How to use point clouds in condition surveys? 5 basics to know before introduction

In the maintenance and management of roads, tunnels, bridges, slopes, retaining walls, building exteriors, and developed land, the importance of condition surveys is increasing year by year. Early detection of abnormalities such as cracks, delamination, settlement, bulging, steps, and deformation, and using that information to inform repair and reinforcement decisions, is indispensable for both safety assurance and extending service life. At the same time, on-site work often faces many challenges—“the target area is large,” “shapes are complex,” “people cannot easily approach,” and “records are hard to review later”—and there are many situations where conventional photographs and visual inspections alone are not sufficient to fully organize the information.

What has attracted attention is the use of point clouds. A point cloud is a method of recording the shape of an object in three dimensions as a collection of many points, and it has the advantage of making it easier to grasp the location, extent, depth of anomalies, and their relationship with the surroundings. However, point clouds are not a万能な手段 that will automatically improve survey quality simply by being introduced. If you do not understand what you are acquiring the data for, how much accuracy is required, and which acquisition method is appropriate for the site conditions, you may end up with a large amount of data that cannot be fully used for decision-making.

This article is aimed at practitioners searching for information with the query "deformation survey point cloud" and provides a clear, practical explanation of how point clouds are used in deformation surveys while organizing the basic things you should know before implementation. Summarized from a field practitioner's perspective and written to be easy to understand, it will be useful both for those considering introducing point clouds for the first time and for those who have already experimented with them and want to review their operations.

Table of Contents

• Why point clouds are required in deformation surveys

• Types of anomalies that can be detected in point clouds

• 5 basics to know before introduction

• Practical Workflow for Leveraging Point Clouds in Deformation Surveys

• Common pitfalls when getting started

• Summary

Why point clouds are required in deformation surveys

A condition survey is not simply a check for the presence or absence of defects; it is the work of objectively recording where, to what extent, and in what form anomalies appear, and of linking that information to future comparisons and decisions on countermeasures. What becomes important here is the reproducibility and shareability of the survey results. It is required that different personnel can have the same understanding, that the findings can be re-verified after time has passed, and that the information can be used for repair design and construction planning.

Traditional condition surveys have focused on visual observation, photography, sketching, dimension measurement, sounding, and close-up confirmation. These methods remain extremely important, but when the inspection area is large, surface irregularities are complex, or close access is difficult due to height or steep slopes, the burden of the survey can suddenly increase. Also, photographs are excellent for recording appearance, but they make it difficult to grasp depth and the overall shape, and impressions can change depending on the shooting position and angle. Later, another person in charge may find it hard to accurately understand the positional relationships and scale of the defects from photos alone.

Point clouds are effective because they can preserve these "positional relationships," "shapes," and "extents" in three dimensions. Because the surface geometry of the object can be recorded as spatial coordinates, slight variations in bumps and depressions that are hard to notice during inspections, and deformation trends that are difficult to see from a distance, can be easily checked in post-processing. For example, bulging of retaining walls, slope deformation, steps in floor slabs and pavements, unevenness of structural surfaces, and changes in surrounding geometry associated with settlement are easier to grasp as surfaces when kept as point clouds.

Moreover, point clouds have the advantage of being well suited for comparison. By acquiring the same location at different times and examining the differences, they help determine whether deterioration has progressed and at what rate. Although age-related changes can also be seen by comparing photographs, simple comparisons become difficult when shooting conditions or positions differ. On the other hand, with point clouds that include coordinates, once spatial alignment is performed, changes are easier to visualize and the continuity of maintenance management can be improved.

Additionally, it contributes to improving on-site safety. In locations where close proximity is dangerous, where traffic restrictions or securing scaffolding are constrained, or where you need to grasp a wide area in a short time, being able to reduce on-site time while bringing back the necessary information is of great value. Of course, not everything can be done remotely, but using it as an initial survey to broadly capture the shape and then performing close-up inspections only on the necessary areas is very rational.

The reason point clouds are sought in deformation surveys is not simply because they are a new technology, but because there is a practical necessity on site to simultaneously enhance the objectivity of records, ease of comparison, safety, shareability, and labor savings. Point clouds are a means that readily meet those demands and are expected to serve as a foundation for improving the quality of surveys.

Types of defects observable in point clouds

Point clouds are not a cure-all, but they are particularly well suited to capturing anomalies that manifest in shape. While not all anomalies addressed in condition surveys can be resolved with point clouds alone, point clouds are an especially effective method for deformations that involve changes in shape. Here, we summarize the relationship between deformations that are of practical interest and point clouds.

