Complete Guide to Applying SfM Processing: Thorough Explanation from Industry Case Studies and Benefits Comparison to Tool Selection and Implementation

What is SfM processing? Basic concept and why it attracts attention

SfM (Structure from Motion) processing is a computer vision technique that reconstructs the three-dimensional structure (3D models) of objects or sites from multiple photographs. Also known as a form of photogrammetry, it automatically analyzes images taken from various angles by drones or cameras to generate point cloud data, 3D mesh models, orthophotos, and more. Traditionally, 3D measurement required expensive laser scanners (LiDAR), but the spread of SfM has made it possible to obtain high-density three-dimensional data easily with ordinary digital cameras or drones.

The reason SfM processing has attracted attention recently is its convenience and cost advantages. Without relying on specialized equipment or expert technicians, site personnel can quickly capture wide areas and automatically create 3D models, dramatically improving work efficiency and productivity. The acquired data can be stored as digital point clouds and models, enabling intuitive site understanding and advanced process management that were difficult with conventional 2D drawings or photographs. Governments are also promoting ICT use to boost productivity in the construction industry (for example, the promotion of *i-Construction*), and 3D surveying using drone aerial photography and SfM is becoming an industry standard. SfM processing, which provides low-cost and high-precision 3D data, is expected to be a key technology supporting today’s digital transformation.

Industry use cases

SfM processing is already being put into practical use across a wide range of fields. Below are representative use cases by industry.

• Construction sites: In civil engineering and construction, aerial SfM surveying with drones enables rapid acquisition of large-area terrain data, which is used for development planning and as-built management. Tasks that used to take several days—such as land surveying and earthwork volume calculations—have been greatly streamlined by drone imaging and automated processing. Comparing generated 3D models with design data allows early detection of construction errors and visualization of progress. It is also advantageous that dangerous slopes and embankments can be surveyed safely without personnel entering them.

• Municipal infrastructure inspections: Local governments use SfM-generated 3D models for maintenance and management of infrastructure such as bridges, tunnels, roads, and dams. High-resolution models captured by drones help identify cracks and displacements and quantitatively record aging progress. Tasks previously performed visually by inspectors are digitized, enabling repair plans based on objective data. Especially for inspections at height or across large facilities, SfM can reduce the cost and risk of erecting scaffolding or using elevated work platforms, contributing to greater efficiency and safety in municipal operations.

• Disaster prevention: SfM is a powerful tool in disaster response. Aerial photos of landslide or flood sites can be used to instantly create 3D terrain models of affected areas, aiding damage assessment and secondary disaster risk evaluation. For example, the volume of collapsed slopes can be measured from models to quickly estimate the amount of debris to be removed. Even during normal times, terrain models are used to create hazard maps for debris flows and to periodically monitor deformation of levees and slopes—SfM data thus contributes to preparedness and prevention. It supports disaster response significantly by enabling safe acquisition of detailed data from the air, even in areas unsafe for people to approach.

• Cultural heritage preservation: In the field of historic buildings and cultural assets, SfM-based 3D documentation is becoming an essential method for preservation and restoration. Ruins, temples, and shrines can be non-contact and precisely scanned and preserved as digital archives for future restoration work and research. Damage that is difficult to see with the naked eye can be enlarged and inspected on the model, providing a complete record of the current state. Because precise dimensional information can be obtained without moving or touching the real object, valuable cultural assets can be protected while enabling high-precision drawing and restoration planning. Generated 3D models can also be used for VR content and online exhibitions, contributing to digital dissemination and tourism promotion of cultural assets.

• Agriculture: In agriculture, SfM processing of drone images is also being applied. 3D maps generated from aerial images of fields can be used to identify low spots with poor drainage and measure field slope, aiding drainage countermeasures and land leveling. In orchards, modeling tree height and canopy volume helps monitor growth status and predict yields. Visualizing the entire farmland enables early detection of uneven growth or disease signs that were previously overlooked, leading to precision (smart) agriculture practices. Covering surveys that used to rely on manual labor with aerial imaging contributes to labor savings in farm work, and usage in this field is expected to expand further.

