What is the allowable tolerance for an as-built heat map? A simple explanation of how to read the Ministry of Land, Infrastructure, Transport and Tourism standards

When you want to know the allowable tolerance of an as-built heat map, many practitioners first want to confirm, “How many mm is acceptable?” “Is it a failure if red or blue appears?” and “Where in the MLIT standards should I look?” But the conclusion is that the allowable tolerance for an as-built heat map is not a single fixed value. Because the referenced standard values change depending on the work type, the measurement item, whether it is surface management or cross-section management, and even the evaluation method, you cannot judge uniformly by looking only at the heat map.

In current 3D as-built management, the basic idea is surface management: converting measured point clouds or surface data into prescribed meshes and comparing the deviations from the 3D design data on a surface basis. In ICT-utilized earthwork projects, if the earthwork quantity is 1,000 ㎥ or more, a heat map is created as an as-built management chart, securing a point density of at least 1 point per 1 ㎡, and the acceptance of the as-built is judged accordingly.

Therefore, in practice it is more important to organize and read “which work type, which measurement item, and which standard value the heat map is referencing” than to just follow the “color of the heat map.” In this article, I’ll explain as simply as possible for practitioners how to read the allowable tolerance of a heat map.

Table of contents

• What an as-built heat map is

• The allowable tolerance is not “a single number”

• Items actually checked in the MLIT standards

• Meaning of heat map colors and how to read them

• The line between pass and fail

• Points that cause confusion in practice

• On-site confirmation procedure

• Summary

What an as-built heat map is

An as-built heat map visualizes the finished construction surface by measuring it, comparing the results to the design data, and showing the magnitude of deviations with colors. In surface management, measured surface data and 3D design data are compared point-by-point to calculate elevation differences or horizontal differences, and these are compared to standard values for evaluation and judgment. In other words, a heat map is not just a cosmetic figure; it is part of an as-built management chart used to evaluate deviations from the design on a surface basis.

What’s important here is that the heat map does not color the “measured values themselves,” but visualizes the “deviations from the design.” For example, work types managed by elevation look at elevation differences, while work types managed by position sometimes look at horizontal differences. Therefore, even if two heat maps appear to have similar color distributions, the meaning changes depending on whether the subject of evaluation is elevation or position. You must first confirm what difference the work type is looking at to interpret it correctly.

Also, MLIT-related 3D as-built management requires sufficient measurement density as a premise for surface management. In the earthwork implementation guidelines, securing a point density of 1 point per 1 ㎡ or more—equivalently, intervals of 1 m (3.3 ft) or less—within the as-built management measurement range is shown as a condition for surface management. Even if the colors are nicely rendered, if the required density is not met, that heat map alone does not constitute a proper as-built evaluation.

A common misunderstanding in practice is, “The heat map is just presentation material; the real judgment is separate.” Of course the figure has strong explanatory power, but in reality it is incorporated into the pass/fail judgment as an as-built management chart, so it is not merely a reference figure. Especially in ICT-utilized projects, because the surface-based acceptance judgment concept is used together with the heat map, it is important that field staff, supervisors, and inspectors align their reading of the map.

The allowable tolerance is not “a single number”

If someone asks, “What is the allowable tolerance of an as-built heat map in mm?” you can’t answer with a single number because the standards differ by work type and measurement item. In the field, the term “allowable tolerance” is easy to use, but in actual standards the wording is often organized as “standard values,” and the figure changes depending on what is evaluated—average, maximum, minimum, range, thickness, reference elevation, etc. In other words, the allowable tolerance of a heat map is not a number intrinsic to the heat map itself; it is easier to understand if you think of it as the standard value that the heat map is comparing to.

For example, in paving work separate standard values are set for individual measurements and the average of 10 measurements. In lower base course work, for individual measurements the tolerance might be ±90 mm (±3.54 in), while for the average of 10 measurements it might be +40 mm (+1.57 in) and −15 mm (−0.59 in); thus, even for the same location the representation of the standard value changes depending on the measurement item. If you look only at the heat map without knowing this, you may judge by feeling—“a little red, so it’s dangerous,” “there’s a lot of blue, so it’s too low”—but actually you must view the average, maximum, minimum, and other specified conditions together to determine pass or fail.

Meanwhile, for earthwork embankment construction, standard values for the average, standard values for individual maximum and minimum differences, required data counts, and allowable rejected point counts are combined. For example, the average might be ±50 mm (±1.97 in), the maximum difference +150 mm (+5.91 in), the minimum difference −150 mm (−5.91 in), required data count at least 1 point per 1 ㎡, and allowable rejected point count 0.3 % or less; if even one of these is out of the standard, the result is a failure. Thus, to say “allowable tolerance of the heat map” is to check a bundle of multiple standard values.

