What is surface evaluation by heat map? Explaining MLIT compliance from four perspectives

Among practitioners investigating MLIT-compliant quality control, many may have heard the term “heat map” but find it hard to understand what is actually being evaluated and how far they need to go. Sites accustomed to traditional cross-section checks often find that, although color-coded diagrams are easier to understand visually, they frequently have not organized and understood the thinking about pass/fail criteria, the role of report forms, and the measurement prerequisites.

In MLIT‑related materials, within the flow of i-Construction, surface-based construction control assuming multi-point measurement and the staged organization of discrepancy evaluation against 3D design data have been developed. In other words, a heat map should not be seen as mere easy-to-read visualization; it must be understood as part of a management method that uses 3D surface data acquired over an area to quantify differences from design surfaces and connect those evaluations to supervision and inspection.

Table of contents

• Basics of surface evaluation by heat map

• Perspective 1: The meaning of evaluating the entire surface

• Perspective 2: How to read heat maps and determine pass/fail

• Perspective 3: Practical workflow to secure MLIT compliance

• Perspective 4: Common points of confusion on site

• Summary

Basics of surface evaluation by heat map

In one sentence, surface evaluation by heat map is the idea of overlaying post-construction 3D surface data with the 3D surface data required by the design, displaying the pointwise differences as a color-distributed map, and evaluating the as-built condition against specification values. In materials from the Kanto Regional Development Bureau of the Ministry of Land, Infrastructure, Transport and Tourism, surface management is organized as converting the surface data into prescribed mesh sizes, calculating elevation differences or horizontal differences between corresponding points created from the evaluation surface data and the 3D design data, and comparing and judging those against specification values. In short, the essence of a heat map is not “adding color” but “reducing the large number of points acquired as a surface into evaluable rules and quantitatively assessing differences from the design.”

It is important to note that heat maps do not exist alone but are handled within the framework of 3D as-built management. The guidance compiles 3D design data, various 3D measurement technologies, evaluation methods for as-built conditions, and implementation flows by work category, and a heat map appears as the output of that process. First there must be 3D design data, then area-based measurement at the site, and only after calculating the differences does the evaluation diagram become meaningful. Put differently, creating a visually pleasing heat map after the fact is weak as an MLIT response if measurement conditions, evaluation ranges, and consistency with design data are insufficient.

Also, surface evaluation is not simply cross-section management with more points. Early i-Construction materials presented the idea of enabling surface evaluation of as-built shapes obtained by photogrammetry or laser scanning so that multi-point information could realize efficient area-based construction control. The traditional method of selecting representative cross-sections to check height or width can miss fine undulations or local anomalies between sections, whereas surface evaluation makes it easier to grasp variability across the entire construction area. Practically, this ability to “see the whole” is the primary value of adopting heat maps.

Perspective 1: The meaning of evaluating the entire surface

The main reason surface evaluation attracts attention is that it allows viewing as-built conditions not as a “representative value” but as a “distribution.” Traditional management focused on checking reference heights, widths, thicknesses at control sections and judging whether those values fall within specification limits. Although this method has been used for a long time, it can fail to reveal local bulges, settlements, or variability as a continuous surface between sections. MLIT materials contrast traditional representative section checks with 3D as-built management that enables surface evaluation of as-built shapes obtained by UAVs and other means. Surface evaluation is therefore a system to confirm construction results as a surface rather than a line.

This difference affects not only supervision and inspection but also contractors’ self-management. If you can view the distribution of differences over the entire surface during construction, you can identify early where excess fill remains and where shortages occur. Even if the mean value is within specifications, concentrated outliers can affect final pass/fail. Thus, heat maps are useful not only as documents submitted at completion but also as intermediate management tools to prevent rework. MLIT trial materials explicitly state that 3D measurement-based as-built management aims to improve quality and streamline work, and that simply collecting many point clouds is not the objective. Surface evaluation should be understood as a tool to balance quality and efficiency.

Furthermore, surface evaluation aligns well with quantity management and future data utilization. i-Construction materials indicate using multi-point measurement results that do not specify locations other than control sections for quantity calculation and the use of surface-based design and as-built data. In other words, data acquired through surface evaluation can lead not only to inspection diagrams but also to quantity calculation and 3D model utilization. For site staff, understanding “why hold data as a surface” is more important than merely producing a heat map, as this understanding is key to adapting to BIM/CIM and data utilization trends.