Comparison of SfM implementation benefits (cost, accuracy, operational burden, compatibility with digitalization)

Here we organize the benefits of introducing SfM technology from the perspective of comparing it with conventional methods.

• Cost: SfM can significantly reduce initial investment and operating costs compared to traditional surveying and measurement methods. Dedicated 3D measurement equipment like laser scanners used to cost millions of yen, but with SfM you can get started with commercial drones or cameras and a PC. Fewer personnel are required for imaging, and a single operator can cover large areas, contributing to reduced labor costs. Software options range from open-source to cloud services, allowing flexible adoption according to the scale of use. For these reasons, low-cost 3D adoption is feasible even for small and medium-sized projects, which is a major attraction.

• Accuracy: While there have been concerns about the accuracy of photogrammetry, current SfM technology can achieve accuracies on the order of a few centimeters if proper procedures are followed. In particular, using drones equipped with RTK-GNSS or placing known coordinate control points on the ground can provide absolute accuracy (coordinate accuracy) for models. This allows the resulting point cloud data to be overlaid directly on public coordinate system drawings or GIS. On the other hand, SfM has weaknesses compared to LiDAR in areas that are hard to capture with photos—such as ground under dense tree cover—where data gaps may occur. However, many cases can be compensated by careful flight planning and additional measurements, making SfM sufficiently accurate for typical civil surveying and design uses.

• Operational burden: The operational hurdles for implementing SfM have decreased compared to the past. Field imaging is faster and simpler than detailed traditional ground surveys. Drone pre-flight settings and flight plans are necessary, but once airborne, photos can be acquired automatically, reducing physical burden and danger for operators. Post-processing used to require high-performance PCs and long computation times due to massive image processing, but nowadays cloud services allow processing without relying on the field PC’s specs. As software automation advances and workflows become more refined, even non-specialist technicians can reach operational proficiency in a relatively short training period. Overall, SfM has matured to a point where it can be integrated into daily operations without undue strain, and operational burdens are steadily decreasing.

• Compatibility with digitalization: The outputs of SfM are digital data, making integration with other ICT tools straightforward. For example, point clouds and 3D models can be imported into CAD or BIM/CIM software for cross-checking with design data, enabling intuitive interference checks and as-built verification that might be missed on 2D drawings. Orthophotos and point clouds can also be overlaid on GIS maps and managed as infrastructure asset information. Shared via the cloud, 3D current-state data can be viewed by remote stakeholders, improving information sharing and decision-making speed. As organizations shift from paper-centric to digital-centric workflows, adopting SfM processing can strongly support DX (digital transformation).

Tool and software selection points (beginner-friendly, task-specific, cloud-based, etc.)

Many software products and services implement SfM processing. Below are points to consider when selecting a tool that fits your organization’s use cases and skill level.

• Beginner-friendly: If you are new to SfM, choose a tool with as user-friendly an interface as possible. Software with intuitive interfaces and comprehensive Japanese support or manuals will make learning smoother. Some tools perform automatic processing simply by dragging and dropping photos, allowing use without specialized knowledge. Try free or trial versions to test the user experience and confirm whether field staff can handle the workflow.

• Task-specific: There are industry-specific solutions for SfM processing. For example, civil surveying tools often provide robust earthwork volume calculation and terrain mapping functions, while cultural heritage applications emphasize high-resolution textures and fine-detail modeling. For agriculture, services with GIS integration and growth-analysis features are available. Compare tools based on whether they offer functions tailored to your intended use cases and which tools are widely adopted in your industry to avoid mismatches.

• Cloud-based: Recently, cloud-based services that complete processing online have become more common compared to traditional PC-installed software. The advantages of cloud solutions include fast processing of large numbers of photos regardless of local PC specs, and no installation or update overhead. Output data is easy to share and view via the web, allowing unified platform access across dispersed field offices. However, uploading many photos requires adequate network connectivity, and subscription or running costs must be considered. Evaluate whether cloud use is appropriate based on project frequency and data confidentiality.

