Table of Contents

• What is 3D design data?

• What is point cloud data?

• What is difference visualization between 3D design data and point cloud data?

• Benefits of point cloud difference visualization

• Key points for using difference visualization on-site

• Simple surveying with LRTK

• Conclusion

• FAQ

In the construction industry, labor shortages due to a declining birthrate and aging population and workplace issues known as the "3 Ks" (hard, dirty, dangerous) have created numerous challenges, making productivity improvement an urgent theme. The Ministry of Land, Infrastructure, Transport and Tourism has set a target to improve on-site productivity by 20% by fiscal 2025 through the use of ICT (information and communication technology) and three-dimensional data (the promotion of so-called i-Construction), so digital transformation on site is imperative. In this context, the technique of combining the design-stage 3D design data with point cloud data acquired on-site and performing "difference visualization" has attracted significant attention. By comparing the 3D model with measured field data to visualize deviations, early detection of construction errors, reduction of rework, and more efficient as-built management can be achieved, promising a dramatic improvement in productivity and quality.

This article, titled "Comprehensive Guide! On-site Utilization Guide for 3D Design Data and Point Cloud Difference Visualization," explains in detail the basics of 3D design data and point cloud difference visualization, its benefits, and concrete points for using it on site. At the end of the article, we also introduce a revolutionary simple surveying system using a smartphone, LRTK, and methods for on-site measurement and difference visualization.

What is 3D design data?

3D design data refers to design models created in three dimensions for buildings and civil engineering structures. Unlike traditional two-dimensional drawings, these are digital data representing the shapes of structures and terrain on XYZ three-dimensional coordinates and are also referred to in the construction field as building BIM (Building Information Modeling) and civil CIM (Construction Information Modeling). 3D design data contains detailed dimensions and shapes of the finished forms of buildings, bridges, roads, and so on, and can be thought of as a "digital rendering of the expected finished product."

In recent years, national and local governments have been promoting the use of BIM/CIM, and in large-scale projects the creation of 3D models from the design stage is becoming more common. Even where only existing 2D drawings are available, creating a 3D model beforehand based on them makes later comparison with point cloud data (difference visualization) easier. 3D design data is an important dataset that serves as the basis for construction planning and as-built management, helping with advance checks for clashes that are hard to see in drawings and with sharing the finished image among stakeholders.

What is point cloud data?

Point cloud data is three-dimensional data that represents the shape of an object or terrain using innumerable points in space. Each point includes X, Y, and Z coordinates (and attributes such as color information and return intensity), and by analyzing this massive collection of points you can digitally reproduce terrain and structures with high accuracy and detail. It is essentially a "digital full copy of the site" captured by countless points, and a major characteristic is that it can record complex shapes that flat drawings or a few survey points cannot capture, in three dimensions just as they are.

Point cloud data is mainly acquired using methods such as laser scanners (LiDAR) and photogrammetry. Laser scanners obtain numerous distance points by emitting laser light and analyzing the returned reflections. There are various methods suited to different applications, such as high-precision terrestrial laser surveys using dedicated equipment, UAV laser surveys that mount LiDAR on drones to measure wide areas from the air, and MMS (mobile mapping systems) that perform mobile measurements by mounting equipment on vehicles. In photogrammetry, photos taken from multiple angles by drones or single-lens cameras are processed by software to generate 3D point clouds. Recent advances in SfM (Structure from Motion) technology make it possible to generate high-resolution point clouds even from photos. More recently, attempts have begun to use built-in LiDAR sensors in devices such as iPhones and iPads to perform easy point cloud scans with smartphones. By utilizing point cloud data acquired through these diverse means, it has become possible to capture the site with vastly greater information density than before.

What is difference visualization between 3D design data and point cloud data?

Difference visualization is a method that overlays the ideal geometry in the design data with point cloud data acquired on-site and visually represents the deviation (difference) between them. Specifically, the as-built point cloud of structures or terrain is displayed in the same coordinate system as the corresponding design 3D model (BIM/CIM model or 3D design drawing), and each part is color-coded or otherwise marked to show how far it deviates from the design values. This kind of difference visualization allows you to grasp at a glance the quality and accuracy of construction and as-built status.

A representative method of difference visualization is heat map analysis. In a heat map, each point on the point cloud is given a color corresponding to the error amount from the design surface. For example, areas finished as designed might be colored green, areas that are overfilled and therefore high colored red, and areas that are over-excavated and therefore low colored blue. This makes it intuitive to see across the entire point cloud where things are within specifications and where they are out of tolerance. By performing difference visualization between the 3D model and the point cloud, you can spatially understand the overall as-built condition without having to compare lists of numbers or 2D drawings.

