Complete Guide to Cloud-Based Point Cloud Generation Services｜Thorough Explanation of the Workflow from Upload to Obtaining 3D Data

Table of Contents

• Challenges of Drone Point Cloud Generation

• What Is a Cloud-Based Point Cloud Processing Service?

• Workflow from Upload to 3D Data Acquisition

• Benefits of Using the Cloud

• Deliverables and Features That Can Be Generated

• Use Cases in Construction and Surveying

• Simple Surveying Using Smartphone-Based LRTK

• FAQ

Challenges in Generating Drone Point Clouds

Recently, the construction industry has been focusing on drone-based three-dimensional surveying as part of digital transformation (DX). Backed by the Ministry of Land, Infrastructure, Transport and Tourism’s i-Construction initiative, the use of 3D survey data has been increasingly promoted year by year. By photographing a site from the air with a drone and generating point cloud data from those images, terrain and structures can be measured accurately to assist in as-built management and earthwork volume calculation. However, traditionally, generating high-precision three-dimensional point cloud data from those photographs required high-performance PCs and expensive specialized software. Point cloud data refers to a 3D model that represents the surface shape of an object as a collection of innumerable survey points (coordinate points). The initial implementation cost was a significant burden, and even when drones were introduced, it was not uncommon for users to merely view the captured photos without taking the step to create 3D models.

Generating a high-density point cloud from over 100 high-resolution images requires lengthy processing that ties up the computer. Operating photogrammetry (SfM analysis) software and configuring its parameters also demands specialized knowledge, making it difficult to handle without experienced technicians in-house. Because of these high hurdles, there were many cases where companies gave up on creating detailed terrain models or performing earthwork volume calculations in-house with drones and were forced to outsource the work to external surveying firms.

When you outsource, it can take several days to complete data analysis, and costs tend to be high. On the other hand, if your company can leverage cloud services, you can complete data processing on the day of the drone flight and check the results that same day. In response to this need, what has emerged in recent years is cloud-based point cloud generation services.

What is a cloud-based point cloud processing service

A cloud-based, end-to-end point cloud processing service automatically converts the aerial photo data you captured on-site into 3D point cloud data simply by uploading it to a dedicated platform in the cloud. By using this service, you can easily generate point cloud data without having to purchase a dedicated high-spec PC or software.

Using the service is simple. First, access the service's website from a web browser and upload the large number of photos captured by a drone. Then, an analysis process of photogrammetry(Structure from Motion, etc.) is automatically executed on high-performance servers in the cloud. This process matches common points appearing in multiple photos and computes the 3D coordinates of each point using the principles of triangulation. Even with hundreds of photos, processing proceeds efficiently by parallel processing on the server side, and in some cases results can be obtained in less time than when processing on a conventional PC. When the analysis is complete, output data such as 3D point clouds and orthomosaic images are produced in the cloud. Users can view and download them from their own PC via a browser. Because users do not need to leave heavy computations to their local PCs, they can proceed with other work during the waiting time and avoid a loss of productivity.

Furthermore, because optimized workflows are provided in the cloud, users don't need to perform complex configurations. As long as basic quality—such as the number of photos and the degree of overlap between shots—is ensured, anyone can obtain high-quality point cloud data without specialized knowledge. Leveraging the cloud is making advanced 3D data processing from drone surveying markedly more accessible.

Flow from Upload to Obtaining 3D Data

With a cloud-based service, you can obtain 3D point clouds from photos in just a few steps. The general process is shown below.

• Imaging (Data Acquisition): First, acquire photographic data of the target area using a drone or similar. Photograph the survey site sufficiently from various angles from above and prepare high-resolution image data. Record information for reference ground control points (GCP) as needed.

• Data Upload: Upload the captured photos to the cloud service platform. Access the dedicated site via the internet and specify multiple image files following the guide. Once the upload is complete, automatic analysis starts in the cloud.

• Automatic Analysis (Point Cloud Generation): Photogrammetry algorithms run automatically on the cloud server for analysis. Because parallel computations are performed on a high-performance server environment, large numbers of photos are processed efficiently. Advanced computations such as matching common points and calculating 3D coordinates are all automatic. Processing status can be checked on the web, and you will be notified upon completion.