• Other considerations: Beyond the above, implementation cost and budget are important. Open-source tools reduce initial expenses but may require self-supported troubleshooting. Paid software offers comprehensive support and features but must be balanced against license costs. Also check whether output data formats (e.g., LAS point clouds, OBJ models, GeoTIFF orthophotos) are compatible with your operational workflow. Consider your organization’s IT literacy and operational structure and select a tool from the perspective of whether it can be sustainably used long-term—this is key to successful adoption.

Precautions when introducing SfM and a checklist for field use (photo shooting, ensuring accuracy, processing load, etc.)

When introducing SfM processing on site, there are points to confirm and on-site precautions to observe. The following is presented as a checklist.

• Photo shooting basics: The quality of 3D reconstruction depends heavily on the source photos. Ensure sufficient overlap (60–80% or more) and photograph the subject from various angles. Blurry or out-of-focus photos will reduce accuracy, so set the camera for sharp images (appropriate shutter speed and ISO). For outdoor shooting, watch lighting conditions—avoid extreme backlighting and aim for uniform illumination to stabilize post-processing.

• Ensuring accuracy: Depending on the required accuracy, additional measures in shooting may be necessary. To increase absolute accuracy, place known coordinate points (control points) on site and include them in photos. These points can be used later to scale and georeference the model, producing more accurate dimensions and coordinates. If the drone is equipped with high-precision GNSS, photo positions along the flight path can be measured, reducing the need for many ground control points. In any case, consider appropriate positioning aids according to the desired accuracy.

• Processing environment and data management: After shooting, you will process a large number of photos, so confirm in advance the PC performance or cloud service availability. Processing hundreds of photos locally typically requires a high-performance GPU and ample memory. If using cloud processing, allow time for large uploads in the schedule. Generated point clouds and model files are often very large, so prepare sufficient storage capacity and backup systems.

• On-site data checks: Before leaving the site, make a habit of checking for missing or defective data. For example, perform a quick review of captured photos on-site to verify there are no blurred images or uncovered areas. If needed, take additional photos on location to avoid later issues such as “holes in the model” or insufficient accuracy in important parts. If possible, perform a partial processing test on a laptop to get an estimate of result quality.

• Compliance with regulations and safety measures: When using drones, flight plans must comply with aviation laws and related regulations. Check for no-fly zones and altitude restrictions, and obtain any required permissions in advance. Site safety around the flight area is also important. In locations where third parties may enter, station observers to prevent accidents. Perform basic checks such as battery level and airframe inspection, and operate with safety first, as this will contribute to smooth SfM utilization.

Conclusion: Integration with simplified surveying by LRTK and implementation effects

SfM processing using drones or cameras brings the various benefits described above, and combining it with other advanced technologies can further enhance its effects. One example is integration with simplified surveying solutions provided by LRTK.

For instance, with a smartphone-integrated surveying device like the "LRTK Phone", anyone can easily acquire centimeter-level positioning points or perform handheld detailed scanning. By overlaying high-precision point cloud data acquired with the LRTK Phone onto point clouds generated by drone SfM, you can create a seamless 3D model that includes areas that are blind spots for drones (such as under tree canopy or behind structures). Combining multiple measurement methods enables comprehensive digitization of all site information.

Implementing such an integrated approach can dramatically boost surveying productivity. Tasks that used to require separate teams and equipment are consolidated, and workflows from fieldwork through data processing and sharing become smoother, directly shortening project lead times and reducing costs. Moreover, the increased reliability of obtained data makes it easier to share the "ground truth" among stakeholders, facilitating optimized construction planning and faster decision-making. In practice, construction companies and municipalities have begun combining drone SfM with LRTK surveys, achieving significant results in rapid disaster response investigations and infrastructure management.

Finally, digital surveying centered on SfM processing is likely to become an increasingly standard tool in everyday operations. By adopting these technologies in a manner that fits your company’s needs, you can steadily advance on-site DX and realize operational reforms that combine safety, efficiency, and accuracy. We hope this guide encourages you to consider applying SfM processing and integrating simplified surveying with LRTK in your organization.