What if, by simply pointing a smartphone on site, the design drawings would appear overlaid on the actual scene? You could dramatically reduce the work of comparing paper drawings while measuring with a tape or driving stakes to verify positions. A "drawing AR overlay" projects design drawings or models into real space through the camera, enabling intuitive and rapid preconstruction checks and inspection of as-built conditions. In recent years, DX (digital transformation) has advanced in the construction and civil engineering sectors, and this AR technology is attracting strong expectations for labor savings and accuracy improvement on site. This article explains in detail the basics of drawing AR overlay, its benefits, tips for achieving high accuracy (using LRTK), practical procedures, precautions, comparisons with other methods, and examples of integration with point cloud data. Finally, we introduce LRTK as an easy-to-start solution and propose points for on-site adoption.

Table of Contents

• Basic concept and background of drawing AR overlay

• Why overlaying drawings in AR speeds up site checks

• Types of drawing data used and preparation methods (2D drawings, 3D models)

• Centimeter accuracy! Importance of high-precision alignment using LRTK

• Drawing AR overlay with a smartphone: practical step-by-step guide

• On-site operational cautions (weather, visibility, satellite reception, etc.)

• Common failure cases and how to avoid them

• Comparison with other methods (paper drawings, marker-based AR, conventional surveying)

• Examples of AR combined with point cloud data and photogrammetry

• Recommendation for introducing simple surveying & drawing overlay using LRTK

• Frequently Asked Questions (FAQ)

Basic concept and background of drawing AR overlay

"Drawing AR overlay" is a technique that displays architectural and civil engineering design drawings (2D CAD drawings or 3D models such as BIM) over the real construction site scene in real time. Through a smartphone or tablet camera, actual scenery and design data are composited so that lines or models from the drawing appear to be drawn on the site, intuitively closing the gap between the drawing and the actual site.

The attention to this technology stems from issues in conventional on-site verification work. With paper drawings or PDFs in hand, confirming positions and dimensions on site required advanced interpretation skills and imagination. Surveying instruments were needed to measure coordinates, and it was necessary to mark out positions on the ground to get accurate alignment; these confirmation tasks consumed time and manpower. Also, plans that looked fine on drawings sometimes revealed clashes with surrounding structures or topography when viewed on site. To discover and correct such drawing-site inconsistencies early, drawing AR overlay has begun to be introduced.

AR (Augmented Reality) itself was originally known in gaming and entertainment, but recently practical use in the construction sector has accelerated. Initiatives such as the Ministry of Land, Infrastructure, Transport and Tourism's i-Construction and the spread of BIM/CIM are also encouraging efforts to utilize digital data on site. However, typical smartphone GPS has meter-level errors, making it unsuitable for precisely overlaying drawings. The combination of high-precision GNSS (RTK) positioning and smartphone AR has emerged to address this. This makes it possible, even without dedicated equipment, to determine current position with centimeter-level accuracy and overlay drawing data on real space almost exactly.

Why overlaying drawings in AR speeds up site checks

Why does drawing AR overlay dramatically speed up preconstruction site checks? The main reason is that it allows you to intuitively compare the design with the actual condition on the spot.

• Significant reduction in surveying and marking work: Traditionally, confirming drawing positions on site required using surveying instruments to set points, staking out boundaries, or stringing control lines. With AR overlay, design lines and models are displayed at full scale on the smartphone screen, eliminating much of that manual work. For example, to check a site boundary or building placement, AR can reveal misalignments at a glance without having to draw provisional lines on the ground. As a result, you can drastically shorten the time spent on preparation and point verification.

• Immediate detection of discrepancies: By overlaying design data on the real scene in AR, any positional or dimensional discrepancies become visible immediately. In conventional workflows where manually measured values are later checked against drawings, finding mistakes could take time and oversights could occur. With AR, you can detect differences at that instant, making checks faster and reducing rework.

