High-precision Earthwork Volume Calculation with Point Clouds: Achieving Within 1% Construction Tolerance with LRTK

Table of Contents

• The relationship between earthwork volume calculation and point cloud data

• Characteristics and challenges of on-site photogrammetry

• The complete flow from point cloud generation to volume calculation and required tools

• Comparison with traditional methods (total station and GNSS alone) — accuracy, personnel, time, and reporting

• Time-series comparison of earthwork volume changes (embankment, excavation, visualization of design differences)

• Smartphone-based point cloud acquisition with LRTK and its benefits (capture assistance, positioning, cloud sync)

• On-site use cases and integration with daily reports and deliverables

• FAQ

The relationship between earthwork volume calculation and point cloud data

In civil construction sites, calculating earthwork volumes is an indispensable task for quality control. Traditionally, terrain survey data were used to create cross-sections and volumes were calculated by the average-end-area method or grid methods. In recent years, however, the use of point cloud data (three-dimensional data composed of countless coordinate points) has enabled more efficient and higher-precision earthwork volume calculations. Point clouds acquired by 3D laser scanners or photogrammetry can record fine surface undulations, allowing reconstruction of the ground surface almost exactly as it is. By comparing point cloud data from before and after construction, embankment and excavation volumes can be calculated directly.

The principle of volume calculation using point clouds is simple: compute the volumetric difference between surface models before and after construction. For example, in excavation work, compare the point cloud of the pre-excavation surface with the post-excavation surface to calculate the volume of removed soil. Because point clouds can reconstruct the terrain as a surface from innumerable measured points, there is no need to interpolate between sparse survey points as in traditional methods, enabling accurate quantity estimation that fully reflects surface undulations. Also, once a point cloud is acquired, you can recalculate volumes with different calculation extents or reference planes without additional surveying, providing flexibility for reanalysis or scenario testing. For these reasons, point cloud–based volume calculation excels in both accuracy and efficiency and is becoming a foundational technology supporting digital construction management.

To achieve higher-precision volume estimates, ensuring the quality of point cloud data is crucial: acquire sufficiently dense point clouds without gaps in the capture area, ensure coordinates correctly match the reference coordinate system, and appropriately remove or process unwanted objects (machinery, trees, etc.). When these conditions are met, low-error volume calculations are possible. Field verifications have reported cases where quantities computed from point clouds differed from traditional manual surveys by about 1%, indicating that with proper procedures the reliability of point cloud volume calculations is high.

Characteristics and challenges of on-site photogrammetry

Because point cloud data are useful for earthwork management, a key onsite issue is how easily the point cloud can be acquired. Traditionally, specialized survey equipment (terrestrial laser scanners or surveying drones) and surveying teams were required, but advances in photogrammetry have increased cases where site staff themselves acquire point clouds using smartphones or drones. On-site photogrammetry refers to the method of easily taking photos in the field and generating point clouds from those images via 3D reconstruction.

The advantage of photogrammetry is the ease with which wide areas can be measured using familiar devices. Drone aerial photography can capture large areas quickly from above, surveying steep terrain or hazardous slopes without personnel entry. Smartphone photography can also produce high-density point cloud models with dedicated software if sufficient images are taken around the target. Being able to generate point clouds with ordinary cameras rather than expensive laser scanners is a major benefit and positions photogrammetry as a promising DX tool for construction sites.

However, there are several challenges with photogrammetric point cloud capture. First, measurement accuracy depends on shooting conditions. If photo resolution or exposure is insufficient, or if subjects contain many shadows or reflections, errors or gaps can arise during point cloud generation. If vegetation or debris is included in the images, extracting only the ground surface from the generated point cloud can be time-consuming. Additionally, the workflow can be difficult to complete fully on site. Even if shooting is quick, photo processing (point cloud generation) may require high-performance PCs or cloud services and take several hours. Traditional photogrammetry can introduce a time lag from surveying to data processing to volume calculation, making it hard to get real-time results on site. Drone use also requires flight permission and pilot skill, and smartphone-only photogrammetry risks misalignment between the model and the real-world coordinate system due to GPS errors. Even if point clouds can be generated from smartphone photos, if the elevation reference and position are unclear, they cannot be used for construction quantity management. Therefore, practical use of photogrammetry in construction has required extra steps to ensure accuracy, such as tying to survey control points or ensuring sufficient photo overlap. While on-site photogrammetry has high potential, conventional technologies alone have not fully realized the ideal of “anyone easily completing high-precision point cloud measurement on site.”

