How to create an as-built heat map from point clouds｜Explained in 5 steps

In the field of as-built management, there is a strong demand to verify as quickly and as clearly as possible whether construction has been carried out according to the design. What has been attracting attention is the as-built heat map using point cloud data. Even in situations where limited survey points were traditionally checked individually, using point clouds makes it easier to understand the as-built condition as a surface and to visually grasp the distribution of excesses and shortages and the areas that require repair.

However, simply acquiring point clouds does not automatically produce a correct heat map. If practical points such as preparing the design surface, unifying reference coordinates, preprocessing the point cloud, and setting the conditions for difference calculations are not addressed, you can end up with diagrams that look nice but cannot be used for decision-making. In this article, aimed at practitioners seeking information on as-built heat maps, we clearly explain the workflow for creating as-built heat maps from point clouds, divided into five practical steps that are easy to use on site.

Table of Contents

• What is an as-built heat map from point clouds?

• Preparations required before creating an as-built heat map from point cloud data

• Step 1 Organize the design aspects and comparison criteria

• Step 2 Acquire and preprocess the on-site point cloud

• Step 3: Align the design surface and the point cloud

• Step 4 Calculate the differences and color-code them

• Step 5 Connect to as-built assessment and report generation

• Points to note when creating as-built heat maps from point clouds

• Summary

What is an as-built heat map from point clouds?

An as-built heat map created from point clouds is a diagram that compares the current point cloud data with the design reference surface or the finished shape and visualizes the differences using color. Typically, areas higher than the design are shown in warm colors, areas lower in cool colors, and areas within the allowable range in intermediate colors, allowing you to see at a glance where there is excess or deficiency on the construction surface. A major feature is that local bulges or continuous sags, which are easy to overlook in numerical tables, become easier to notice when displayed as surfaces.

In as-built management, the important thing is not merely to show differences in elevation. It is necessary to present the information in a state where decisions can be made, including which reference the differences are measured against, what area is subject to evaluation, and how much deviation is acceptable. The purpose of creating as-built heat maps from point clouds is not to make the site look flashy, but to support inspections and rework decisions, verify construction accuracy, and drive improvements in construction methods. Therefore, the validity of the comparison criteria is as important as readability.

Additionally, as-built heat maps are not limited to earthwork. They can be widely used to evaluate surface geometry—not only for checking large areas corresponding to developed faces, slopes, subgrade, and roadbed, but also for the top surface after concrete placement, graded surfaces, finish after backfilling, and interference checks around structures. On some sites they are also effective as supplementary documentation to capture construction irregularities that are hard to discern from plans and cross-sections alone. Especially at sites where multiple stages overlap, the meaning changes depending on which point cloud is used as the reference, so operation that keeps the relationship to the construction stages in mind is essential.

Furthermore, point-cloud–based as-built heatmaps have the advantage of helping to bridge differences in experience among personnel. While an experienced worker may notice something off just by looking at the site, less experienced staff can take longer to read anomalies from numerical tables. By showing deviations in color, it becomes easier to share priorities about where to check first, improving the efficiency of meetings and internal reviews. In other words, creating as-built heatmaps from point clouds is not only about measuring, but also about establishing a system for communicating.

Preparations required before creating as-built heat maps from point clouds

To produce a stable as-built heat map from point clouds, it is no exaggeration to say that the result is largely determined by the preparatory work. First you need the design data that will serve as the comparison target. If the three-dimensional reference surface—including design elevation, design surfaces, and slope shoulders and toes—is left ambiguous, the assumptions for the difference calculation cannot be established. Even when you only have two-dimensional drawings, you must first clarify which sections to connect to create the comparison surface and where to cut off the edges of the construction area. If the method for creating the comparison surface is unclear, the difference results will look unnatural even if the current point cloud has no problems.

Next, it is important to unify the coordinate system and vertical datum. On site, it is not uncommon for horizontal positions to match while the vertical datum differs, or for temporary local coordinates to be mixed with the official site coordinates. In this situation, even if you overlay the point cloud and the design surface, any color discrepancies will reflect a mismatch in reference rather than construction errors. Because heatmap colors are intuitively readable and can therefore appear convincing under incorrect assumptions, you must always check the reference points, how elevations were determined, the origin setting, and whether any rotation has been applied.