• Improved communication efficiency: Projecting drawings on site lets all stakeholders share the same "completed image" or "design intent." Instead of construction managers and workers huddling over paper drawings to explain things, reviewing the AR view together makes understanding intuitive and speeds up meetings and instructions. When clients or designers attend site inspections, viewing the image rather than verbal explanations generally leads to quicker agreement and faster decision-making.

In short, drawing AR overlay makes confirmation work real-time by "visualizing drawings on site," dramatically reducing time previously spent on surveying and reading drawings. This shortens preconstruction checks and as-built inspections, directly contributing to shorter schedules and improved quality.

Types of drawing data used and preparation methods (2D drawings, 3D models)

Drawing data for AR overlay can be broadly classified into 2D drawings and 3D models. Each has different preparation methods and use cases, so it’s useful to understand their characteristics.

• 2D drawing data: Refers to conventional two-dimensional CAD drawings such as plan views and cross-sections. CAD data like DXF/DWG or scanned images/PDFs of paper drawings can be used. To display a 2D drawing in AR, the drawing must basically have the correct scale and coordinate information set. For example, a CAD-generated plan should be drawn to real scale and positioned to match the survey coordinate system. Doing so allows on-site coordinates and drawing coordinates to correspond, enabling projection without additional on-device alignment. - Handling drawings without coordinates: Even drawings like paper plans that lack survey coordinates can be overlaid by performing on-site calibration in the AR app. For instance, using smartphone RTK to measure on-site feature points corresponding to drawing features (such as an intersection center or building corner), you can align the drawing by translating and rotating it to match those points. Matching two or more points also corrects scale errors, allowing local-coordinate or unscaled source drawings to be used for AR overlay by linking feature point coordinates obtained on site. - Use cases: 2D drawing AR overlay is effective for checking site boundaries and plan layouts. For example, at a residential land development site you can display plot lines in AR to confirm boundaries with adjacent lots, or overlay road centerlines to visually check curves and widths. Because you can reproduce a state where drawing lines appear on the ground as in the drawing, these checks become intuitive without staking or string lines.

• 3D model data: Includes BIM models of buildings and infrastructure, 3D CAD, and terrain surface models. Formats like IFC, LandXML, or vendor-specific 3D outputs are usable. For 3D models as well, it is important that they are created and placed in the correct coordinate system. If the model is positioned relative to reference points during design, the life-size model will be projected onto the site correctly. If the model was created with an arbitrary origin, preprocessing to translate/rotate it to the survey coordinate system is needed before AR use. - Model preparation tips: Highly detailed 3D models can be heavy for mobile devices, so export only the necessary portions or simplify models. Also align the zero-elevation reference appropriately—decide where ground elevation 0 m is. For a building model, align the GL (ground line) with the known elevation; for terrain models, ensure the model’s elevations match site elevations. Aligning vertical as well as horizontal directions prevents models from appearing to float or sink relative to the ground. - Use cases: 3D model AR projection is powerful for displaying a predicted completed structure on site at full scale. For instance, you can project a bridge or building BIM model onto the site to check clearances with surroundings, or render buried piping transparently to warn during excavation. Visualizing unseen parts (like rebar or internal structures) via AR helps prevent construction errors and supports as-built inspection.

Summary of drawing data preparation: In either case, the key to success is to "prepare drawings or models aligned with the site’s survey coordinate system." Ideally reflect public coordinates or reference point coordinates in drawings/models from the design stage. If only existing materials are available, you can still handle them by performing on-site calibration. Check units, coordinate system, and scale in advance, and simplify or hide unnecessary elements to make data AR-friendly for smoother operation.

Centimeter accuracy! Importance of high-precision alignment with LRTK

To perform AR overlay at a practical level, device positioning accuracy is critically important. Built-in smartphone GPS typically has errors of several meters, which makes precise overlay impossible. The key technology here is high-precision positioning by the real-time kinematic (RTK) method, and making that usable with a smartphone via technologies like LRTK.

