMU makes it easier to achieve both a positional reference and stable measurement operation, increasing the reliability of differential evaluations.

Furthermore, point clouds and as-built verification are often used not as one-off measurements but for continuous monitoring at each stage of the process. In this case, it is important to be able to acquire data with the same quality every time. Operations that combine RTK and IMU are well suited to continuous management because they make it easier to maintain consistent quality even when site conditions vary.

Effect 4 of Combining RTK and IMU: Easy to Use for Construction Guidance and Machine Guidance

The fourth effect is improved usability for construction guidance and machine guidance. On construction sites, it is necessary not only to know the current position but also information such as which direction you are moving, how much the orientation is changing, and when you should operate. Particularly in situations where decisions must be made while moving, RTK alone is not sufficient, and continuous attitude information from an IMU becomes useful.

For example, in tasks such as excavation, shaping, spreading and leveling, slope works, and positioning assistance, there are many situations where you want to grasp not only the positional relationship with the target but also the tilt, turning, and changes in the direction of travel of the machine or equipment. RTK can clarify the relationship to design coordinates, but it can be difficult to adequately represent the smoothness of operation and instantaneous behavior. Combining an IMU makes it easier to track fine changes during operation and helps stabilize the response of guidance displays and auxiliary calculations.

Also, at construction sites the reception environment often temporarily worsens, and if the RTK status becomes even slightly disturbed and guidance becomes unstable, operators are likely to lose trust. In real-world sites, as important as accuracy itself is that behavior does not jump suddenly and that displays and guidance are smooth. An IMU also proves effective in these aspects of usability. Because it makes it easier to bridge behavior over short unstable intervals, it tends to reduce the sense of discomfort during operations.

Furthermore, in construction guidance, it is necessary not only to know the current position but to operate while predicting movements slightly ahead. Because sharp turns, stops, re-starts, and passing over steps frequently occur, an IMU that can capture changes in motion at high frequency is well suited. By maintaining a positional reference with RTK while supplementing instantaneous behavior with an IMU, it becomes easier to support on-site operational decision-making.

Of course, the quality of construction guidance and machine guidance is not determined solely by sensor performance. Mounting position, the geometric relationship with the machine, calibration, latency, display design, and so on also have a major impact. However, configuring the system to handle both position and orientation at minimum greatly expands operational flexibility. In particular, on sites that require both accuracy and stability while in motion, using both RTK and an IMU is effective.

Effect 5 of combining RTK and IMU: Easier to improve site-wide work efficiency and repeatability

The fifth effect is that it becomes easier to improve overall on-site work efficiency and repeatability. The combined use of RTK and IMU not only improves positioning accuracy and measurement quality, but also contributes to site-wide productivity by reducing re-measurements, shortening post-processing time, reducing variability between operators, and improving data integration between processes.

First, when the continuity of positioning and measurement improves, work interruptions are reduced. With RTK alone, even slight occlusion can make the solution unstable, and each time you may need to stop to check or re-acquire a fix. If an IMU works complementarily and increases the sections where you can continue on the spot, it leads to a reduction in work time. Even just reducing the need to walk back or redo a pass significantly changes the burden at the site.

As a result, the reproducibility of outcomes becomes easier to achieve. If both position and orientation are acquired using a consistent approach, it becomes easier to achieve similar quality even when personnel change. This also contributes to labor savings, in the sense that it makes it easier to maintain consistent quality even without skilled operators. In practical work, it is more important that anyone can obtain the same results than the performance of the equipment itself. The combined use of RTK and IMU contributes to that standardization.

Furthermore, it becomes easier to improve data integration with downstream processes. Point clouds, drawings, as-built data, construction history, current-condition photos, and map data are easier to integrate the better organized their coordinate, time, and orientation information is. While using only RTK or only an IMU tends to leave gaps, combining both makes it possible to handle both positional references and movement history. As a result, it becomes easier to produce data that is usable across the entire workflow—from site assessment to construction, inspection, and maintenance.

Also, the ease of explaining things on site should not be overlooked. When explaining why this data can be trusted and why quality checks are necessary for this particular segment, it becomes easier to assess if you separate the roles of RTK and IMU. Because it makes it easier to determine whether the cause is the reception environment or attitude variations, it also becomes easier to plan countermeasures when problems occur. This is a benefit not only technically but also from a management perspective.

In this way, the combination of RTK and IMU contributes not only to improved single-measurement performance but also to overall on-site operational efficiency and quality stability. The more a site aims to reduce staffing and leverage data, the more important it becomes to manage position and orientation as an integrated whole rather than treat them separately.

How to Decide Between Standalone Use and Combined Use

RTK and IMU do not always need to be used together. In practice, it is important to use them selectively based on purpose, site conditions, required accuracy, working time, and equipment configuration. In fact, distinguishing situations where single use is sufficient from those where combined use is effective makes it easier to increase the benefits of adoption.

Representative examples where RTK alone is sufficient include observations of static points and point-by-point position checks in relatively open environments. For control point verification, acquiring the current condition at single points, determining boundaries or the positions of structures, and spot checks of as-built measurements, absolute positional accuracy is important while information about attitude and continuous movement is relatively less important. In such cases, it is more effective to operate RTK carefully and to optimize observation conditions.

On the other hand, an IMU by itself is most useful for applications that focus on capturing short-term behavior and recording relative motion. For example, when checking equipment vibration, monitoring changes in orientation, or acquiring machine motion, absolute coordinates are not necessarily required. However, if the results are to be retained as survey outputs or construction management data, an IMU alone is often insufficient, so you should be mindful to limit its use to specific applications.