In addition, dividing the evaluation target into appropriate units is a practical point. FAQs organize that for earthwork excavations, the evaluation range should basically be a continuous single surface, while adjacent slopes separated by small benches may be evaluated together or separately. For arrangements like slope–flat–slope, it is acceptable to evaluate grouped by management item. This means that making one heat map is not the end; practical judgment must include which surface to evaluate, under which management items, and how to subdivide. The essence of surface evaluation lies not only in measuring widely but also in appropriately partitioning evaluation units.

Perspective 2: How to read heat maps and determine pass/fail

When looking at a heat map, it is easy to focus only on red and blue distributions, but for MLIT compliance what really matters is the meaning of the differences and the judgment conditions, not the colorfulness. The Kanto Regional Development Bureau materials present a heat map that shows the percentage of the calculated deviation relative to the specification range from −100 percent to +100 percent, plotting results by evaluation-data point with a color legend made explicit. In other words, color is not the absolute truth but an intuitive means to show where a point lies relative to specification bands. Do not decide pass/fail by the visual impression of red or blue alone; read what band that color represents relative to the specification.

For pass/fail judgment, multiple indicators checked in summary tables are important, not just the color distribution of the heat map. In the paving example, six items—mean value, maximum value, minimum value, number of data points, evaluation area, and number of rejected points—are all checked, and only if all meet specification values is the result judged pass. For example, even if the mean is within specification, exceeding the maximum allowable value is a fail, and if the number of out-of-specification points causes rejected points to exceed 0.3 percent, that is also a fail. Heat maps are excellent for getting an overview visually, but final decisions should be made based on the entire set of management documents including these indicators. Practitioners should first grasp that “heat maps are easy to read, but pass/fail is determined by numbers.”

A common misunderstanding is thinking that “a few black or dark points won’t be a problem overall.” However, the materials set fail conditions such as exceeding 100 percent of the specified range, insufficient data points, or rejected points over 0.3 percent. Thus, even if the distribution looks gently varying overall, exceeding those conditions at out-of-specification points can result in failure. Conversely, an apparently heterogeneous color pattern can be acceptable if values fall within specification, required data density is met, and rejected-point conditions are satisfied. On site, always treat “visual unevenness” and “specification-based anomalies” as distinct concepts.

Understanding how differences are calculated for the heat map also helps prevent misreading. The Shikoku Regional Development Bureau operation guide shows an example of extracting one point per 10 cm (10 cm (3.9 in)) grid cell to compute differences for a heat map and explains methods such as averaging, nearest neighbor, TIN, and inverse distance weighting. It also states that chamfered areas cannot be evaluated and that a 5 cm (2.0 in) strip at the edge of 3D design data should be excluded from evaluation. Thus, heat-map colors are not an all-purpose truth; results vary depending on the difference-calculation method and excluded areas. Practically, save not only the color distribution map but also the calculation conditions used to produce it.

Perspective 3: Practical workflow to secure MLIT compliance

The first thing to grasp in MLIT-compliant practice is that heat map creation is not “the last task of measurement” but a management act that must be prepared by backwards planning from the construction planning stage. The guidance suggests organizing in a plan the applicable work categories and areas, measurement locations for as-built management, as-built management standards and specification values, and equipment and software to be used. In other words, whether to produce a heat map should not be decided at the end; you must determine before construction which work types, which ranges, and by what methods you will manage surfaces—otherwise later measurement density or evaluation ranges may be insufficient and the data unusable.

Next is selecting measurement methods appropriate to the work type and site conditions. The guidance organizes multiple measurement technologies by work type, including aerial photogrammetry, terrestrial laser scanners, vehicle-mounted laser scanners, UAV-mounted laser scanners, total station non-prism methods, total station optical EDM methods, RTK‑GNSS, construction history data, terrestrial photogrammetry, acoustic sounding equipment, and mobile-device-based 3D measurement techniques. Thus, there is no single “correct” MLIT device. What is needed is choosing a reasonable method in light of the work type, specification limits, construction extent, presence of obstructions, safety, and required density. Matching required accuracy and application conditions is emphasized over naming specific device models.

It is also necessary to meet measurement density and accuracy requirements. For example, in earthwork surface management, accuracy and density for pre-construction surveys and as-built management measurements are organized by device, and an as-built evaluation point density of at least one point per 1 square meter is indicated. Some device examples require densities equivalent to 0.5 m (1.6 ft) mesh or 0.1 m (0.3 ft) mesh. The key point is that even if final reports look correct, if the source data density does not meet requirements the basis for evaluation collapses. A heat map is a post-processed representation, but its quality is largely determined at the on-site measurement stage.