Keeping this idea in mind changes how the field appears. Knowing the heat map’s allowable tolerance means confirming “for this work type, which measurement item is being judged by which standard value.” Conversely, it is dangerous to judge from color alone without knowing the work type and measurement item. This is especially important on sites where multiple work types progress simultaneously; do not import the sense of one work type into another. Because heat maps are created as relative evaluations against standard values, this is a fundamental, unavoidable point.

Items actually checked in the MLIT standards

The most important point in reading the MLIT standards is that “the heat map alone does not determine pass/fail.” As-built management charts list items such as average, maximum (difference), minimum (difference), data count, evaluation area, and rejected point count, and you confirm whether all of these satisfy the standard values. The surface color distribution is very conspicuous, so people tend to look only at that, but formal judgment assumes reading the summary table and the distribution figure together.

The paving example illustrates this well. In a passing example, the average is 12 mm (0.47 in), the maximum is 70 mm (2.76 in), the minimum is −45 mm (−1.77 in), the data count is 8,000 points, the evaluation area is 7,000 ㎡, and the rejected point count is 0—each of which meets the corresponding standard values. However, in a failing example, even if the average itself is within the standard, the maximum might be 100 mm (3.94 in), exceeding the standard of 90 mm (3.54 in), or the rejected point count might be 23 points, exceeding the 0.3 % upper limit of 21 points, causing failure. Even if most of the appearance is within the acceptable color range, local standard violations or excess rejected points can sway the overall judgment.

The same applies in earthwork. In a failing example for embankment construction, the average might be −60 mm (−2.36 in), which falls outside the average standard of ±50 mm (±1.97 in), and there may be six points shown in black as outside the standard, exceeding the rejected point limit of three points. Thus, even if only part of the heat map appears problematic, if any single item in the summary table is outside the standard, the entire as-built fails. Without this understanding, the mistaken judgment “most of it is green so it’s fine” easily happens.

Another point to keep in mind in practice is the interpretation of “all measured values satisfy the standard.” Supervision and inspection guidelines using 3D data indicate the idea that if 99.7 % of the as-built evaluation data satisfy the standard for individual measured values, it is treated as “all measured values satisfy the standard.” Conversely, the allowable rejected point upper limit is 0.3 %, and exceeding this may result in failure. The common field impression that “a few points outside is okay” often arises from misunderstanding this point, so caution is required.

Meaning of heat map colors and how to read them

Heat map colors do not simply map mm differences directly, but are basically shown as a ratio to the standard value. Supervision and inspection guidelines require that deviation calculation results be expressed as a percentage of the standard value, color-coded point-by-point within a range of −100 % to +100 %, and that a legend be clearly shown. It is also desirable to use separate colors to distinguish around ±50 % and ±80 % and to display out-of-standard ranges in a different color.

Therefore, the same green color may indicate a margin of several centimeters in one work type and several millimeters in another. Although the color itself is the same, its meaning changes if the standard value differs. In other words, a heat map should be understood as a visualization of relative margin or risk compared to the standard value, not an absolute-value figure. It is inappropriate to place heat maps of different work types side by side and assume the same color means the same condition. This is the practical way of reading derived from the distribution figure being drawn by standard-value ratio.

Also, paying attention to report layout makes things easier to understand. In example reports, the summary table on the left or top lists numbers such as average and maximum in mm, while the distribution figure on the right or bottom shows the color coding of the percentage to the standard. Thus, the summary table is “the section for numeric judgment,” and the heat map is “the section to grasp distribution and bias.” The most important MLIT standard practice is to never judge by color alone but always confirm together with the numeric table.

It is also organized that, at the client’s request, it is desirable to indicate in the figure the number of measured points within 50 % of the standard and within 80 % of the standard. This helps not only for pass/fail but also to understand the margin of finish and stability of quality. In practice, this is useful not only at completion but also for mid-construction as-built checks to early detect tendencies like “it’s passing now but the margin is small” or “there is local bias.”

The line between pass and fail

If you summarize the pass conditions for a heat map in one sentence: “The ratio to the standard must fall within the prescribed range, the required number of points and area must be secured, and the rejected point count must be below the upper limit.” In the paving example, passing conditions are: standard value ratio within 100 %, data density at least 1 point per 1 ㎡, and rejected point count 0.3 % or less. Failing conditions are standard value violations, data density less than 1 point per 1 ㎡, or rejected point count exceeding 0.3 %. The same logic applies to earthwork.

A particularly easy-to-overlook practical point is that even a small number of out-of-standard points can directly lead to failure. In the earthwork example, with an individual standard of ±150 mm (±5.91 in) there were six points shown in black (out of standard), exceeding the rejected point upper limit of three points and thus failing. In the paving example, even though most of the area was within acceptable colors, there were 23 black points exceeding the upper limit of 21 points, resulting in failure. Even if the heat map looks relatively neat, you must always check whether black or outlier points exist or are concentrated.