Benefits of point cloud difference visualization

Introducing point cloud difference visualization brings various benefits to on-site operations. The main effects are summarized below.

• Visualization of construction accuracy and improved quality: Point cloud differences allow you to quantitatively visualize construction accuracy. If you acquire point clouds during construction and compare them with the design model, you can detect minor dimensional errors or missed work on-site immediately and prevent major rework by early correction. Also, because you can comprehensively grasp the as-built condition rather than checking only a few measurement points as was done traditionally, you can significantly reduce quality defects caused by oversight.

• Efficiency of surveying and measurement work: Earthwork volume calculations for large embankments and excavations can be computed quickly from point cloud data. While conventional methods estimated volumes partially from cross-section surveys, point clouds allow a computer to calculate the volume difference between the as-built terrain and the design terrain in bulk, enabling same-day results on-site. For example, cases have been reported where embankment volumes were recalculated on the spot from drone photogrammetry point clouds and dump truck dispatches were adjusted immediately. Frequent surveying no longer consumes as much manpower and time, contributing to shorter schedules and cost savings.

• Improved safety: Point clouds are effective for verifying as-built conditions in hazardous areas where people cannot approach. Even steep slopes or the backsides of high bridge locations can be measured remotely by laser scanning or drone imaging, eliminating the need for personnel to take risks. This allows both safety assurance and as-built data acquisition to be achieved.

• Smoother consensus building and reporting: The visualization results of point cloud differences facilitate communication among stakeholders. Presenting colored 3D data to clients and designers makes it easier for them to intuitively understand the finished condition that was difficult to convey with paper drawings or numerical tables. Point cloud data can also be retained as objective evidence, making it easier to convincingly demonstrate quality during inspections and reducing the time required to prepare as-built inspection documents. Additionally, it is now possible to confirm differences between design models and point clouds on site with AR display on smartphones or tablets and take photos for records, simplifying the preparation of reporting materials.

• Progress management and remote sharing: Regular point cloud scans enable visualization of construction progress and remote monitoring. For example, if the entire site is scanned weekly by drone and point clouds are accumulated over time, you can compare as-built conditions at different times in 3D to intuitively grasp construction progress. Sharing point cloud data on the cloud allows headquarters or clients to check detailed conditions without visiting the site, speeding approval and reporting processes. This reduces the burden on site supervisors and makes it possible to perform real-time quality control even for remote projects.

Key points for using difference visualization on-site

To effectively use difference visualization on real sites, here are several key points to keep in mind.

• Unify coordinate systems: To correctly compare design data and point cloud data, both must be located on the same reference coordinate system. If you establish reference points beforehand and align measurements to those coordinates when acquiring point clouds, you can overlay the obtained point cloud and the design model without additional adjustments. Recently, methods that use RTK-GNSS to assign absolute coordinates at the time of point cloud acquisition and automatically convert to a known coordinate system have become widespread, allowing design data to match the point cloud without cumbersome post-processing.

• High-accuracy on-site measurement: The accuracy of the point cloud data itself is important to effectively perform difference visualization. While acquiring data with a high-precision laser scanner is ideal, select appropriate measurement means according to site scale and budget. For broad earthwork sites, drones; for detailed checks of structures, terrestrial scanners or smartphone LiDAR—using equipment appropriate to the purpose will efficiently yield high-quality point clouds. Timing of measurement is also important; scanning at key construction milestones allows later verification of the as-built condition at any chosen point in time.

• Use difference analysis tools: To compare the acquired point clouds with design data, use specialized software or cloud services. A variety of tools have appeared recently, from point cloud processing software that runs on PCs to cloud services accessible via browser, some of which automatically perform noise removal and compute differences from the design with button operations. By utilizing such tools, even those without specialized knowledge can obtain visualizations like heat maps in a short time, allowing site personnel themselves to perform difference analysis.

• Data sharing and on-site utilization: Share and utilize point cloud data and the results of difference visualization with the site and stakeholders. Uploading point clouds to a cloud platform lets distant managers or clients view the same 3D difference data in real time. Advanced use cases have also emerged where the difference between the design model and point cloud is displayed in AR on tablets or smartphones for direct on-site checking. Make effective use of the obtained 3D data to support on-site decision-making and communication.