Combination use is particularly effective for tasks that acquire data while moving, sites with intermittent occlusions, operations that handle continuous data such as point clouds and images, and situations that require both position and orientation, such as construction guidance and machine guidance. In these applications, RTK alone tends to lack continuity and orientation information, while an IMU alone tends to lack absolute coordinates, so their complementary relationship directly translates into operational effectiveness.

When considering how to differentiate their use, the important thing is what you want to leave as the deliverable. Depending on whether you need survey results aligned with public coordinates, stable guidance during operations, improved point-cloud alignment accuracy, or the ability to absorb short-term occlusions, the elements you should prioritize will change. If you choose equipment while the objectives are unclear, you may end up with a system that is more feature-rich than necessary or, conversely, lacking the performance you need.

Therefore, when deploying in the field, it is important to first clarify which of position accuracy, pose estimation, continuity, resistance to occlusion, and post-processing efficiency you will prioritize. Based on that, deciding whether to use RTK and IMU individually or in combination makes it easier to arrive at a configuration that is neither excessive nor insufficient.

Practical Considerations When Using Point Clouds in Surveying and Construction

To leverage RTK and IMU on-site, simply providing the equipment is not enough. You must have coordinates, mounting, corrections, quality checks, and operational procedures all in order, otherwise it will be difficult to realize their full effect. In particular, for practical work involving point clouds, surveying, and construction, advance preparation and operational rules greatly affect the results.

First and foremost, it is important to clarify which coordinate system will be used to handle the deliverables. RTK is a technique that excels at absolute coordinates, but if the coordinate system setup is ambiguous you cannot leverage that strength. You need to sort out at the outset whether a site-based reference is acceptable, whether to align to known control points, or whether the data will later be overlaid with drawings and design data. Even with point cloud surveys, data may look clean at acquisition but become difficult to use in practice if it does not match other datasets later.

Next, it is important not to underestimate the IMU's mounting conditions and initialization conditions. An IMU can capture attitude and motion at high frequency, but if the mounting is misaligned, the fixation is loose, or the device's reference axes are not aligned, the reliability of the estimated results will decrease. Especially for vehicle- or machine-mounted systems, differences in mounting position can easily affect the interpretation of behavior, so the installation state should be managed in a reproducible way.

Furthermore, it is important not to rely too heavily on interpolation during signal blockage. Even when using RTK and an IMU together, prolonged reception degradation or complex multipath environments can cause quality to deteriorate. To distinguish sections that can be supplemented from those that require rechecking, field procedures should incorporate checks of quality indicators and observation logs. Even if the results look fine at first glance, discrepancies can appear later when comparisons are made.

In point cloud work, it is also effective to appropriately establish reference points and validation points. Even if combining RTK and IMU increases stability, the validity of the results still needs to be verified separately. Especially on large sites or sites with complex structures, quality can drop in only certain sections, so it is reassuring to have points that allow checking across the entire area. In surveying and as-built verification, the presence or absence of this verification mechanism greatly affects the reliability of the results.

On construction sites, equipment behavior and the way operators use it also affect quality. Sudden swinging, excessive tilting, repeated stopping and restarting, loosened attachments—operational habits like these can show up in the results. Therefore, standardizing operational rules, not just equipment performance, is necessary. Deciding on walking speed, how to handle the equipment during stops, conditions for reinitialization, and methods for passing through shielded sections, for example, makes it easier to maintain consistent quality even when personnel change.

Finally, it is important to clearly define the division of responsibilities between post-processing and on-site checks. Using RTK and IMU together tends to improve continuity in the field, but relying solely on on-site judgment to complete everything is risky. By separating the checkpoints—such as ensuring obvious anomalies are not missed in the field, and verifying coordinates, trajectories, attitude, and overlap during post-processing—you can reduce rework.

Summary

Put simply, the difference between RTK and IMU is that RTK excels at acquiring high-precision absolute positions, while IMU is a technology that excels at continuously estimating orientation and movement. RTK makes it easy to handle correct positions tied to a coordinate frame, but it is vulnerable to signal blockage, short-term reception degradation, and changes in attitude during continuous motion. IMU can track fine changes in motion and orientation, but on its own it is difficult to maintain an accurate absolute position over long periods and it is susceptible to drift.

Therefore, in practice it becomes more meaningful to combine the two. The main combined benefits are that it is easier to improve the continuity of positioning, easier to determine the pose (orientation) of moving objects, easier to enhance the stability of point-cloud measurements and as-built verification, easier to use for construction guidance and machine guidance, and that it tends to increase overall site work efficiency and reproducibility.

In particular, it is important to understand separately the position reference provided by GNSS positioning and the attitude and motion estimation provided by an IMU. On sites where not only position but also orientation and movement are required, relying on only one of them tends to be insufficient. In tasks such as point clouds, surveying, construction, as-built verification, and mobile measurement, assigning roles based on the differences between the two makes it easier to improve both data quality and work efficiency.

When planning deployment, rather than simply choosing a high-performance configuration, it is important to clarify what outcomes you want, what the site conditions are, how much signal obstruction there is, and how closely alignment with the coordinate system is required. Based on that, if you can determine where RTK alone is sufficient and where an IMU should be combined, it becomes easier to apply the system in practice without overreach. RTK and IMU are not similar technologies; they fulfill different roles. Correctly understanding that difference and using them to complement each other is the quickest way to achieve results on site.