• Generation of Deliverables: As a result of the analysis, deliverables such as 3D point cloud data and orthophotos are generated in the cloud. Point cloud files in standard formats such as LAS and high-resolution orthomosaic images are output, and, if necessary, DSM (digital surface model) and contour data are also created.

• Data Review and Download: Users can view the deliverables in the cloud from a browser. You can freely change viewpoints to inspect the point cloud in a 3D viewer and perform operations such as measuring distances and areas. Download the necessary data to save locally and use it with your company’s CAD software or BIM tools.

As described above, you can obtain high-precision 3D point clouds from on-site photos simply by uploading them to the cloud. No special equipment or complicated procedures are required; the cloud point-cloud generation service is designed to be easy for anyone to use.

Benefits of Using the Cloud

Generating point cloud data in the cloud offers various advantages compared to conventional in-house processing. The main benefits are summarized as follows.

• Reduction of initial costs: Because purchasing high-performance PCs or expensive analysis software becomes unnecessary, capital investment can be significantly reduced. Traditionally, equipment investments of several million yen were required, but removing that burden makes it easier for small and medium-sized companies to adopt the solution.

• Reduced load on local PCs: Because the processing of vast photo datasets is performed on the cloud side, you don’t need to worry about your PC freezing for long periods or its fans running loudly. You can continue other tasks while waiting for processing, avoiding a drop in work efficiency.

• High-speed processing and scalability: Powerful server resources and parallel computing allow heavy analyses to be completed in a short time. Even when the number of photos or the area scale is large, the cloud can handle it. For example, analyses that take more than half a day on a typical PC can, in some cases, finish in a few hours on the cloud. The cloud also offers the flexibility to process multiple sites’ data concurrently within a single company.

• Accessible from anywhere: If you can connect to the internet, you can use the service from the office, a site, or while on the go. For example, you can upload photo data to the cloud immediately after shooting at a site, use travel time to advance the analysis, and have results by the time you return to the office.

• Easy data sharing: Generated point clouds and ortho images are stored in the cloud, so you can immediately share data with stakeholders via a URL. The latest 3D models can be viewed through a browser from remote locations. Because large files are accumulated in the cloud, there is no need to worry about insufficient disk space on your PC. Also, without installing a dedicated viewer on each user’s PC, 3D data can be displayed directly in the browser.

• Always benefit from the latest technology: Since the service provider continuously updates analysis algorithms and features, users can always use the latest version of the technology. With each update, accuracy improvements and new features are automatically applied, so users do not need to perform cumbersome software updates themselves.

• Comprehensive support: If it is a Japan-based cloud service, you can have the reassurance of thorough support and implementation assistance in Japanese. You can quickly consult about operational questions or troubleshooting, allowing even beginners to start implementation with confidence.

Deliverables and Capabilities That Can Be Generated

A cloud-based point cloud generation service can automatically generate a variety of deliverables useful for surveying and design, not just provide point cloud data. Representative examples of outputs are as follows.

• 3D point cloud data: Generates high-density, colored point clouds from drone photos. By utilizing high-precision positioning information from RTK-enabled drones and ground control points, you can obtain high-precision point clouds in which each point is assigned accurate absolute coordinates. The generated point cloud data can be exported in standard formats such as LAS, allowing import into other CAD software and analysis tools for further use.

• Orthomosaic images: Aerial photos are stitched to create distortion-free ortho images viewed from directly above. These map-like images allow an overview of large areas and are suitable for understanding the overall picture or for use as survey drawings. Because the images include position coordinate information, they can be easily overlaid on drawings for use.

• Digital Surface Model (DSM) and contour lines: Elevation data of the ground surface are extracted from the generated point cloud to automatically create gridded DSMs and contour maps at specified intervals. These can be used directly as basic maps for understanding site topography and are useful for civil design and calculation of earthwork volumes during disasters.

• Various measurement functions: In a cloud-based point cloud viewer, you can measure the distance between any two points, calculate the area of an enclosed region, and compute volumes, elevation differences, and slopes. For example, volumes of embankments or excavations can be automatically calculated simply by setting a reference plane on the point cloud. Because required dimensions can be read immediately on the web without special CAD software, report creation and verification work are streamlined.