Defining the evaluation scope is essential. For example, for a slope, whether the evaluation target is from the slope shoulder to the slope toe, whether to include small benches, or whether to exclude interface areas with structures will change how the results appear. While an as-built heat map allows comparison of the entire surface, including areas that should not be compared increases noise. Deciding in advance how to handle puddles, material storage areas, parked positions of heavy equipment, and unconstructed sections will improve the reliability of the difference maps.

Additionally, the method for acquiring the point cloud should be decided during the preparation stage. Whether you measure from the ground, acquire from the air, generate a point cloud from photos, or create it based on distance measurements, each approach has different strengths and weaknesses for different subjects. Surfaces that are easy to see from above are suited to aerial acquisition, but areas that are hard to view—such as alongside walls or under overhangs—are more likely to be missed. Conversely, ground-based acquisition can capture fine details more easily, but it requires more time over wide areas. Consider which method is most appropriate based on the objective and the shape of the target.

Finally, clearly defining in advance what the heat map will be used for helps prevent inconsistencies in downstream processes. Whether it is for internal review, extracting areas for rework, presentation materials for the client, or as an aid to inspection forms will change the required resolution, the granularity of the color coding, and how it is mapped onto drawings. On site, attention tends to focus on the data processing itself, but if you start from who will ultimately view it and how they will judge it, you can avoid producing unnecessarily heavy data and are more likely to create an as-built heat map that is practical for use in the field.

Step 1 Organize design aspects and comparison conditions

The first step is to clarify what you are comparing. When creating an as-built heat map from a point cloud, if the comparison target is ambiguous, even carefully carrying out the subsequent steps will not yield correct results. The current point cloud represents the form that actually exists on site, but design-side information can include multiple candidates: the target surface at completion, the control surface during construction, or a reference surface for a specific process. First, you need to decide which surface you want to check at that time.

For example, the design surface you use changes depending on whether you are evaluating the developed/graded surface or the finished top surface. If you compare a surface that is still under construction to the final completed surface, unintended discrepancies will appear across the entire area. Conversely, if you use the reference surface from the temporary construction stage after completion, you will not be able to correctly assess the level of accuracy. Because heat maps make differences visually obvious through color, choosing the wrong comparison surface can easily result in misleading documentation.

It's important to be able to explain the comparison conditions in writing. If you organize the assumptions—such as which design surface was used, which area was compared, when the point cloud was acquired, and whether any areas were excluded—internal sharing and explanations during inspections will go more smoothly. In field work, it's no use if only the person handling the processing understands the conditions. Putting the comparison conditions into words so that anyone will interpret them the same way underpins the reliability of the as-built heat map.

Also, it's a good idea to align how you define the direction (sign) of deviations at this stage. For example, will values higher than the design be treated as positive, or will they be treated as negative to indicate insufficient excavation relative to the design? If you don't standardize notation rules on site, the same color can be interpreted differently. If each person defines it differently, comparing with past data becomes difficult. Because the color representation in heat maps, while improving readability, also carries a risk of misinterpretation, it's important to decide on the sign convention from the outset.

Furthermore, you should clarify the concept of comparison resolution. Point clouds can contain very fine detail, but if the design surface is coarse or such fine discrimination is unnecessary as a finishing tolerance, producing unnecessarily fine differences can make surface roughness appear as anomalies. Conversely, averaging under conditions that are too coarse will miss local bumps or depressions. Depending on the site’s management objectives, assume whether you want to inspect at the centimeter level (cm, in) or the millimeter level (mm, in), and deciding the granularity of the comparison is the first key.

Step 2 Acquire and preprocess the on-site point cloud

The next step is to acquire the on-site point cloud and prepare it for use in comparison. Point cloud data are voluminous and at first glance appear information-rich and reassuring, but in reality they contain many unnecessary points, missing data, reflection noise, and the inclusion of extraneous objects. The quality of the as-built heat map is influenced far more by the quality of preprocessing than by the difference calculation. In other words, before comparing the correct surfaces, you must carefully reselect what to adopt as the current condition.

When acquiring data, the first thing to be aware of is whether the surface under evaluation has been sufficiently observed. On wide developed/graded surfaces, point density decreases with distance, and on slopes portions can be missing depending on the angle. Around structures, points are often lacking in shaded areas, and after rain or near water surfaces unstable points can be mixed in due to reflections. If you create a heat map while such gaps and disturbances remain, areas that are actually unobserved can appear as anomalies. It is important to identify blind spots during the acquisition stage and, if necessary, perform supplementary measurements.