• Centimeter-class positioning achievable with smartphone RTK: RTK-GNSS improves positioning accuracy by having both a base station and the rover (smartphone side) observe multiple satellite signals simultaneously; the base station computes error corrections that are used to correct the rover’s solution, achieving centimeter-level accuracy. Traditionally, dedicated survey GNSS equipment was required, but recently compact high-performance GNSS antennas attachable to smartphones have appeared. LRTK is one such solution: an RTK receiver integrated with a smartphone enabling centimeter-level positioning without external equipment.

• Effects of centimeter-level accuracy: RTK can achieve planar positioning accuracy around ±1–2 cm and vertical accuracy within a few centimeters. With this accuracy, differences between coordinates from design drawings and on-site measurements are minimized so much that the digital data and the real-world positions are visually indistinguishable. When drawing overlays are projected in AR, design lines and actual structures or terrain line up almost perfectly, yielding the truly practical "non-drifting AR."

• High accuracy eliminates the need for markers: Many AR apps use markers (like QR codes) as reference points for alignment, but these require placement effort and are only effective near the marker. Combining RTK-based precise self-positioning with calibration of the smartphone’s orientation enables absolute coordinate-based AR, fixing models in their correct locations without placing markers. Users can walk around and view from different angles without the model sliding or rotating — a stable AR projection enabled by high-precision positioning.

Thus, centimeter-class positioning is indispensable for accurate drawing AR overlay, and smartphone RTK solutions like LRTK satisfy that need. The importance of high-precision alignment becomes obvious when you experience AR display: meter-level offsets render AR unusable for practical work, whereas centimeter-level accuracy makes AR reliable for site checks, making LRTK-enabled high-precision AR a trump card for on-site DX.

Drawing AR overlay with a smartphone: practical step-by-step guide

Now let’s go through the concrete steps for displaying drawing overlays in AR using a smartphone (+ high-precision GNSS device). This presents a general workflow; details may vary depending on the app and equipment used.

• Necessary equipment and app preparation: Attach a high-precision GNSS antenna (e.g., an LRTK device) to the smartphone and launch a compatible surveying & AR display app. Pre-import drawing data (2D or 3D) into the app. As described above, the drawing data should be prepared in the survey coordinate system. Some apps allow project data to be synchronized via the cloud so that you can call it up on site.

• RTK setup: Upon arrival on site, receive correction information and start RTK positioning. If deploying a base station, place it in an open location and set a known coordinate (or connect to a virtual reference via network services), then establish communication with the rover (smartphone). If using a network RTK service, connect the smartphone to mobile data and log in to Ntrip, etc. Check the positioning status on the smartphone screen and ensure RTK has a Fix solution for centimeter accuracy. If it shows "Float" or "DGPS," accuracy is insufficient—improve signal conditions or reconfigure to obtain a Fix.

• Alignment operations: Align the imported drawing or model with the site as needed. If the drawing data already contains absolute coordinates, it should already appear roughly in the correct position. For finer adjustments or for drawings without coordinates, use methods such as: - Aligning using known points: Move the smartphone to a site point corresponding to a point on the drawing (e.g., a design reference or building corner) and perform the operation to link "this position = drawing point X." This translates and rotates the entire drawing to fit that point. - Correcting with multiple points: If possible, match two or more points. One point corrects translation and rotation, but a second point can help correct scale error as well—useful for scanned paper drawings that may have scale inaccuracies. - Checking height: For 3D models, adjust the model to align with known vertical references (benchmarks or existing structure elevations). Confirm that the model’s ground plane matches the actual ground; some apps allow Z-axis shifts to correct vertical misalignment.

• Start projection in AR mode: When ready, switch the device to AR display mode. The camera feed will show the drawing data overlaid in the previously aligned position. Lines may appear on the ground or a completed model may appear in space. First, look around 360 degrees to confirm that the model is displayed at the intended position. If there are nearby landmarks, compare their positions to the corresponding drawing positions to check for errors.

• Walk the site to verify: With the AR setup, walk around the site while holding the smartphone to verify various locations. Check whether building corners align with projected corners, verify road widths versus the overlay, and confirm piping routes are unobstructed—inspect by superimposing the AR drawing on the actual scene. Occasionally check that GNSS remains in Fix state on the screen; if buildings or trees obstruct the sky, accuracy may degrade, so pause and carefully check for display drift in such areas.