The complete flow from point cloud generation to volume calculation and required tools

What steps and tools are needed to generate point cloud data and calculate earthwork volumes? Here is a typical workflow.

• Data acquisition (measurement): Record the terrain and soil conditions by photographing the site with a smartphone camera, LiDAR scanner, or drone-mounted camera. For photogrammetry, take sufficient images from all directions around the subject and ensure overlap between photos. For drones, use an automated flight plan to capture images at set intervals; for smartphones, walk around the subject to capture images without gaps. Some modern smartphones (for example, models equipped with LiDAR) can complete point cloud capture in real time.

• Point cloud generation (processing): Use dedicated photogrammetry software or cloud services to create point cloud models from photos. When photos are input to dedicated software, feature point matching and angle calculations recover the camera positions and 3D point coordinates, producing point clouds numbering in the millions to tens of millions of points. This processing requires significant computational resources and time, but high-performance PCs or GPUs can speed up processing and improve accuracy. Recently, cloud services that automatically generate point clouds simply by uploading photos have appeared—if images are sent from the field, the point cloud may be ready by the time you return to the office.

• Georeferencing: To use generated point clouds for volume calculations, they must be properly positioned in the real survey coordinate system. Georeferencing means assigning scale, orientation, and elevation references to the point cloud model. For drone photos, pre‑placed ground control points (GCPs) are captured in images and used to align the model to real-world coordinates. For smartphone photos, you can photograph known site points and match them in post-processing, or, as described later, acquire data with inherent high-precision positioning that provides absolute coordinates from the start. For comparing point clouds taken at different times or checking against design data, having all data on a common coordinate foundation is essential. Properly performing this step makes point clouds useful measurement data in GIS or CAD workflows.

• Volume calculation: With aligned point clouds, specify the area to calculate and perform the computation. Typically, a TIN (triangulated irregular network) terrain model derived from the point cloud is used for volume calculations. Specifically, volumes of fill or cut are integrated relative to a reference plane, or two different terrain point clouds are overlapped as differential point clouds to compute volumetric differences. Civil-oriented 3D software or point cloud processing tools are used for these calculations, but once the procedure is set up, software can automatically output volume values. For example, specifying a polygon for an embankment area lets software quantify the vertical volumes within that polygon. Open-source point cloud tools with volume functions and web-based cloud services that display volumes in a browser have emerged, enabling volume checks in a browser without specialized software.

• Sharing deliverables and reporting: Share calculated volumes on-site or compile them into reports for clients and stakeholders. Traditionally, results were organized in Excel and cross-section drawings and tables were prepared manually. With point cloud utilization, you can attach visualizations from a 3D viewer or share cloud-hosted interactive links so stakeholders can intuitively inspect the site. Being able to share the full 3D data that underlies the calculation is a major advantage of point clouds. The toolset supporting this workflow includes capture devices (smartphones, drones), photogrammetry conversion software, point cloud processing tools or cloud services, and viewers/sharing platforms. Recently, integrated solutions that combine these previously separate tools have appeared.

Comparison with traditional methods (total station and GNSS alone) — accuracy, personnel, time, and reporting

Approaches using point clouds differ substantially from traditional methods that rely only on total stations (TS) or GNSS surveying equipment. Let’s compare accuracy, staffing, time required, and reporting.

