How to streamline deformation surveys with point clouds? Six checks to avoid failure

In the maintenance and management of road structures, slopes, areas around bridges, retaining walls, tunnels, reclaimed land, and building exteriors, it is crucial how quickly, accurately, and reproducibly defects such as cracks, bulging, settlement, displacement, material loss, and spalling can be identified. Conventional condition surveys have focused on on-site visual inspections, photographic records, measurements taken with scales, and recording on hand-drawn sketches, but when conditions such as large inspection areas, highly uneven surfaces, significant elevation differences, and limited access coincide, work time increases and oversights and variability in the records become more likely.

What has been attracting attention is the use of point clouds. Because point clouds record the surface geometry of an object as a collection of many points, a major feature is that they can preserve the site’s shape as a three-dimensional form rather than as surfaces. By incorporating point clouds into deformation surveys, many advantages can be gained, such as reduced on-site time, easier comparative verification, easier sharing, and easier later rechecks. On the other hand, simply acquiring point clouds does not mean the survey will go well. If data are captured at an accuracy that does not meet the purpose, if comparison conditions are not aligned, or if blind spots remain, it can actually make judgment more difficult and cause rework rather than improving efficiency.

Many practitioners who search for "deformation survey point cloud" want to know what becomes easier by using point clouds, where failures are likely to occur, and what they should check when introducing them. In this article, after organizing the key ideas to keep in mind for streamlining deformation surveys with point clouds, we clearly explain six on-site checks that are particularly important from a practical perspective.

Table of Contents

• Why using point clouds for condition surveys leads to increased efficiency

• Confirmation item 1: First define what the survey intends to verify.

• Check item 2: Determine acquisition conditions appropriate for the type of deformation

• Confirmation item 3: Standardize how coordinates and reference points are determined

• Checklist item 4: Ensure the measurement plan leaves no blind spots or gaps.

• Confirmation item 5: Align comparison methods and evaluation criteria in advance

• Confirmation item 6: Implement operations that allow point clouds to be used as on-site records

• Summary

Why deformation surveys using point clouds lead to increased efficiency

Efficiency in condition surveys is not simply about finishing measurements more quickly. Efficiency should be considered to include shortening on-site work while recording data with the required level of accuracy, making later re-verification easy, and creating a situation in which stakeholders can make decisions while viewing the same information. From this perspective, point clouds are a very well-suited method.

First and foremost, a major advantage is that shapes can be preserved both as surfaces and in three dimensions. With photographs alone, appearance easily changes depending on the shooting position and field of view, making it difficult to quantitatively trace afterward where and to what extent displacements or level differences occurred. By contrast, point clouds retain the surface positions of the object in three dimensions, allowing you to check not only a single point but also its relationship with the surrounding geometry. Another practical strength is that they make it easier to get an overview of not only local changes but also wide-area deflection, tilting, protrusion, and tendencies toward settlement.

Also, making comparisons on revisits easy directly improves efficiency. In condition surveys, there are many situations where confirming the difference from the previous inspection is more valuable than a single record. If the previous condition and the current condition can be managed in the same coordinate system, it becomes easier to visualize displacements and shape differences and to determine whether progression is occurring. This makes it easier to provide objective explanations without relying solely on personnel with extensive field experience.

It's also worth noting that you don't need to stay on site for long. In locations that require traffic control, sites near high elevations or slope shoulders, or places where maintaining separation from third parties is difficult, being able to acquire the necessary information in a short time is itself highly valuable. If you capture a point cloud thoroughly once, you can perform cross-section and distance checks back at the office, making it easier to reduce the number of re-measurements required on site.

However, these advantages only come into play when the workflow—from acquisition through comparison to management—is properly established. It is not simply a matter of capturing more detailed geometry or having a larger number of points. What is important in a deformation survey is designing the conditions, aligned with the survey's objectives, so that the necessary changes can be reliably detected. For that reason, the following six items to check are important.

Confirmation Item 1: Define in advance what you want to confirm in the survey

What you should confirm first in a deformation survey using point clouds is the survey objective itself—what you want to understand. If you enter the site with this unclear, the capture density, measurement range, required accuracy, comparison method, and the format of deliverables tend to become half-baked, and you may later find that "the information you wanted was not captured."

There are various types of condition surveys. For example, the requirements for point cloud data vary depending on whether the primary objective is to determine crack locations, to measure the amount of surface bulging, to monitor settlement or differential displacements over time, or to assess the volume of missing sections and changes in cross-sectional area. Whether you want to precisely track the minute width of cracks using point clouds alone, or to identify the areas where cracks are concentrated and detect signs of shape change, will also require different acquisition methods.