In preprocessing, we first remove obvious unwanted objects. If items unnecessary for assessing the finished surface—such as heavy machinery, workers, materials, temporary structures, traffic vehicles, or swaying vegetation—remain, only those parts will show extreme colors. In practice, these extreme colors can be so conspicuous that they make it difficult to see trends in the actual constructed surface. Although point clouds appear to record everything at the site, only the points related to the surface being evaluated are needed for as-built assessment. Removing unnecessary points according to the purpose leads to clearer, more usable heat maps.

On top of that, balance noise removal and smoothing. If you remove points with large variability too aggressively, even real irregularities will be erased; conversely, if you do nothing, minute noise will appear as color unevenness. What you should be mindful of here is not making the appearance prettier, but faithfully representing the actual condition of the constructed surface without excess or omission. If the smoothing used to make it look smooth is too strong, localized steps that actually exist can be averaged out and become invisible. Especially since as-built heat maps are used for inspection and rework decisions, excessive prioritization of appearance should be avoided.

Also, thinning point clouds must be done carefully. If the data are too large, processing efficiency drops, but non-uniform thinning can cause the stability of difference calculations to vary by location. Because the density required differs between areas with gentle surface slopes and areas with large variations, simply reducing the number of points is not sufficient. It is important to suppress only the excess density that is unnecessary for comparison while maintaining the required resolution. In practice, one is expected to achieve both lightweight processing and reliable results.

Furthermore, extracting the surface to be evaluated is also at the core of preprocessing. Whether it’s the slope, the roadbed, or the top surface, isolating only the surfaces you want to compare makes subsequent difference calculations and color-coding much clearer. Conversely, processing multiple surfaces together can produce unnatural differences at the boundaries and leave color-range settings ambiguous. When creating as-built heat maps from point clouds, it is easier to interpret the results if you handle areas separately according to the evaluation unit rather than processing a wide area all at once.

Step 3 Align the design surface and the point cloud

The third step is to correctly align the design surface with the as-built point cloud. This alignment is one of the most critical processes that determines the success or failure of the as-built heat map. Even if the site's point cloud is high quality, a slight misalignment with the design surface can produce an overall color bias. That makes it impossible to tell whether the issue is a construction defect or a coordinate offset, greatly reducing the value of the heat map.

When aligning, you need to pay attention not only to planar (horizontal) offsets but also to vertical (height) offsets. In practice, even when the planar overlap looks plausible, a slight difference in the height reference can make the entire surface appear uniformly higher or lower. When such a uniform trend appears, rather than immediately questioning construction accuracy, it is important first to check consistency with control points and known points, the reference elevation of the design surface, and whether height corrections have been applied to the point cloud. You should examine the color distribution to distinguish whether the issue is a field error or a reference shift.

When checking alignment, verifying not only the entire surface but also representative feature points helps stabilize accuracy. For example, confirming several locations that are easy to identify both in the design and in the as‑built condition—such as edges, break points, slope change points, and locations where elements meet structures—makes it easier to notice local offsets. Even if the overall average appears to match, slight rotation in one direction or misalignment in only a portion can be more readily detected by examining feature points.

Also, instead of creating a heat map immediately after alignment, inserting cross-section checks can reduce failures. Data that appears to match on a plan view can show offsets between the design lines and the as-built points when viewed in cross section. This is especially true for sloped surfaces and embankments, where planimetric alignment alone is not sufficient. By checking the relationship between the design surface and the as-built point cloud on several representative cross sections before proceeding to compute differences, the heat map colors become more convincing.

What must not be forgotten in the alignment process is to leave it in a state that can be reproduced later. If there are no records of who used which reference points, what corrections were applied, and which version of the design data was adopted, recalculation and additional verification become difficult. As-built management is not finished by issuing drawings once; conditions may need to be reviewed later or comparisons made with other work sections. Keeping reproducible alignment procedures reduces rework on site and improves the quality of data operations.

Step 4 Calculate the differences and color-code them

The fourth step is to compute the differences between the aligned design surface and the point cloud and color-code them as a heat map. Only at this stage does a figure resembling the as-built quality heat maps commonly seen on construction sites take shape. However, while the difference calculation and color-coding are processes concerned with visualization, they also reflect the decision criteria. It is not enough for colors to simply appear; what to show, over what ranges, and at which thresholds is critically important.