• Share findings and record them: When you find discrepancies between drawings and the site via AR, you can mark them and consider corrective measures on the spot. If stakeholders are nearby, show the live AR view for discussion. Record items to share later by taking screenshots or using the app’s photo/video capture to capture AR-displayed photos or videos. Position and timestamp metadata are often recorded automatically and can be used in reports.

• Finish up: After completing checks, close the app and pack equipment. Turn off GNSS devices and remove any base stations. Upload screenshots and positioning logs to the cloud or designated storage. Back at the site office, summarize and share any noted discrepancies so they can be incorporated into subsequent construction plans.

That is the basic workflow. While this is a simple smartphone-based surveying and AR setup, beginners may find it unfamiliar at first. Be mindful of the precautions described in the next chapter and start with trial operations to build familiarity. Once accustomed, it will be a powerful, ready-to-use tool for site checks.

On-site operational cautions (weather, visibility, satellite reception, etc.)

When using smartphone AR overlay outdoors, be aware of several important points. Understand the environmental conditions and device limitations to prevent trouble.

• Ensure suitable GNSS reception environment: RTK positioning requires good reception of satellite signals. In urban canyons surrounded by tall buildings, tunnels, or dense tree cover, satellites may not be sufficiently visible and positioning can degrade or stop. On site, try to position yourself where the sky is as open as possible and choose the base station location accordingly. For indoor spaces where GNSS cannot be used, consider alternative methods such as using known point coordinates to manually align the model.

• Maintain a Fix solution: Keeping a high-quality Fix RTK solution is essential. If the solution falls to a Float, accuracy may degrade to several tens of centimeters to a meter. Monitor the RTK status frequently on the app and do not proceed with work when the solution remains Float. If Float persists, reset the positioning or move to a more open area to reacquire Fix. If corrections are received over the network, also check mobile connectivity for outages or delays.

• Calibrate device sensors: The smartphone or tablet’s built-in sensors (gyroscope, accelerometer, magnetometer, camera-based pose estimation) are critical for AR. If the compass is off, the model’s orientation will be incorrect. Calibrate the compass by performing figure-eight motions and periodically reset gyroscope drift by placing the device on a level surface. Because bearing accuracy varies by device, verify orientation against multiple reference directions or use the sun or known azimuths to fine-tune.

• Deal with weather and lighting: Smartphone screens are hard to see in bright sunlight. On clear sunny days, reflections and insufficient contrast can make AR overlays hard to discern. Increase screen brightness, use a simple sunshade, or position the screen in the shade. Rain complicates handling even for waterproof devices; lens water droplets reduce recognition accuracy. Avoid using AR in severe weather—perform checks when weather is stable.

• Battery and heat management: Running GNSS and AR simultaneously increases power consumption. Carry a mobile battery for extended use. In cold climates battery performance can decline, while in hot weather the device can overheat and throttle or shut down. Avoid leaving devices in direct sunlight for long periods, periodically cool them, or use a small fan to manage temperature.

• Data size and rendering load: Displaying dense point clouds or highly detailed 3D models can cause long load times or app crashes. Pre-thin point clouds and simplify models to reduce load. Since smartphone performance varies, prepare range-limited models to ensure smooth operation on site.

• Safety considerations: Being absorbed in AR may distract you from situational hazards. Always watch your footing and be aware of nearby equipment and vehicles. Work in pairs where possible—one person operates the device while the other observes the surroundings. Avoid use at heights or on scaffolding; instead, conduct checks from the ground and move as needed.

By observing these points you can reduce site-specific troubles and obtain a more stable AR overlay experience. As the technology is relatively new, errors due to unfamiliarity may occur—understanding common failures and countermeasures, covered in the next chapter, is important.

Common failure cases and avoidance strategies

While drawing AR overlay is convenient, first-time operation can lead to certain common mistakes. Here are typical onsite failure cases and how to avoid them. Check these in advance to prevent repeating the same mistakes.