The most typical examples are deformations and displacements. For instance, bulging of a retaining wall toward the front, extrusion of a slope, deformation of the tunnel interior, settlement of floors or road surfaces, and deflection or tilting of structures all appear as slight differences in surface geometry. These changes are easier to discern in terms of trends when viewed in cross section or as differences from a reference plane. Local deformations that tend to be overlooked by single measurement values are easier to detect as anomalous areas when examined with surface (areal) data.

Next is the detection of level differences and surface irregularities. On pavements, floor slabs, walkways, stairs, and around joints, even slight level differences can affect accessibility and safety. Features that may be visible in photographs but hard to quantify become easier to capture as height differences in point clouds. This is particularly effective when you want to find localized anomalies within a large area.

Point clouds are also useful for capturing bulges and sags. Concrete surfaces, stonework, exterior surfaces, temporary structures, and panel faces can develop local bulging or deformation. These are difficult to grasp from frontal photographs alone, and their appearance changes depending on the viewing direction. If recorded as three-dimensional shapes, it becomes easier to determine which direction and to what extent they have deformed.

On the other hand, some caution is required when it comes to cracks. Cracks with large openings or cracks involving offsets can sometimes be captured easily with point clouds, but extracting thin linear cracks with high accuracy is strongly affected by acquisition density, surface condition, and lighting conditions. Therefore, if crack inspection is the primary objective, it is realistic to combine point clouds with close-up photographs and visual records rather than relying on point clouds alone. Considering point clouds as a supplementary means of understanding the spatial relationships of cracks and the surrounding deformation makes failure less likely.

Delamination and missing material can also be effectively captured by point clouds, depending on their scale. Deteriorations such as parts of a surface being chipped away, rounded-off corners, or peeled surface layers involve depth and a sense of volume, so three-dimensional records can sometimes convey the condition more clearly than two-dimensional images. They can also be used for before-and-after repair comparisons and for grasping quantities.

Furthermore, it is important to consider the relationship with the surrounding topography and nearby structures. Anomalies may appear to occur in isolation, but in reality they can be related to poor drainage, displacement behind the structure, ground movement, interactions with adjacent structures, and the like. If you capture a wider area with point clouds, it becomes easier to check not only the abnormal locations but also the surrounding conditions. This is a perspective that is difficult to obtain from single-point measurements or close-up photographs alone.

In other words, what point clouds readily reveal are abnormalities that manifest as changes in shape. Conversely, internal voids, the progression of material degradation, and early cracks that cause only minor surface changes need to be investigated in combination with other methods. The first step in deciding whether to adopt point clouds is to avoid overestimating them and to position them as a means of broadly, objectively, and comparably recording shape.

5 Basics You Should Know Before Implementation

When incorporating point clouds into deformation surveys, what you should clarify first is the basic design—namely, for what purpose and to what extent you require them—rather than the equipment or processing methods themselves. If you start while this is vague, you may only increase the workload on site without achieving results. The basics you should know before implementation are organized here from five perspectives.

The first fundamental principle is to clarify the purpose of point cloud acquisition. Whether you want a quick check for the presence of damage, to quantify deformation for repair planning, or to retain baseline data for comparison over time will change both the required accuracy and the acquisition scope. For example, wide-area primary screening and local, detailed assessment of damage demand different point densities and different levels of care in fieldwork. If you proceed with an unclear purpose and simply "3D-ize everything for now," post-processing becomes heavier and the essential information for decision-making becomes harder to discern.

The second fundamental is not to confuse required accuracy with required density. On site, people tend to think "the finer the sampling, the safer," but a greater number of points is not necessarily effective. What is required is a density and positional accuracy sufficient to determine the deformations you want to observe. Something that is sufficient for understanding broad trends may be insufficient for detailed crack evaluation. Conversely, sampling a large target at excessively high density increases processing time and data volume and reduces operability. It is important to first consider the minimum unit of deformation you wish to observe and the error tolerance needed for judgment.

The third fundamental principle is not to try to complete an investigation using point clouds alone. In deterioration surveys, three-dimensional shape alone is often insufficient for judgment. Changes in color, traces of water leakage, rust staining, surface delamination, the state of material degradation, and internal conditions detectable by sounding (tap testing) all require separate verification. Point clouds are a very powerful means of documentation, but survey quality becomes stable only when they are combined with photographs, field notes, close-up inspections, existing drawings, and past inspection records. Before implementation, clarifying what point clouds will be responsible for and what will be supplemented by other methods makes the division of roles clear.