What is important here is to separate and consider four stages: confirming the presence of deformations, grasping the extent of the deformations, quantifying the amount of deformation, and comparing changes over time. If the goal is presence confirmation, a plan to cover a wide area without omissions is important; if the goal is quantification, coordinate management and accuracy control become more important. If comparisons over time are the main objective, it is not enough to get good results this time only—you must also consider being able to reproduce the same conditions in future surveys.

For example, if you want to check for bulging in a retaining wall but collect data intermittently from positions where only part of the wall face is visible, it becomes difficult to discern the overall trend of the surface. Conversely, if you want to observe the deformation trend of an entire slope but only take high-density measurements in local areas, the relationship to the whole becomes unclear. Clarifying the survey objectives means not simply "confirming deformations," but verbalizing "which deformations," "over what area," "to what accuracy," and "how you will judge them."

In practice, failures are reduced if you prepare so that the investigation objective can be written in one sentence before going to the site. For example, if you make it specific to the level of "record the outward bulging of the front face of the retaining wall and the horizontal displacement of the crest in a form comparable to the previous survey" or "verify surface collapse traces on the slope and map the areal extent of erosion progression," it becomes easier to decide the necessary coverage and how to set standards. Point clouds are a versatile recording method, but if the objective is vague, only the data volume increases and judgment actually becomes harder. The first step toward efficiency is to define the investigation objective concretely.

Confirmation Item 2: Decide acquisition conditions that match the type of deformation

The next important step is to determine the acquisition conditions based on the anomalies of the object under investigation. The quality of the point cloud is affected by many factors, including the distance to the target, acquisition angle, point density, overlap, lighting and weather, surface material, and surrounding obstructions. If this is left to on-site judgment, you are likely to encounter problems such as only the areas you later want to inspect being captured at low resolution, excessive reflections and noise, loss of fine detail, and indistinct contours around missing sections.

In condition surveys, the changes you want to observe vary depending on the type of object. For concrete structures, surface delamination, spalling, loss of material, steps or level differences, and out-of-plane displacement can be important. For earthworks and slopes, erosion, collapse, scour, and the accumulation or movement of sediment are the main items to check. For stone masonry and blockwork, local bulging or overhangs, opening of joints, and displacement of individual units are important. Because whether you need to focus on fine details or the overall shape of a surface differs for each target, the required point density also changes.

A common misconception here is the idea that taking more detailed measurements will solve everything. Indeed, higher point density makes fine details easier to see, but on sites with a wide acquisition area the data volume becomes large, and processing and sharing take more time. Also, even with high density, a poor acquisition angle can prevent you from correctly capturing the shape of a deformation. For example, to correctly observe bulges or chips on a wall surface, it is important to capture from as appropriate an angle as possible to the target surface and, where necessary, overlay captures from multiple directions. Even if you capture at high density when only part of a surface is visible, the result may not become a practical point cloud.

Furthermore, if a comparative survey is intended, it is insufficient for conditions to be good only this time. To make it easy to acquire data under the same conditions next time, you need to record from which position, at what height, and which direction was the focus when capturing. Without that, the areas visible in the previous and current data will differ, making comparison difficult. To track the progression of deformation, it is important not merely to create point clouds but to continue recording the same subject using the same methodology.

On site, deciding beforehand which surface of the object will be the primary target, how much of the surrounding area to include, and where to set the reference plane or stable parts used for comparison will make acquisition conditions less likely to vary. Efficiency gains from point clouds are not determined solely by the performance of the equipment. Acquiring data under conditions appropriate for the type of deformation and capturing only the information that is necessary—no more and no less—is ultimately the most efficient approach.

Checklist Item 3: Standardize How Coordinates and Reference Points Are Taken

In deformation surveys, the way coordinates and reference points are defined is critically important for making effective use of point clouds. In particular, if you plan to compare data over time, cross-check it with other drawings or records, or share it among stakeholders, acquiring data using inconsistent references each time will make it difficult to use in practice. Even when you believe you measured the same location on site, a change in reference can make it hard to determine whether a difference is actual displacement or simply a misalignment in positioning.

One reason point cloud comparisons fail is aligning based on the object itself. For example, if you align using the entire surface of a wall or slope that may have displaced, the true deformation can be absorbed, making the differences appear smaller. Conversely, using an unstable portion as the reference can make the intact areas appear shifted. In either case, you cannot correctly determine whether deformation is present.