• Large offsets due to incorrect coordinate system: If drawing data or base station coordinates are set incorrectly, the AR display can be greatly offset. For example, selecting the wrong zone in a plane rectangular coordinate system can cause tens of meters of offset. Counter this by double-checking coordinate systems and reference point information, and if in doubt, test alignment with a known site point before full-scale use.

• Proceeding with Float RTK: Doing work despite insufficient positioning ("it’s probably fine") is a common mistake. Float solutions can yield tens of centimeters to meter-level errors, potentially requiring rework later. Always monitor RTK status and wait for Fix before proceeding.

• Incorrect device orientation: Compass errors or gyroscope drift can cause subtle orientation offsets. On wide sites, a 2–3° bearing error can produce tens of centimeters of offset at the far end of a long structure. Calibrate sensors before use and verify alignment against a straight reference feature. Use app-based yaw adjustment if available.

• Overlooking vertical alignment: Even if horizontal position matches, models can appear to float or be buried if vertical references are inconsistent. For instance, forgetting to set the design GL to the site’s known elevation can cause height differences of tens of centimeters to meters. Check the model’s vertical reference beforehand and perform Z-axis adjustments on site if needed, or at least know which elevation the displayed model references.

• Loading wrong data: Displaying an outdated drawing or a file from a different project can cause confusion. Ensure that the app contains the latest, correct files and clarify file and project names. When syncing via cloud, confirm uploads are up to date.

• Model too complex causing performance issues: Trying to display overly detailed 3D models or huge point clouds can freeze the app or overheat the device. Test with lightweight data beforehand to find device limits, then simplify or segment the model as needed to optimize data volume.

• Human errors in operations: Mistakes like selecting the wrong calibration point or mixing up points happen when operators rush. Before AR operations, reidentify reference points and confirm them verbally ("this point corresponds to drawing point X") to avoid mistakes. Cross-checks with multiple people are effective.

• Neglecting safety checks: AR-related accidents from focusing on the screen have been reported (nearly falling into ditches, missing heavy equipment). Prevent this by holding safety briefings on AR use areas and procedures, marking exclusion zones, and assigning spotters to guide safe movement.

Most of these failures can be largely prevented through preparation and verification. Start with small trial areas, identify problems, and use PDCA cycles to improve. Despite being advanced technology, good practice and careful checks remain essential—if in doubt, stop and verify.

Comparison with other methods (paper drawings, marker-based AR, conventional surveying)

To clarify the characteristics of drawing AR overlay, compare it with common traditional methods: paper drawings, marker-based AR, and human-centered conventional surveying.

• Site checks with paper drawings: This long-standing method is accessible to everyone, but mapping drawing positions to the physical site relies on the person’s internal spatial interpretation. Even skilled technicians can misread drawings, and translating drawing information into work on site takes time. AR overlay directly displays drawing information on site, significantly reducing the cognitive load for spatial understanding. Paper drawings are convenient because they don’t need power, but in terms of work efficiency and reducing human errors, AR has the advantage.

• Marker-based (or markerless) AR: Many AR apps use printed markers or plane detection as reference anchors to display models. For example, placing a printed marker sheet on the floor fixes the model’s position relative to the marker. While this maintains high relative accuracy around the marker, it requires precise marker placement and is unsuitable for large areas; if the user moves away the marker may no longer be visible. RTK-enabled AR eliminates the need for markers by using absolute coordinates, allowing wide-area inspection without re-alignment. Marker-based AR is still convenient indoors or for small-scale tasks, so choose the method that fits—RTK+AR for wide outdoor infrastructure sites, marker-based for small indoor objects.

• Conventional surveying and inspection: Traditional surveying crews using transits or total stations to stake out points and perform as-built checks are reliable but time-consuming and resource-heavy. For example, measuring 50 height points for an as-built inspection can take more than a full day. AR overlay allows real-time on-site checks so that obvious defects can be identified immediately and remedial work scheduled the same day—yielding faster response times. Precise numerical verification by survey instruments remains necessary for final checks, but AR is effective for upstream checks that reduce rework. AR also enables non-survey personnel like site supervisors or craftsmen to intuitively understand conditions, widening who can perform initial checks.