• Accuracy: For individual point positioning, TS and high-precision GNSS generally excel. TS distance measurement to prisms can achieve millimeter-level accuracy, and RTK-GNSS with a base station can limit horizontal errors to within a few centimeters (within a few in). However, “accuracy” in earthwork volume calculation is not just about small point errors but also about how well the overall terrain shape is captured. TS/GNSS surveys typically sample points on a grid at intervals of several meters (several ft) or place points along critical cross-sections; while measured points themselves are accurate, fine undulations between survey points must be interpolated and may be missed. Point clouds, by contrast, cover the terrain continuously with a massive number of points, capturing details everywhere. Therefore, small depressions or bulges can be detected by point clouds but might be missed by coarse traditional meshes. Overall volumetric error depends on such undetected features, so under comparable conditions point cloud–based estimates can be as accurate as traditional methods. In fact, some large embankment quality checks reported that volumes calculated from point clouds differed from average-end-area methods by about 1%, confirming adequate accuracy of point clouds.

• Personnel and time: Traditional workflows require many manual steps from surveying to drawing creation and volume computation. For a large site, a survey team of four might spend a full week (20–30 person-days) surveying and producing cross-sections and volumes. Switching to photogrammetry plus point cloud processing, some cases completed the task with two people in one day (2 person-days). A drone captured images in about 15 minutes and point cloud generation and volume computation were done the same day. Point cloud workflows dramatically reduce personnel and time requirements: a single operator can operate a smartphone or drone, allowing others to handle different tasks. With a shortage of experienced surveyors, easy-to-use digital measurement methods help address manpower shortages.

• Reporting: Report preparation also becomes much more efficient. Traditional reporting required manual compilation of drawings and tables and often involved internal verification and client inspections. Point cloud data itself can serve as evidence: presenting the 3D model lets stakeholders intuitively understand site conditions and visually verify calculation bases. Sharing a colored differential map or overlay of volumes in a point cloud viewer immediately communicates where material is lacking or in excess. Automatically generated cross-sections and plan views from point clouds can be directly embedded into reports, greatly reducing workload. In short, reporting can be semi-automated. Point cloud utilization accelerates the entire cycle from surveying to reporting, enabling faster on-site decision-making.

Time-series comparison of earthwork volume changes (embankment, excavation, visualization of design differences)

Another strength of point cloud–based volume calculation is that it can capture terrain changes over time. Construction sites change daily, and regularly acquired point clouds allow quantitative tracking of embankment and excavation progress.

• Progress management: For example, if a site is drone-scanned every weekend and models are saved weekly, you can graph weekly increases in embankment volume or display color-coded maps showing where fill increased compared with the previous week. Easier time-series comparisons let you manage construction progress with objective data. If progress lags, you can promptly arrange additional equipment or trucks, improving the PDCA cycle in construction management.

• Design-difference verification: From the perspective of as-built control, comparing point clouds with design models is important. With point clouds and the design completion model (e.g., design ground surface), you can check whether the site matches the design. For excavation, you can identify areas excavated to the required depth versus areas where residual soil remains; for embankments, you can detect locations that exceed specified heights. Differences can be visualized as heat maps (color distribution) for intuitive understanding—for example, red for higher-than-design and blue for lower-than-design—making it easy to identify spots requiring rework. Spatially indicating excesses and shortages of cut and fill smooths instruction of corrections and helps prevent rework and material loss.

• Maintenance and disaster response: Accumulated time-series data are useful not only during construction but also for maintenance and disaster response after completion. Storing as-built point clouds at project handover allows quantitative assessment of long-term changes when comparing future inspections to the original dataset. For example, settlement of embankments or earthworks can be quantified from differences with prior point clouds, and collapse volumes in slope failures can be estimated from pre- and post-event point clouds. Traditionally, post-disaster surveys required on-site measurements, but remote point cloud measurement can rapidly quantify material volumes even in hazardous areas, improving safety and response speed. Thus, differential use of point clouds provides value across preconstruction, construction, and post-construction stages beyond mere volume management.