In difference calculations, you quantify whether the as-built point is above or below the design surface and by how much. What you should be aware of here is the difference in approach between taking the difference as the shortest distance and taking the difference in the vertical direction. The appropriate way to compute differences changes depending on whether the surface you want to evaluate is nearly horizontal or has a slope. On slopes such as embankment faces, evaluating based solely on simple height differences can give a result that deviates from reality, so it is important to choose a comparison method suited to the nature of the surface.

When using color coding, setting the tolerance range is the practical key. If the tolerance range is made too wide, problem areas will be buried under neutral colors. Conversely, if the settings are too strict, even minor surface roughness will appear abnormal and the entire map will be covered in warning colors. To make an as-built heatmap that can be used on-site, you need to decide—based on actual management standards and construction objectives—how much difference you want to highlight. It is important to choose a color span that viewers can intuitively understand while also standardizing the numerical ranges that form the basis for judgment.

Furthermore, the impression changes greatly depending on how you set the color scale’s upper and lower limits. If you widen the color range to accommodate extreme outliers, most areas will appear the same color and important variations will become hard to see. Conversely, if you make the color range too narrow, even slight differences will look like major problems. In practice, it is easier to understand if you separate a figure that shows overall trends from a figure that zooms in on areas of concern. In other words, rather than trying to explain everything with a single heat map, it is important to use different visualizations according to the purpose.

Also, attention is needed to how parts with no detected differences or with insufficient points are represented. If unobserved areas or excluded regions are colored in normal tones, they may be mistaken for areas without problems. Conversely, making invalid values too conspicuous can draw attention away from the information that is actually needed. Clearly indicating parts that cannot be evaluated is important for maintaining the trustworthiness of the heatmap. When creating as-built heatmaps from point clouds, it is essential to make the meaning of colors clear and to avoid mixing colors that indicate “do not view” with colors that indicate “safe to view.”

Finally, it is important to adopt a perspective that does not complete judgment based solely on the heat map. Color-coding is well suited to detecting anomalous areas, but it will not identify the cause. Whether a localized elevation is a construction tolerance issue, point-cloud noise, or the inclusion of an extraneous object must be determined in conjunction with cross-sectional views and a review of the raw data. Rather than treating the heat map as the final conclusion itself, regard it as a powerful intermediate deliverable for visualizing inspection priorities, which makes it more practically useful.

Step 5 Connect to as-built evaluation and report generation

The fifth step is to connect the generated heat map to as-built assessment and report generation. Even if a difference map has been created up to this point, it will not fully deliver business value unless it can be translated into on-site decision-making. What matters to practitioners is organizing, in an explainable form, which areas are within tolerance, which require confirmation, and which are candidates for rework. In other words, a heat map is not useful merely because it was created; it only becomes a practical document once it is connected to judgment and record-keeping.

The first thing to do is to examine the color distribution and distinguish overall trends across the surface from localized anomalies. If the entire surface is slightly shifted in the same direction, you should suspect issues with reference setting or alignment rather than the construction itself. On the other hand, if only a specific banded area shows a color change, it may be caused by uneven compaction, spreading and leveling, or finishing work. A heat map not only highlights abnormal areas but also provides clues for interpreting patterns in the construction process.

Next, narrow down the areas that require rework. In conventional inspections, it was necessary to infer which parts of the site to repair from numerical checks at each measurement point, but with a heat map the continuity of color makes it easier to visually grasp the repair extent. However, you should not decide the extent based on color alone; it is important to judge in combination with section checks, on-site photos, and construction history. In particular, boundary areas and interface points are prone to point cloud gaps and the influence of extraneous objects, so it is more reliable to make the final confirmation carefully with an on-site visual inspection.

When preparing reports, rather than simply pasting the heat map by itself, arranging it in a format that conveys the underlying conditions increases the report's value. Organizing information such as which design aspect it was compared to, when the data were acquired, what the evaluation range is, how the color legend is defined, and which areas require verification will make it easier to judge when reviewing it later. As-built heat maps are visually strong materials, but because insufficient explanation of the conditions can easily cause misunderstandings, it is important to present the figure and the explanatory text together.