Overall, drawing AR overlay excels in "speed" and "intuitiveness," complementing and resolving issues with traditional methods. It’s not a total replacement—final precision measurements and official inspection records will still rely on conventional surveying—but AR streamlines the preliminary stages and updates site management practices.

Examples of AR integration with point cloud data and photogrammetry

In addition to design drawings and models, combining AR with on-site point cloud data and photogrammetry (SfM) enables more advanced uses. Here are representative examples of AR × point cloud integration.

• Difference checks between as-built point clouds and design models: It is increasingly common to measure completed structures or terrain with 3D scanners or drone photogrammetry to obtain point clouds. By overlaying these as-built point clouds with design 3D models and displaying deviation heatmaps, you can visualize where deviations are significant. With smartphone RTK like LRTK, point cloud points are tagged with absolute coordinates during acquisition, enabling comparison with design data without further alignment. Loading the generated heatmap into the smartphone and viewing it in AR lets you visually identify large deviations on site—for example, areas where fill is higher than design shown in red, making remediation points obvious. This heatmap AR allows immediate pass/fail judgment on site, speeding quality control cycles by enabling instant marking and repair.

• Visualizing buried utilities and photogrammetry: AR is also effective for visualizing underground pipes and cables. Create a 3D model of buried utilities from design data or past surveys and project it as a transparent subsurface image on site to reduce the risk of damage during excavation. After excavation reveals piping, measure its position with smartphone RTK, photograph it and convert to a point cloud to create data for future asset management. The resulting point clouds or photogrammetric models can be reprojected in AR after backfilling so that "invisible items become visible." Photogrammetry can be done with large-scale drone flights or simply multiple smartphone photos for lightweight point clouds. Using high-precision coordinates from LRTK ensures those photogrammetry results have accurate georeferencing, easing later comparisons and AR visualization.

• Progress monitoring and digital twin: Combining AR with point clouds is a stepping stone to digital twin implementation. Periodically scanning the site and updating the point cloud, then overlaying it in AR, allows on-site verification of progress and changes over time. For example, obtaining weekly point clouds and comparing them with design models in AR lets stakeholders discuss the current as-built status on site. This real-time, simplified process replaces the conventional flow of taking photos and returning to the office for analysis. Accumulated high-precision 3D data can eventually form a digital twin for remote monitoring or AI-based automatic detection; AR acts as the interface that connects the real world and the digital one.

As shown above, AR overlay is not just for design comparison—it is a tool for immediately leveraging various on-site digital data. Combining point clouds and photogrammetry with AR greatly increases the amount and accuracy of information available on site, improving construction management quality. Rather than a gadget, AR is becoming a data hub for the field.

Recommendation for introducing simple surveying & drawing overlay using LRTK

The drawing AR overlay and smartphone RTK surveying techniques described so far are no longer limited to cutting-edge sites. Required equipment and services have become affordable, and an environment where anyone can start on site is emerging. A representative example, repeatedly mentioned in this article, is LRTK.

LRTK is a system composed of an ultra-compact RTK-GNSS receiver and a smartphone app that turns a handheld smartphone into a versatile surveying instrument. Without costly specialized equipment, simply attaching an LRTK device to your smartphone enables centimeter-level positioning, and you can perform point cloud capture and AR overlay of design data. Tasks that used to require outsourcing to a surveying team can now be handled by site engineers themselves, reducing subcontracting costs and schedule coordination.

When introducing it for the first time, concerns like "Is the accuracy really sufficient?" or "Can it be used reliably on site?" are natural. We recommend starting with small-scale verification: trial LRTK in a limited area, compare discrepancies with conventional measurements, or run AR checks in parallel with traditional methods. Initial unfamiliarity is expected, but reviewing results with stakeholders builds understanding and trust in the system. LRTK providers typically offer support to address initial questions, and hands-on experience speeds staff training.