• Creation of cross-sections (DXF output): There is also a function to extract cross-sectional shapes along arbitrary section lines from point cloud data and automatically convert them into CAD drawings. The generated cross-sections can be downloaded in DXF format and used as-is for comparison with design drawings or for preparation of as-built reports.

• Overlay with design and BIM data: Design models and drawing data (BIM/CAD) can be overlaid on the point cloud to visualize differences between the as-built condition after construction and the design plan. Color difference maps, recommended in the Ministry of Land, Infrastructure, Transport and Tourism’s as-built management guidelines (draft), can be output in the cloud, helping early detection of construction defects. Advanced inspection analyses, such as comparing as-built 3D data with BIM models and color-coding deviations between the final design and current conditions, can also be performed easily, strongly supporting site management and as-built inspections.

Use cases in construction and surveying

Cloud-based point cloud generation services are expected to see broad adoption across a wide range of fields, particularly in the construction and civil engineering industries. They deliver significant benefits in the following scenarios.

• Construction & Civil Engineering: Drone surveying can be used for as-built management and earthwork volume calculations at construction sites. For example, calculating the volumes of fill and excavation from point clouds captured by drones during land development work dramatically streamlines progress management and as-built reporting. As-built measurements that traditionally took several days by manual surveying can be completed from capture to cloud processing within the same day using drones, enabling rapid decision-making. This 3D surveying workflow aligns with the i-Construction initiative promoted by the Ministry of Land, Infrastructure, Transport and Tourism and contributes to the advancement of site management.

• Surveying Firms & Construction Consultants: Because 3D surveying deliverables can be produced quickly and inexpensively, this leads to improved operational efficiency and service competitiveness. Even without owning expensive workstations or software in-house, data processing for large-scale surveying projects can be entrusted to the cloud, enabling short delivery times and cost reductions. In addition, automated processing stabilizes data quality, allowing less-experienced technicians to provide high-quality results.

• Municipalities & Public Agencies: In infrastructure inspections and disaster response, current-condition data acquired by drones can be rapidly converted into 3D. By simply uploading images captured on site to the cloud, detailed terrain models and orthophotos can be available within a few hours, quickly aiding damage assessment and facility maintenance. For example, even in a large-scale landslide, drones can record conditions from above and the volume of collapsed soil can be calculated that same day and reflected in recovery plans, enabling rapid response. Even municipalities without specialized departments will increasingly be able to respond in-house without outsourcing.

• Other Fields: The range of applications for drone × cloud point clouds is expanding: in agriculture and forestry for wide-area measurement of field and forest topography, and in education and research for easily creating 3D models for practical training and experiments. They can also be applied in many fields depending on ideas, such as recording hazardous areas that are difficult for people to enter and creating 3D archives of cultural heritage.

Simple Surveying with LRTK Using Smartphones

In addition to cloud-based drone point cloud generation, in recent years attention has also focused on simple surveying using smartphones as an easy way to perform on-site 3D surveying. Our company offers LRTK Phone, an innovative system that enables 3D surveying with cm level accuracy (half-inch accuracy) using only a smartphone. By using a smartphone equipped with a dedicated compact antenna and combining high-precision GNSS positioning with the built-in LiDAR scanner and camera, a single person can easily acquire high-precision point cloud data. Simply launch the dedicated app and walk around the site holding the smartphone, and real-time 3D point cloud generation is completed on the spot.

For example, areas that are difficult to capture with drones—such as the undersides of bridges or the ground surface within forests—can be scanned in detail simply by walking while holding an LRTK Phone. Because the ground point cloud data obtained in this way have absolute coordinates, they can be automatically overlaid with point cloud models generated by drones. By using a smartphone to supplement blind spots that are hard to capture from above, it becomes possible to obtain a more complete understanding of the current conditions.

For routine, simple measurements, there’s no need to fly a drone every time. Since you can quickly capture on-site 3D data with just a smartphone when needed, this will greatly improve productivity at sites with frequent surveying. By using a cloud-based drone point cloud service for wide-area mapping and the LRTK Phone for detailed measurements, you can establish a system that enables seamless 3D surveying across every corner of large sites.