Using heat maps to quickly align stakeholders’ understanding is also effective. When the construction team, surveying team, and supervision team each look at numbers separately, interpretations can vary, but materials that present information as an area make it easier to share the discussion points. Because they make it easy to show not only where the problems are but also where they are not, they can help avoid unnecessary rework. For keeping site schedules, as-built heat maps have value as materials that speed up decision-making.

Furthermore, operating by comparing heat maps from multiple points in time is also effective. Rather than performing a single as-built check, comparing heat maps during construction and after finishing, or before and after rework, allows you to share improvement effects quantitatively. With this accumulated data, it becomes less likely that the same defects will be repeated on the next site, and it leads to reviewing construction plans and work procedures. If you treat heat maps not as one-off deliverables but as records of site improvement to be cultivated, the benefits of implementation become even greater.

Considerations when creating as-built heat maps from point clouds

One common mistake when creating as-built heat maps from point clouds is assuming that a visually tidy figure is necessarily a correct result. While color visualization makes interpretation easier, it also tends to conceal mismatches in the underlying assumptions. For example, if the design surface’s edge treatment differs, unconstructed areas are mixed in, extraneous objects remain, or the height reference is inconsistent, you may be able to compute the differences themselves but end up with a heat map that cannot be used for as-built evaluation.

Particular attention should be paid to overconfidence in the density and quality of point clouds. A large number of points does not necessarily mean they are correct. If there are not enough points on the surface you want to evaluate, or if the surface looks rough because it was captured from an angle, many small color variations will appear. Interpreting these directly as construction errors can lead to unnecessary corrective actions. It is important to judge, based on the point cloud acquisition conditions and the nature of the surface, to what extent the data can be trusted.

Also, consistency in color settings should not be overlooked. If the color range and color order differ from site to site, comparing with past projects becomes difficult and viewers’ judgments can be swayed. If one site shows a slight difference in red while another site still shows a large difference in an intermediate color, confidence in the heat map itself declines. Aligning basic color rules and the approach to display ranges within the company or department greatly improves the readability of materials.

Insufficiently delineating evaluation targets is a common challenge in practical work. Combining a wide area onto a single map may appear efficient at first glance, but when locations with different surface characteristics are mixed, the criteria for color-coding become inconsistent. Treating horizontal surfaces, slopes, and areas around structures under the same conditions will force compromises somewhere. While heat maps are well suited for wide-area display, dividing evaluation units appropriately improves decision accuracy. You need to adopt the approach of using an overall map and detailed maps.

Furthermore, it is important not to make final decisions based solely on the heat map. Color anomalies are only candidates for irregularities, and to determine their causes and priorities you need to combine cross-sections, site photographs, construction records, and visual inspections. To make the documentation trusted in practice, it is essential to use the heat map not as an all-purpose answer but as an entry point to streamline verification. If this role is correctly understood, the practice of generating as-built heat maps from point clouds will steadily improve the speed of on-site decision-making and the ability to explain findings.

Summary

Creating an as-built heat map from point clouds is not simply a matter of colorizing the points. You need to prepare the design surface to be compared, preprocess the as-built point cloud, perform correct alignment, color-code the differences so their meaning is clear, and link the results to as-built judgment and reporting—only then does it become a deliverable usable in practice. By following the five steps explained here, you can avoid the common situation with as-built heat maps of being visually striking but unusable for decision-making.

What matters on-site is not so much creating the heat map itself as stabilizing the comparison conditions and positional accuracy. If these remain ambiguous, no matter how advanced the processing, you cannot have confidence in the results. Conversely, if you organize the design surfaces, standardize the coordinates, and clarify the evaluation range, an as-built heat map becomes a very powerful field management tool. Because you can grasp the as-built condition as a surface, it greatly aids in determining rework areas, explaining the situation to stakeholders, and streamlining inspection responses.

To make as-built management using point clouds easier to run on site, it is important to organize operations not only for point cloud processing but also for control point verification, supplemental surveying, and stakeout. If you want to streamline such field operations, combining an iPhone-mounted high-precision GNSS positioning device like LRTK makes it easier to carry out daily positioning and verification tasks while ensuring positional reliability. Rather than treating the workflow of creating as-built heat maps from point clouds as a one-off analysis task, establishing it as a system that improves the overall quality of on-site as-built management will become increasingly important in practical work going forward.