Sites that have introduced LRTK report outcomes such as "time spent on surveying and drawing checks has been cut by more than half," "we noticed defects that were overlooked on paper drawings," and "we could share AR images with distant supervisors and obtain quick instructions." One small smartphone device can directly improve productivity across the site. This approach aligns with the MLIT's DX and i-Construction initiatives, and such technologies are expected to become industry standard.

First, take DX’s first step at a familiar site. With LRTK you can incorporate cutting-edge centimeter-accurate surveying and drawing AR overlay into daily operations without special effort. Experience the benefits: faster preconstruction checks, more efficient as-built inspections, and reduced human errors. The future of construction management will feel much more accessible.

Frequently Asked Questions (FAQ)

Q: *What equipment is required to perform drawing AR overlay?* A: *Basically you need a smartphone, a high-precision GNSS receiver (RTK-capable device), and a compatible AR display app. A smartphone-mountable RTK antenna and dedicated app such as those used with LRTK are typical examples. The smartphone does not have to be the latest model, but it should support AR frameworks (ARKit or ARCore) and have adequate performance. A mobile battery is recommended for prolonged use.*

Q: *Can anyone handle it with just a smartphone? Is specialized training required?* A: *Compared to traditional surveying instruments, the operation is relatively intuitive, but pre-training and practice are recommended. Learning the app, RTK basics, and precautions in advance reduces confusion on site. The system is designed so that non-surveyors can operate it; there are cases where site supervisors or managers mastered it after a few hours of practical training. Start by trialing it with experienced personnel and expand usage as confidence grows.*

Q: *Do I need to set up a base station every time? Can it be used on sites without Internet?* A: *It depends on how you obtain RTK corrections. If there is a nearby public benchmark or you can use a network virtual reference station (VRS), you can achieve centimeter accuracy without setting up your own base, provided the smartphone has mobile Internet. In areas with poor connectivity, you can deploy your own base station and broadcast corrections wirelessly. In Japan, the quasi-zenith satellite "Michibiki" offers CLAS (centimeter-class augmentation) which, with a compatible receiver, can provide corrections without Internet. LRTK supports multiple correction methods to match site conditions.*

Q: *What accuracy can I expect when overlaying drawings? I’m worried about errors.* A: *Under good conditions, planar errors are typically within 1–2 cm, and vertical errors are within a few centimeters, so discrepancies are usually not noticeable to the eye. This assumes RTK is maintained in a Fix solution and device orientation sensors are properly calibrated. In environments with poor satellite reception or with sensor orientation errors, accuracy decreases. Therefore, validate accuracy using known points or landmarks to verify performance during use.*

Q: *Can AR overlay be used at night or in dark areas?* A: *Positioning itself remains available at night, but if the camera feed lacks sufficient light the AR overlay becomes hard to see. AR pose estimation relies on tracking visual features in the camera feed, so complete darkness hampers stability. For dark sites, provide lighting such as work lights, or use devices with LiDAR (e.g., some iPad Pro models) which can maintain better AR stability in lower light. For safety and practicality, daytime use is preferable; in dark conditions rely less on AR and more on previous daytime inspections.*

Q: *What about indoor use or places where GPS is unavailable?* A: *Indoors or underground where GNSS is unavailable, you cannot get absolute RTK positions. Alternatives include using known-point local-coordinate AR, where points established by total station are used as references and the smartphone’s position is manually set relative to them. UWB or Visual SLAM-based indoor positioning systems are also emerging. Achieving centimeter-level accuracy indoors is still challenging, but plane recognition or marker-based AR can substitute depending on the application. The key is to combine other positioning methods where GNSS is not available.*

Q: *I’m worried about introduction costs—will it be expensive?* A: *Compared with traditional surveying instruments or 3D scanners, a smartphone plus an RTK receiver is a relatively cost-effective option. Prices vary by model and service, but it’s often only a fraction of the cost of a full survey GNSS set. Savings also come from reduced outsourcing and fewer reworks. Return on investment is often realized quickly when considering labor and time savings. Subscription-type services can lower initial costs and allow trial introductions before scaling up. Start small, evaluate benefits, and expand gradually.*