Therefore, in deformation surveys, it is important to decide in advance where to place reference points that will move as little as possible. Securing elements that can be used as references for comparison—such as nearby stable structures, known points, or fixed points that are easy to reproduce—will make subsequent processes more stable. What matters is not taking a reference per se, but ensuring that references can be set using the same approach each time. You should record the rationale for coordinates and the positions of fixed references so that the same judgments can be made even if personnel change.

Also, even for a localized survey, leaving records that show positional relationships with the surroundings will broaden its usefulness later. On site, it’s easy to default to “let’s just capture this surface for now,” but when evaluating deformations it is often important to consider continuity with surrounding areas and the relationship to adjacent parts. For example, what you thought was a localized settlement may actually be part of a much more extensive deformation. Rather than limiting the work to the local area alone, including at least minimal surrounding reference points will ultimately reduce the effort required for re-surveying.

Furthermore, it is important to establish workflows that link point clouds not only with other point clouds but also with photographs, sketches, existing drawings, and inspection records. Condition surveys are not something that can be completed with point clouds alone. They often only become meaningful when cross-checked against on-site observations and historical records. Therefore, you need to align location information and naming management so that which record corresponds to which location can be traced later. Unifying coordinates and reference systems is the foundation not only for comparison but also for enabling point clouds to function as operational records.

Checklist Item 4: Ensure the measurement plan leaves no blind spots or gaps

A common failure in deformation surveys using point clouds is that necessary areas were not visible. Even if it appears that data were captured on site, when checked in the office there may be portions missing where shadows occurred, the backsides of nearby objects may not have been captured, or important surfaces may be missing because they were obscured by vegetation or temporary structures. Because deformations often appear in locations that are hard to see, addressing blind spots is central to improving efficiency.

Particular attention should be paid to the corners of wall surfaces, areas close to backfill, the undersides of projections, around equipment, behind handrails and piping, stepped sections of slopes, gutters and edges. These locations are difficult to capture adequately when viewed from a single direction, tend to concentrate deterioration, and are also prone to being omitted from records. Point clouds are often assumed to be reliable because they can record three-dimensional geometry, but in reality surfaces that are not visible are not recorded. To reduce blind spots, it is necessary to plan on the assumption of acquiring overlapping data from multiple directions.

What’s important here is not to capture the entire surface at the same uniform density, but to concentrate on areas prone to distress and linear features that are important for comparison. For example, you should ensure views that make later checks easy for places where displacement tends to occur—such as the top edge, joints, areas around openings, connection points, and locations affected by drainage. Trying to shorten on-site acquisition time by taking a single frontal shot and calling it done actually increases the likelihood of a revisit. Efficiency is not about reducing the number of shots, but about reducing the need for follow-up investigations.

Noise from the surrounding environment can be just as problematic as occlusions. Vegetation swaying in the wind, reflections on water surfaces, unstable representation of wet surfaces, and the inclusion of passing vehicles or people can degrade the quality of the point cloud and lead to misinterpretation of comparison results. You must avoid situations where, although you intend to examine deformation differences, you are actually comparing environmental noise. It is important to assess site conditions and adjust acquisition timing and vantage point to choose conditions that make it less likely for unwanted noise to be captured.

Furthermore, on-site it is important to have the mindset of "confirming on the spot." Perform a quick check after acquisition to verify whether the main target is sufficiently captured, there are no missing parts, and the surrounding areas needed for comparison have been secured — doing so can greatly reduce rework after returning. Rather than thinking of acquisition and verification separately, an operation that completes the primary check on-site increases the efficiency of point cloud surveys.

Checkpoint 5: Align comparison methods and evaluation criteria in advance

The value of utilizing point cloud data in deformation surveys lies not only in preserving geometry, but also in making it easier to compare differences from previous surveys or from the design geometry. However, if the comparison methods and evaluation criteria are ambiguous, the same data can be interpreted differently by each person in charge. This undermines the reliability of the records as objective documentation.

For example, if you have not decided which part to treat as the fixed reference for comparison, what range to target when examining differences, whether to regard local protrusions as deformations or as noise, or at what degree of change to flag as noteworthy, evaluations will fluctuate each time. In particular, although difference representations of point clouds are visually easy to understand, impressions change depending on how the reference surface is taken and on processing conditions, so operating based on appearance alone is risky.