Simple surveying with LRTK

Finally, we introduce the innovative simple surveying system LRTK, which uses a smartphone. LRTK (pronounced L-R-T-K) is a solution provided by Refixia Inc., a startup based in Minato Ward, Tokyo. By attaching a compact RTK-GNSS-compatible antenna to an iPhone or other smartphone and using a dedicated app, it transforms a smartphone into a centimeter-level accuracy (half-inch accuracy) surveying instrument. Combining the smartphone's built-in LiDAR sensor with high-precision positioning makes it possible for anyone to easily acquire high-accuracy 3D point clouds on site.

With LRTK, acquired point cloud data is synchronized to the cloud immediately, facilitating smooth comparison and analysis with design models. If you upload the design 3D data to the LRTK cloud, automatic matching with on-site measured point clouds and heat map generation can be done with a few clicks. Because the point clouds have absolute coordinates attached, there is no need for time-consuming alignment, and difference visualization results can be obtained in a short time. Furthermore, if you download that difference heat map to a smartphone and display it in AR, you can overlay it on the actual site to confirm construction accuracy. For example, by holding up a smartphone showing the heat map, you can immediately see color distributions indicating areas that are too high or too low, allowing you to instantly identify defective areas that used to require staking out and begin corrective work right away. In addition, the LRTK app lets you measure distances, areas, and volumes from point clouds and share acquired data with stakeholders via the cloud, enabling end-to-end completion of on-site measurement, analysis, and reporting on a smartphone. LRTK, which can be operated without specialized equipment or advanced skills, is truly a reliable tool for easily putting the point cloud difference visualization discussed in this article into practice on-site.

Conclusion

3D design data and point cloud difference visualization are digital technologies that greatly contribute to improving on-site productivity and ensuring quality. By visualizing deviations between the design model and the as-built condition, you can achieve reduced rework and shorter schedules, advanced as-built management, and shared understanding among stakeholders, among other benefits. 3D measurement and analysis, which previously relied on specialized companies and expensive equipment, can now be practiced easily on site by anyone using solutions like LRTK. If you have not yet introduced difference visualization, consider adopting cutting-edge 3D technologies on your sites and taking the first step toward construction DX.

FAQ

Q: What preparations and data are required to perform difference visualization? A: You need a 3D design model and on-site point cloud data. Prepare the design 3D data (BIM/CIM, etc.) in advance and acquire point clouds on-site. You must ensure the two datasets are in the same coordinate system during measurement or later adjust the point cloud to the design data based on reference points. Also prepare software or cloud services to compare them. Recently, services that automatically analyze differences simply by uploading design data and point clouds to the cloud have become available.

Q: In what situations is point cloud difference visualization useful? A: It is mainly useful for construction management and as-built inspection. For example, measuring as-built conditions of concrete structures with point clouds and comparing them to the design model verifies dimensional accuracy of finishes. It is also broadly effective in situations requiring comparison with design drawings, such as calculating embankment or excavation volumes in earthworks, checking pavement thickness in road works, and confirming pipe positions in equipment installation. In addition, point cloud difference comparison is applied to progress management during construction and to post-completion maintenance management (monitoring aging changes).

Q: Is smartphone LiDAR scanning sufficiently accurate? A: It depends on the application, but using the latest smartphones plus RTK enables acquisition of point clouds with an accuracy on the order of several centimeters (several inches). While it does not match conventional high-precision laser scanners' millimeter-level accuracy (0.04 in), smartphone LiDAR is sufficiently practical for as-built management in civil engineering and earthwork volume measurements. Systems like LRTK that improve positioning accuracy with RTK-GNSS make it easy on-site to achieve accuracy approaching that of conventional specialized equipment.

Q: Point cloud data files are large; how should they be managed and shared? A: Point cloud data can contain millions to hundreds of millions of points, so file sizes become very large. When using PC-based dedicated software, a high-performance machine and large storage are needed, but recent cloud services allow handling large datasets regardless of local PC specifications. Uploading point clouds to the cloud enables viewing and measurement via web browsers and sharing 3D data simply by sharing a URL. Storing data on an internal NAS or establishing data management rules per project is also important.

Q: Is it difficult or expensive to introduce point clouds and difference visualization? A: It used to require costly equipment and specialized knowledge, but recently easy measurement methods using drones and smartphones have emerged and tool usability has improved. For example, with a smartphone plus LRTK you can greatly reduce initial costs compared to traditional laser scanners while enabling your in-house engineers to perform 3D measurement and difference analysis. Start by trialing it on small sites to experience the benefits and then roll it out internally. Manufacturers also offer support and training, so the barrier to entry is not high.