Consistency with design data is also crucial. 3D design data are defined as surface data output as TIN or similar, including road centerline geometry, cross-section normals, as-built cross-section shapes, construction control point information, and coordinate system information. A heat map only makes sense when comparing this design surface with evaluation surface data, so if control points or coordinate systems are ambiguous, the map may look tidy but lack a sound evaluation basis. In sites with multi-day measurements or multiple devices, maintaining coordinate consistency and unified data rules is more important than the measurement itself.

Finally, consider submission materials for supervision and inspection. 2024 trial materials present reference cases where 3D-measured models with embedded measurement data or AR technologies are used to confirm as-built conditions on site directly, streamlining or omitting traditional as-built management tables. However, in general practice the standard flow remains producing heat maps by overlaying design data and evaluation data and submitting them as as-built management tables. Thus, the pragmatic stance is: “Although submission formats may change in the future, for the time being proceed on the assumption of submitting diagrams and tables while confirming contract requirements and whether trials are in effect.”

Perspective 4: Common points of confusion on site

The most important practical caution is that “surface evaluation by heat map cannot be applied identically to all work types.” FAQs explain that for soil‑retaining frame works, as-built management is dimensional control using measured point clouds rather than heat-map evaluation using 3D design data, and preparing 3D design data for as-built management is not necessary. In other words, while heat maps are a convenient evaluation method, certain work types or structural conditions may retain approaches closer to traditional dimensional control. If site staff assume “MLIT compliance = always a heat map,” they may do unnecessary work or prepare incorrectly. First check the applicable work-type guidance and determine the expected evaluation method.

Next, a heat-map report is not always uniformly mandatory. The FAQ notes that the guidance does not require heat maps as mandatory as-built management forms and advises consulting the client. There are also examples where intermediate inspections use traditional management, and at completion the entire area is measured for surface as-built management with heat maps used for uninspected areas while inspected areas are treated as reference values. This shows that in practice it is more important to agree with the client on “at what stage, over what range, and how to handle heat maps” than simply whether to create them. Do not focus solely on the presence or absence of a heat map; prepare by reviewing the overall inspection-time operations.

There are pitfalls in setting evaluation ranges. As mentioned earlier, chamfered areas or edge strips may be excluded from evaluation, and whether to separate slopes and flat areas changes report thinking. A common mistake is combining all acquired data into a single heat map so that it becomes unclear which parts were judged by which criteria. While surface evaluation provides an overview, ambiguous evaluation units and exclusion criteria make the documentation hard to justify. Early in the process, organize on drawings “what is included in the evaluation” and “what will be treated as separate surfaces” to prevent downstream confusion.

Another important point is confirming the latest versions. Related materials show that MLIT’s as-built management guidance is updated, and regional bureau guidance and FAQs are continuously supplemented. Searchable materials include FAQs for the March, Reiwa 6 version and the March, Reiwa 8 version of the guidance. Therefore, do not simply reuse past site practices; check the bid timing, contract documents, special technical provisions, content of client consultations, and the latest guidance together. Rather than using “MLIT compliance” as a formal phrase, preparing according to the rules in effect at the time of the work is the most reliable approach.

Summary

Surface evaluation by heat map is not about creating a pretty colored diagram but about quantitatively grasping differences between 3D design surfaces and post-construction surface data over the entire area and judging as-built conditions against specification values. Key MLIT compliance points are: first, unlike cross-section management, it can capture variability and local anomalies across the entire construction area; second, pass/fail is judged not by color impression but by indices including mean, maximum, minimum, data count, evaluation area, and rejected points; third, plan measurement methods, ensure accuracy and density, align with design data, and set evaluation ranges before construction; and fourth, some work types are subject to dimensional control rather than heat-map evaluation, and the handling of reports may change depending on contract and consultation. Seen this way, a heat map is not a standalone diagram but an output document that helps run 3D as-built management correctly.

Implementing this approach on site requires not only large-scale measurement and final report production but also making routine control-point checks, small-area positioning checks, and as-built confirmations as quick, light, and reliable as possible. Especially if coordinate consistency and on-site positioning underpinning surface evaluation are unstable, later heat-map production will be inefficient.

In such cases, using LRTK, an iPhone-mounted high-precision GNSS positioning device, can facilitate on-site coordinate checks, simple surveys, and efficiency around control-point work. To ensure smooth final surface evaluation, establishing a system that enables easy, routine high-accuracy positioning on site makes a significant practical difference.