Conversely, the presence of red or blue does not automatically mean failure. If the percentage relative to the standard falls within −100 % to +100 %, it is treated as within the acceptable range on the legend. That is, warm or cool colors may appear, but if they only indicate bias within the standard, they are not failures. What matters is not the presence of colors but which percentage band the color falls into, whether it reaches an out-of-standard band such as black, and whether the items in the summary table satisfy the standard values.

Furthermore, surface management is not always the fixed top priority. The earthwork implementation guidelines state that when surface management is inefficient due to limited single-measurement area (for example, when as-built management timing is divided multiple times), or when surface management cannot be performed due to snowfall/accumulation, management by measured cross-sections or change points may be chosen in consultation with the supervising officer. Depending on site conditions, choosing an appropriate management method in consultation may be a standards-compliant response rather than insisting on producing a heat map.

Points that cause confusion in practice

The first common misconception is “if it’s mostly green it’s probably passing.” In reality, if any one of average, maximum, minimum, data count, evaluation area, or rejected point count is outside the standard, the result is a failure. In particular, the maximum and rejected point count shown in the summary table can have a stronger impact on the judgment than the overall color impression. Heat maps are excellent for grasping overall trends, but never break the sequence: final judgment must always be checked against the numeric table.

The second is assuming “the color bar represents mm.” In fact, the colors in the distribution figure represent the ratio to the standard, so the actual mm difference for the same +50 % varies by work type and item. Therefore, it is inappropriate to compare forms from different work types and conclude “this site has stronger blue so it’s more dangerous.” First check what standard value the report uses, then read the colors.

The third is thinking “if there’s a heat map, there’s no need to re-measure on site.” A heat map is only a figure organizing measurement results, and it presumes appropriate point density and correct correspondence with design data. If point density is insufficient, measurement conditions were poor, or design data handling was incorrect, a nice-looking figure is weak as a basis for judgment. The reliability of as-built management is determined by measurement conditions and consistency of the report, not by the figure’s appearance.

The fourth is “if the color scheme worked in past sites, it will work this time too.” Management standards and operation documents are updated annually, and regional bureau pages continuously organize revised editions. Therefore, do not directly reuse legends or judgment instincts from past reports; confirm the version for the current project year, special specifications, and discussions with the supervising party before judging. Even when referring to MLIT standards, always confirm the latest edition.

On-site confirmation procedure

The first practical step is to identify which work type and which measurement item the heat map targets. The reference standard changes depending on whether it is earthwork or paving, whether it is reference elevation or elevation difference, thickness or width. Without clearing this up, red or blue have no meaning. First identify the work type and measurement item and verify the corresponding standard value. Simply following this order prevents many misreads.

Next, confirm the prerequisites for surface management. Check whether a point density of 1 point per 1 ㎡ or more is secured, whether the required data count for the evaluation area is met, and whether the comparison conditions with the design data are set correctly. If these are inadequate, it is a problem prior to color distribution. Especially when receiving outsourced deliverables, check not only the heat map image but also the summary table, evaluation area, measurement point count, and if necessary the availability of original data so that explanations are easier later.

Then review the items in the summary table in order. Confirm that the average, maximum, minimum, data count, evaluation area, and rejected point count are all within the standard values; if any are outside, that immediately affects pass/fail. Next look at the heat map to check whether out-of-standard colors are locally concentrated or whether there are biases in the 50 % or 80 % bands. Use the summary table to confirm pass/fail and the heat map to grasp bias and construction tendencies. This two-step reading is the most practical approach.

If surface management does not suit site conditions, do not force the creation of a heat map; instead consult with the supervising officer to choose an appropriate management method. On sites where construction timing is divided or where weather conditions have strong influence, cross-section management or change-point management may be more rational. The essence of the standard is not producing a heat map itself, but properly confirming that the construction is as designed. That is why it is necessary to consider drawings, measurements, reports, and consultations together.

Summary

The most important thing in understanding the allowable tolerance for an as-built heat map is to know that “the allowable tolerance is not a single number set on the heat map.” In practice, standard values exist by work type and measurement item, and the heat map colors are expressed as percentages of those standard values. Moreover, pass/fail is not decided by color alone but by the summary table read together with the heat map, including average, maximum, minimum, data count, evaluation area, and rejected point count. In other words, a simple way to put the MLIT standard is: “Before looking at the color, check what the figure is comparing.”

In practice, it is more important to correctly grasp deviations from the design, and to quickly detect where margins are sufficient and where values are approaching the standard, than to produce a pretty heat map. To do that, capturing the current state quickly and accurately on site and having a system to re-measure immediately if necessary are strengths. If you want to streamline from simple surveying to the pre-stage of as-built confirmation, using an iPhone-mounted high-precision GNSS positioning device like LRTK can facilitate on-site positioning, recording, and preparation for comparison with the design. Understanding how to read heat maps and making measurement itself lighter, faster, and field-driven will become increasingly important in future as-built management.