Smartphone-based point cloud acquisition with LRTK and its benefits (capture assistance, positioning, cloud sync)

As noted, on-site photogrammetry had challenges, but recent technologies have overcome many of them and are creating environments where “anyone can easily acquire high-precision point clouds.” A representative example is LRTK. LRTK is a smartphone-integrated high-precision positioning system provided by Refyxia Inc. By attaching a dedicated small antenna to an iPhone and using network RTK corrections, LRTK improves smartphone positioning to centimeter-level accuracy (half-inch accuracy). This enables each point in point clouds acquired by a phone’s camera or LiDAR to have accurate coordinates, allowing smartphones to achieve survey-grade precision. Previously, high-precision 3D surveys required drone+GNSS reference setups or expensive laser scanners, but LRTK can substitute with a single handheld smartphone—revolutionary for field workflows. It requires neither specialized equipment operation skills nor complex setup, so field technicians can use it as an extension of familiar workflows. Compared with other 3D measurement methods, LRTK has lower introduction costs, requires no vehicle mounting or power setup, and offers the agility to measure immediately when needed, making it suitable for frequent routine measurements.

Organizing LRTK benefits from the perspectives of capture assistance, high-precision positioning, and cloud synchronization:

• Capture assistance: The LRTK smartphone app generates point clouds in real time and displays them on-screen, enabling operators to fill gaps while capturing. If a slope is scanned and an occluded area does not appear in the point cloud, the operator can immediately capture additional images to complete coverage. Even with multiple scans, datasets are automatically aligned, so high-quality point clouds can be obtained without specialized post-processing. The app also provides guidance for optimal capture trajectories and features for repeatable fixed-point observations, allowing anyone on site to capture reliably.

• High-precision positioning: LRTK’s main feature is improving smartphone GNSS accuracy. The LRTK antenna, compatible with network RTK, reduces typical smartphone GPS errors of several meters to within a few centimeters (within a few in). This adds global coordinates to all captured point clouds and photos, enabling immediate use for volume calculations and drawing comparisons without GCP correction. Elevation information is also precise enough for measuring height differences and generating cross-sections. For example, LiDAR scans from an iPhone Pro combined with LRTK produce point clouds that already have high-precision coordinates, allowing direct use in as-built management. Eliminating the traditional georeferencing step simplifies the entire workflow.

• Cloud synchronization: LRTK integrates a field app with cloud services, enabling automatic sharing and storage. Point clouds and georeferenced photos captured in the app are uploaded to the cloud immediately after capture, removing the need to copy files via USB or perform manual conversions in the office. Uploaded point clouds can be viewed in a 3D viewer on the cloud, and web-based tools can compute volumes or generate diagrams. For example, the workflow “smartphone scan on site → cloud automatic processing → instant volume results” is achievable end-to-end, minimizing lag between capture and analysis. Storing data in the cloud centralizes terrain changes across the project and makes time-series comparisons easy. Past point clouds can be retrieved for reference, and stakeholders can share links for joint review. Instant sharing of site information that was difficult with paper reports enhances construction management speed and reduces labor.

Thus, LRTK is an integrated system that enables “smartphone point cloud capture → on-site high-precision volume calculation → cloud data sharing.” It packages solutions to the technical hurdles of on-site photogrammetry (capture skills, positioning accuracy, data processing) and directly supports labor reduction and efficiency gains in the field.

On-site use cases and integration with daily reports and deliverables

Finally, here are practical on-site use cases for point cloud earthwork measurement and how to leverage them in daily reports and quality-control reports.

• Progress management and daily reports: On a certain development site, the site supervisor scanned spoil piles with a smartphone and LRTK each evening to immediately quantify that day’s removed soil. The values were recorded in the daily report and used to decide the next day’s equipment and dump truck allocations. Daily soil quantities, previously roughly estimated from truck counts and capacities, can now be recorded as measured values, improving daily report reliability and facilitating smoother coordination among contractors. Weekly or monthly accumulated point cloud data can also serve as objective evidence of progress and productivity for internal and external reporting and help reassess construction schedules.