Thus, with the emergence of cloud-based end-to-end services and smartphone surveying technologies, 3D measurement using drones is becoming even more accessible. Why not introduce a new on-site workflow that isn’t tied to high-spec PCs and leverage field data with unprecedented speed and efficiency? Our LRTK platform strongly supports the digital transformation (DX) of your surveying operations through these cutting-edge technologies. Cloud analysis technologies are advancing rapidly, and we plan to further expand the features of our LRTK platform going forward. By continually adopting the latest solutions, we will continue to support the improvement of your operational efficiency.

FAQ

Q: What kind of users is this service suitable for? A: It is suitable not only for construction site engineers and civil surveying personnel, but also for a wide range of people who need to utilize 3D data from drone surveys, such as construction consultants handling spatial data and municipal staff responsible for infrastructure management. Because it can be easily adopted by small businesses without dedicated departments, it is recommended for any site that wants to make effective use of drones. It is also useful for individual drone operators as an affordable means to achieve advanced 3D surveying.

Q: What deliverables can be obtained from this service? A: In addition to the basic 3-dimensional point cloud data (color point cloud), a variety of deliverables can be automatically generated, such as high-resolution orthophotos (vertical nadir images), digital surface models (DSM) of the ground surface, contour maps, and arbitrary cross-sectional drawings (DXF format). Since outputs required for surveying and design are provided, the acquired data can be directly used for a wide range of business applications. For details, please refer to the "Deliverables and Features" section in the body of this article.

Q: How long does the analysis take? A: Processing time varies depending on the number of photos and the condition of the cloud servers, but as an example, for around 100 photos point cloud generation is often completed in about an hour. Even for large projects on the order of 500 photos, processing finishes in a few hours to about half a day. As a guideline, results are usually available within the same day, and in most cases you can check results faster than when processing on a standalone conventional PC.

Q: Do you need any special equipment or software to use it? A: All you need is a standard personal computer (or tablet device), an Internet connection, and a drone for capturing images. Because the service is cloud-based and accessible via a web browser, there is no need to install any dedicated software. It can be used without issue on any OS that runs a browser, whether Windows or Mac, and all analysis processing is executed on the server side, so a high-performance PC is not necessary.

Q: Which types of drone images are supported? A: Photos taken with commercially available aerial drones can, in principle, be processed regardless of the model. However, camera resolution and shooting method are important for generating high-quality point cloud models. Take photos at the highest possible resolution and with good sharpness, and plan flights so that images have sufficient overlap to achieve accurate results. Also, using an RTK-equipped drone improves the accuracy of positioning information, which further increases the surveying accuracy of the generated point cloud data.

Q: What is the accuracy of the generated point cloud? A: The accuracy varies depending on the drone's GPS accuracy, flight altitude, and whether ground control points (GCP) are used. Even with a standard GPS-equipped drone alone, under favorable conditions you can obtain relative accuracy on the order of several centimeters to a dozen or so centimeters (a few in to about 4–8 in), which is sufficient for many practical applications. Furthermore, using an RTK-capable drone or combining with known ground points can reduce errors to a few centimeters (a few in). With proper accuracy control, it is possible to produce deliverables that meet the accuracy requirements of the Ministry of Land, Infrastructure, Transport and Tourism's 3D Surveying Guidelines.

Q: Is advanced specialized knowledge required to operate? A: No, no special skills are necessary. As long as you know the procedure for uploading photo data, all advanced analysis processing is automatically performed on the cloud, so you do not need expertise in image analysis. The service interface and manuals are also provided in easy-to-understand Japanese, allowing even first-time users to operate intuitively by following the guide. If you have any questions, our support team will assist you, so you can use the service with confidence.

Q: I'm uneasy about storing a large volume of photo data in the cloud. Is security okay? A: Our service implements robust security measures for data in transit and for data stored on our servers. Uploaded photos and generated point cloud data are encrypted and kept in secure cloud storage. You can also set access permissions for each user so that only authorized members can access the data. The cloud servers themselves are operated in highly reliable domestic data centers, so you can use the service with confidence even for highly confidential projects.