The fourth principle is to be conscious of preserving data in a form that is comparable. One of the values of introducing point clouds into condition surveys is their usefulness for future comparison. To achieve this, operations should be set up so that handling of coordinates, the concept of reference points, acquisition range, viewpoint bias, recording timing, site conditions, and the like can be kept consistent. Even if you re-capture the same location later, if the conditions differ greatly each time the accuracy of comparisons will decrease. If you are considering long-term monitoring, you should design the initial recording with future comparison in mind.

The fifth principle is to look ahead to how the survey deliverables will be used. On site, acquisition itself tends to become the goal, but what really matters is how they will be used afterward. If you don’t consider how to reflect them in reports, who will view them when shared internally, who will perform cross-section and change-detection checks, and how to link them to the maintenance management ledger, point clouds are likely to end up stored without being utilized. Whether you organize them into formats that are easy to view, manage areas of deterioration with attributes, or build a mechanism for time-series comparison, the design of the outputs determines the effectiveness of implementation.

These five basics are not particularly difficult. However, many of the reasons why point cloud implementation fails in practice stem from skipping these fundamentals. Purpose, accuracy, complementary methods, comparability, and how results will be used. If you address these five points first, discussions about implementation become immediately more practical. Treating point clouds not simply as a "new method" but as a "design element to improve condition surveys" is the shortcut to success.

Practical Workflow for Utilizing Point Clouds in Deformation Surveys

To make use of point clouds in condition surveys, how they are integrated into the overall survey workflow is more important than the acquisition method itself. Here, we organize the process according to a practical, easy-to-reproduce workflow.

The first thing to do is to clarify the inspection targets and the items to be checked. The distresses you want to observe differ depending on whether the target is a bridge, a slope, or a building exterior wall. By deciding in advance which components, which distresses, and at what level of accuracy you want to capture, you can reduce missed shots and unnecessary data collection on site. Even simply separating areas that require a broad overview from those that require detailed inspection can greatly improve work efficiency.

Next, confirm the site conditions. There are many factors that affect point cloud quality, such as whether a clear line of sight can be secured, whether consideration is needed for vehicle and pedestrian traffic, whether solar radiation or reflections have a large impact, whether there are areas that are difficult to approach closely, and whether there is abundant water or vegetation. In condition surveys, it is important not simply to create clean three-dimensional data but to capture data in a state that can be used for decision-making. Therefore, taking site conditions into account, you need to plan acquisition positions, acquisition sequence, and how to record the necessary auxiliary information.

On-site data acquisition is most effective when you consciously focus on both overall understanding and key areas. First, cover the target broadly so the overall spatial relationships of its shape are clear, and then concentrate on recording locations suspected of abnormalities and known deformations. This reduces the risk of making an incorrect judgment by looking only at local areas. It is also important on site to properly keep photographs and notes in addition to the point cloud. Even if there are areas of concern on the point cloud, it can be difficult to judge them without information on surface color changes or material feel.

In post-acquisition processing, first verify the consistency of positional relationships. If the data were acquired in multiple passes or the target covers a wide area, check as early as possible for any seams or misalignments, because failing to do so can lead to major rework downstream. After that, proceed to clean up unnecessary parts, extract the required area, inspect cross-sections, and check differences. In deformation surveys, merely viewing the entire point cloud rarely yields useful insight, so it is important to convert it into views that are easy to compare. Depending on the survey objectives, you need presentations such as viewing deviations from a reference plane, checking cross-sections at regular intervals, and comparing differences with previous years' data.

In the reporting and sharing stage, it is important not to simply hand over the point cloud itself and stop there. Clarify where the deformation locations are, how they appear, and what can be interpreted, and, as needed, present planar positions, cross-sectional information, and surrounding context together so that stakeholders can make decisions more easily. In particular, when stakeholders are unfamiliar with point clouds, three-dimensional data alone may not convey the information well. Produce location maps of the deformations, focused cross-sections, explanations of the magnitude of change, and correspondences with photographs, and translate these into a form that can be used for decision making.

Furthermore, organizing materials in preparation for ongoing surveys is indispensable. Deterioration surveys often do not finish after a single inspection, and their value increases through long-term monitoring. Therefore, standardizing items such as the acquisition date and time, scope of inspection, how reference baselines are set, and methods for identifying abnormal locations will make comparisons in subsequent surveys easier. The more thoroughly these details are organized in the initial survey, the lower the operational burden will be in later rounds.