• As-built verification and report generation: At embankment completion, compare LRTK-acquired point clouds with the design surface to verify as-built volumes. Automatically computed cut-and-fill volumes can be used directly as as-built quantity tables, and cross-sections or 3D views extracted from point clouds can be attached to reports for client explanations. With the immutable evidence of a point cloud, clients can immediately confirm quantities without re-measuring on-site, eliminating repetitive on-site checks. Because the LRTK cloud can generate required diagrams from captured data, on-site inspections can lead to instant confirmation and agreement.

Introducing LRTK-based quick surveys on site enables immediate capture and sharing of results. This saves significant personnel and time while maintaining precision and reliability in as-built control, contributing to productivity gains, cost reduction, and enhanced safety across the project. By internalizing measurement tasks previously delegated to specialists, construction management speed dramatically improves. This is truly the effect of “on-site completion of earthwork quantity calculation and point cloud measurement.” Smartphone-based point cloud capture and real-time as-built management are becoming standard processes in future civil construction. Please try smartphone scanning and LRTK-enabled streamlined earthwork management at your site.

FAQ

Q1. Is the accuracy of earthwork volume calculation using point clouds acceptable? A1. Yes—if point cloud data are properly captured and processed, accurate volume calculations are possible. Generally, point clouds from photogrammetry or laser scanners, when calibrated with control points and captured at sufficient density, yield volumetric errors comparable to traditional surveying (within a few percent). Field comparisons have reported only very small differences (around 1%) between point cloud–derived quantities and traditional methods. However, ensuring high accuracy requires no gaps in the capture area, removal of non-ground points, and correct coordinate alignment. When these conditions are met, point cloud volume calculations are sufficiently reliable for site use.

Q2. Do I need specialized skills to handle photogrammetry or LRTK? A2. Traditional photogrammetry sometimes required expertise, but recent solutions have become easy to operate. Smartphone scanning apps are intuitive—follow on-screen instructions to capture images and generate point clouds—so special shooting skills are not essential. LRTK simply requires attaching the antenna to a supported smartphone, launching the app, and following guided movements; the system then automatically performs positioning, point cloud capture, and volume calculations. Users don’t need to master technical terms or complex settings, and short on-the-job training is sufficient. Because data processing and analysis are automated in the cloud, users mainly need to review results. Modern photogrammetry tools and LRTK are designed for ease of use and can be operated by typical site personnel.

Q3. What equipment and environment are required for on-site point cloud measurement? A3. Essentially, a device with a high-quality camera (smartphone, tablet, drone) and supporting software/services are sufficient for point cloud measurement. For smartphones, LiDAR-equipped models such as the latest iPhone or iPad are preferable, though photogrammetry can work with ordinary cameras. Using LRTK requires a compatible smartphone, the LRTK antenna, and a mobile network connection to receive RTK corrections. Drone surveys require a GPS-equipped drone, camera, flight permissions, and a qualified pilot. In all cases, cloud services or PC-based point cloud software are needed to process captured data. However, all-in-one services like LRTK can complete capture-to-cloud storage with just a smartphone and antenna, eliminating the need for a dedicated PC. For capture environment, good visibility and sufficient shooting positions are important to cover wide areas. Photogrammetry is affected by weather and lighting, so performing captures in clear, well-lit conditions improves accuracy. For safety, avoid having personnel enter risky slopes—use drones—or use monopods or poles to measure from unsafe spots, ensuring the measurer’s safety.

Q4. Which should I use for earthwork measurement: smartphone scanning or drone surveying? A4. Choose based on site scale and purpose. For large sites or areas with many high elevations, drone surveying is efficient for quickly obtaining an overview point cloud. For confined sites, indoor areas, or detailed measurements, smartphone scanning provides agility. Smartphones require no takeoff/landing area and can be used in no-fly zones, making them convenient for frequent, routine measurements. In practice, both are often combined: use drones for a site-wide baseline point cloud and smartphone+LRTK to supplement and monitor detailed or changing spots. Both point clouds can be merged on a common coordinate system, so select the optimal method according to site conditions, desired accuracy, and measurement frequency. Either way, volumes can be calculated from the resulting point clouds, so choose according to your needs.