In practice, it is necessary to standardize the way evaluations are viewed for each subject. The appropriate comparison method varies depending on whether you want to look at surface displacement, linear changes, primarily use cross-section comparisons, or primarily use areal differences. Continuous cross-section comparisons can be effective for bulging of a wall face, while areal comparisons over a wide area can be effective for slope erosion or collapse. Deciding in advance which method will serve as the primary criterion makes explanations consistent when reporting.

Also, in deterioration surveys it is important not to rely too heavily on point clouds. While point clouds make it easy to quantify geometric changes, they do not directly represent information such as the state of material degradation, fine surface textures, traces of water leakage, or the results of sounding tests. Therefore, it is practical to use point clouds as the basis for understanding shape and displacement, and to combine them with photographic records and on-site observations when making judgments. Corroborating areas where changes are seen in point-cloud comparisons with photographs, or linking them to field notes, improves the accuracy of assessments.

Furthermore, it is important to organize the report with the recipient in mind. Even if the surveyor is familiar with point clouds, the client or other related departments may not be able to interpret three-dimensional data in the same way. Therefore, you need to be able to explain which defects or anomalies changed, where they occurred, and how they evolved, linking this information to cross-sections, plan-view locations, before-and-after comparison images, fixed-point photographs, and so on. Efficiency is not just about making things easier for the survey team; it means streamlining the entire process, including decision-making and explanations. Standardizing comparison methods and evaluation criteria in advance contributes to achieving that.

Checklist item 6 Establish operations to enable point clouds to serve as on-site records

Lastly, an often-overlooked perspective is how to operate the acquired point clouds as site records. If a condition/deformation survey using point clouds ends with a one-time check, ad hoc file management might suffice. However, if you assume maintenance and long-term monitoring, you need to design everything from how to store the data, how to name files, and how to share them, to how to link them with photos and drawings; otherwise, the point clouds you worked hard to capture will end up buried.

A common issue in practice is that acquired file names contain only the date, only the location, or only the responsible person's name, so that when viewed later it is unclear which structure, which face, or under what conditions the data were collected. This makes them difficult to use for comparison during re-inspections and causes problems with handovers. At a minimum, you should ensure that the subject name, location, acquisition date, measurement range, the basis for reference, whether related photos exist, and the correspondence with previous data can be traced.

Also, point clouds are not valuable merely by being captured. It is important that the people who need them can view them when they need to. To achieve this, the field personnel, office staff, report authors, and managers need to agree on which formats to use for sharing, how much to reduce file size, and which information to attach. If files are too large to open, reduced so much that the essential parts are unclear, or related photos are managed separately and cannot be linked, the efficiency gains will be diminished.

Furthermore, deformation surveys become more useful the more closely they are linked to site coordinates. If it is clear which deformation is located where, the information can be more readily used for repair planning, re-inspections, and coordination with nearby construction activities. If location information remains vague, simply finding the same spot on the next site visit can take a lot of time. It is important to treat point clouds not merely as geometric data but as georeferenced site records.

From this perspective, it becomes clear that handling location information is critically important to streamlining condition surveys. If there is a system that can reliably capture defect locations on-site at the same time as acquiring point clouds, post-survey organization becomes much easier. This is especially true on large sites or when inspecting multiple points in succession; operational practices that prevent photos, point clouds, survey notes, and target locations from becoming separated will affect outcomes. If on-site position verification and recording can be integrated, a condition survey stops being a mere measurement task and becomes management information that leads to the next decision.

Summary

To streamline condition assessments using point clouds, simply acquiring three-dimensional data is not sufficient. It is important to first define what you want to check, determine acquisition conditions appropriate to the type of deterioration, unify coordinates and reference frames, plan measurements to avoid blind spots and missing data, align comparison methods and evaluation criteria, and establish workflows that allow continued use as on-site records. If you address these six checkpoints, point clouds will not be "data that was just collected and left as is"; they will become practical information that objectively captures and explains the progression of deterioration and supports subsequent surveys and repair decisions.

Especially in the field, whether you can handle not only recording the shape itself but also where deformations occur together with position information greatly affects subsequent work efficiency. When you need to manage multiple inspection points across a large site or rapidly reacquire the same position during follow-up surveys, integrating position identification and recording is essential. In such situations, leveraging LRTK, an iPhone‑mounted high‑precision GNSS positioning device, streamlines on‑site position checks and simple surveying. To further enhance the value of deformation surveys using point clouds, it is increasingly important on future sites to organize not only shape recording but also the capture and operational use of site coordinates.