Thus, point clouds are not determined solely by on-site acquisition techniques; they are effective as part of a workflow that includes survey planning, on-site decision-making, processing methods, report organization, and ongoing operation. To successfully incorporate point clouds into deterioration surveys, it is important to focus not on the creation of three-dimensional data itself but on how to connect it to deterioration assessment.

Common Pitfalls When Getting Started

More and more sites are showing interest in adopting point clouds, but in actual operations there are several typical stumbling blocks. Understanding these in advance makes it easier to improve the effectiveness of the implementation.

The most common issue is that point cloud acquisition becomes an end in itself. Even if you see value in being able to preserve things in three dimensions, if what you want to view is vague the results won't be useful. For example, it's not uncommon to capture a wide area at high density only to end up using photos in the report, or to find that data taken for difference comparisons can't be compared because the reference conditions weren't consistent. In condition assessments, the design intent before acquisition determines the value of the results.

Next is the tendency to expect more versatility than is warranted. While point clouds are strong at capturing shapes, they have limitations in interpreting fine surface anomalies and assessing internal conditions. There are many situations that require other investigation methods, such as precise verification of crack widths, determining the causes of leaks, and detecting internal cavities. Trying to substitute everything with point clouds alone can actually lower diagnostic accuracy. Point clouds can be the main tool in some cases, but that presupposes combining them with supporting records.

Another major challenge is that operations cannot keep up with the large volume of data. Point clouds, while rich in information, place significant burdens on management, sharing, viewing, and storage. If large volumes of data are accumulated without considering who—site staff, inspection staff, design staff, or the client—will use them in which situations, they end up not being looked at again. In condition surveys, it is important to organize the necessary scope, at the required accuracy, and in the required form of presentation; there is no need to always aim for the maximum volume.

The challenges of comparison-based operations are often overlooked. Examining differences is a strength of point clouds, but comparisons require that prerequisites be consistent. If acquisition coverage, control points, the condition of the target surface, seasonal factors, or the presence of obstacles change, you may observe differences caused by changing conditions rather than actual deformation. Especially outdoors, changes in vegetation, puddles, temporary structures, and other obstructions affect comparison accuracy. If you don't keep "future comparisons" in mind from the initial acquisition, the valuable data you collected will be less useful for ongoing management.

Furthermore, a disconnect between field staff and office staff is also a cause of implementation failure. If the processing team has not been informed about what was intended to be captured on site and which locations were important, they cannot analyze the data as required. Conversely, even if the office side has decided on the required cross-sections and comparison methods, if those assumptions are not shared in the field, re-acquisition may become necessary. When using point clouds in condition surveys, coordination between field personnel and office personnel is particularly important.

To avoid these stumbling blocks, it can be effective to start small when introducing the technology. Rather than rolling it out across all sites at once, begin with cases where deformation is easy to detect, where ongoing comparisons have high value, or where you want to reduce proximity to hazards—starting in situations where the effects are clearly visible makes it easier to establish an operational model. The value of point clouds changes far more depending on how they are embedded than on whether they are introduced. The challenge is less about the technology itself than about how you root it in your business workflows.

Summary

The use of point clouds in condition surveys is not just about creating three-dimensional representations; it is a practical means to objectively record, compare, and share the location, extent, shape, and relationship to surrounding features of anomalies. It is particularly effective for anomalies that involve shape changes such as deformation, steps or offsets, surface unevenness, bulging, and missing parts or defects. However, judging fine surface anomalies and internal conditions requires combining point clouds with other methods, and it is important not to place excessive expectations on point clouds alone.

Before introduction, you should clarify the objectives, sort out the required accuracy and required density, decide the division of roles with other recording methods, design the recording with future comparisons in mind, and set up operations that include how the results will be used. If you stick to these basics, point clouds are much more likely to help achieve both improved quality and labor savings in deformation surveys.

In practice, it's extremely important not only to record the defect location itself but also where it was captured, how its positional relationships were recorded, and how it can be overlaid for later review. In that sense, handling site coordinates, checking reference/control points, and linking photographic records with location information form the foundation for using point clouds. When you want to streamline the organization of this basic information, systems that leverage high-precision positioning like LRTK are effective. If coordinate checks and geotagged records can be carried out smoothly on site, the initial steps of a damage survey become easier to organize, and comparing and sharing point cloud data becomes simpler. If you want to carry out damage surveys in a more practical and reproducible way, it's important to focus not only on using point clouds but also on the accuracy of simple surveying and position records that support the work before and after. As an option that supports creating that onsite foundation, LRTK helps streamline daily maintenance operations.