Four Precautions and Countermeasures to Avoid Failure in Cultural Heritage Photogrammetry

Interest in photogrammetry is increasing year by year in the fields of cultural heritage documentation, investigation, and public utilization. Methods that can reproduce three-dimensional shape from photographs are easy to use in situations where contact with the object is difficult, and they make it easy to visualize shape and surface conditions, so they have become an attractive option for many practitioners. Especially for cultural heritage, where opportunities for re-acquisition are limited and on-site conditions or conservation policies impose significant constraints, the ability to record large areas in a short time is a major advantage.

On the other hand, cultural heritage photogrammetry does not succeed simply by taking a lot of photographs. Unlike general product or real estate photography, recording cultural heritage requires careful consideration in advance of many issues: protection of the object’s condition, requirements for reproduction accuracy, on-site constraints, and data design with future uses in mind. It is not uncommon to discover after shooting that images are insufficient, that required accuracy cannot be met, or that the data produced are usable for public display but not for research. Even when models look good, failures that make them impractical for actual use are especially common in the field of cultural heritage.

Moreover, difficulty varies greatly depending on the subject in cultural heritage photogrammetry. Stone monuments, wooden buildings, Buddhist statues, wall surfaces, excavated artifacts, site topography, stone walls, ruins, and decorated surfaces each require different shooting distances, focal lengths, light handling, shooting sequence, and approaches to control points. Whether the object is outdoors or indoors, on public display or not, in a stable conservation environment or not, or requires scaffolding also affects the results. In other words, to succeed in cultural heritage photogrammetry, design tailored to the object and objectives is often more important than the technology itself.

This article organizes four points where cultural heritage photogrammetry is particularly prone to failure, and explains, for each, the common on-site problems and concrete countermeasures. The content will be useful not only for those considering implementation for the first time, but also for those who have already carried out shoots but could not utilize the results as expected, and for those who want to clarify check items when outsourcing. Decisions made in cultural heritage documentation have long-lasting effects. That is precisely why it is essential to reliably check the precautions that should be addressed before work begins, from a practical viewpoint.

Table of contents

• Why failures are likely to occur in cultural heritage photogrammetry

• Precaution 1: Don’t proceed with shooting objectives and deliverables left ambiguous

• Precaution 2: Don’t underestimate on-site conditions and cause omissions or gaps

• Precaution 3: Don’t postpone accuracy management and alignment

• Precaution 4: Don’t produce data without considering post-delivery usage

• How to approach cultural heritage photogrammetry for success

Why failures are likely to occur in cultural heritage photogrammetry

The main reason failures commonly occur in cultural heritage photogrammetry is the large gap between photography and data use. On site, attention tends to focus first on taking photographs, and the number of images and coverage rate are often emphasized. However, what actually determines success or failure is how the data will be used after shooting. Whether it will be used as survey documentation, for a conservation ledger, for exhibition or education, or as material for restoration planning will change the required accuracy, presentation methods, and information that must be preserved. If these points remain ambiguous, the completed model may not meet the intended purpose and require re-shooting or reprocessing.

Moreover, cultural heritage objects are not always easy subjects. Glossy surfaces, low color contrast surfaces, densely detailed carvings, dark or confined spaces, high locations, and areas with access restrictions often combine to create conditions that photogrammetry struggles with. Small camera shake or subtle changes in shadow that would not be problematic for general subjects can manifest as missing geometry or distortion in cultural heritage. Especially when you want to record surface weathering, cracks, or fine tool marks, merely producing a visually plausible 3D representation is insufficient.

Another frequently overlooked issue is that re-shooting cultural heritage is difficult. For ordinary buildings or products, it is relatively easy to shoot additional images later if parts are missing. But for cultural heritage, conditions for a revisit—display periods, storage location, permission from managers, transport, lighting, weather, scaffold installation, safety considerations—are not easily met. Thus, how much can be reliably captured in the initial shoot is extremely important. Underestimating this premise increases the risk that data that seemed sufficient on site will become unusable later.

In addition, the longevity of recorded data is important in cultural heritage practice. It is not enough for data to be viewable immediately after creation; they may be revisited years later, sometimes more than a decade hence. If file names are not organized, coordinate or scale information is not retained, it is unclear what area was photographed when, or correspondence between original images and outputs cannot be traced, the value of the data declines significantly. Cultural heritage photogrammetry requires thinking not only about shooting technique but also about record design and operational planning.

Thus, failures in cultural heritage photogrammetry usually do not arise from lack of advanced technical skill. In many cases they stem from omissions in basic matters such as clarifying objectives, on-site checks, accuracy management, and delivery design. From the next section, we will look concretely at four particularly important precautions in practice, the common failures that occur, and countermeasures.

Precaution 1: Don’t proceed with shooting objectives and deliverables left ambiguous

The first thing to confirm in cultural heritage photogrammetry is the purpose of shooting. This may seem obvious, but in practice it is the point most likely to become ambiguous. On site, projects often start with the phrase “let’s just digitize it in 3D for now,” but that is insufficient as a requirement. Digitizing in 3D is a means, not an end. If the purpose is unclear, necessary shooting range, image resolution, accuracy level, and delivery format will not be decided.

For example, whether you want an attractive 3D model for exhibition, measurable data for inclusion in an investigation report, or a baseline record with references for future change comparisons will drastically change required specifications. For exhibition, overall appearance and material texture representation are emphasized. For research, accuracy of shape, reliability of scale, handling of missing parts, and relation to control points are prioritized. The same cultural heritage object will require different shooting and processing approaches depending on purpose.

A common failure is proceeding while stakeholders have different expectations for deliverables. Managers may expect preservation records, researchers may expect measurement-ready data, and public relations staff may expect content for publication; in such cases, a single shoot is often expected to satisfy all needs. The result is often data that are mediocre for all purposes: the overall model looks good but lacks detail for measurement; detailed parts exist but are weakly registered to the whole; there are images for publication but no management information for records.

A useful countermeasure is to verbalize at least five items before shooting: intended use, target scope, required accuracy, deliverables, and storage format. For example, clarify whether you need a full-view record or a partial record, whether you need to capture surface patterns and incised lines, whether dimensional measurement is required, or whether the data will be used as baseline material for future comparisons. Then concretize the finished image. Decide in advance how much you need—not only the 3D model, but also orthophotos, images for dimensional checks, high-density partial scans, a full set of original photos, and management materials that clarify positional relationships.

In cultural heritage, balancing whole and detail is especially important. Capturing only the overall shape is insufficient if detailed information is missing. Conversely, photographing details at high resolution is of limited use if the positional relationship to the whole is ambiguous. Therefore, plan separate shoots for overall capture and for detailed recording. Trying to cover everything with a single type of image from the start tends to result in neither being sufficient.

It is also important to align the granularity of deliverables. For built heritage, for instance, prioritize separate records not only for the overall exterior but also for base of columns, joinery, decorative elements, and visibly damaged areas. For stone objects or artifacts, inscriptions, missing parts, join traces, and heavily weathered faces should be consciously recorded separately; otherwise they will not be usable later. Under time pressure on site, operators tend to prioritize full views, but detailed information is often what proves necessary later.

When outsourcing, do not judge by deliverable names alone. “3D data delivery” does not reveal what is actually included. Specify whether the 3D model is lightweight for display or high-density for analysis, whether texture images are included, whether original images are retained, and in what format scale and coordinate information are provided. In cultural heritage records, well-managed information that can be referred to later often has more value than visually impressive displays.

The first step to success in cultural heritage photogrammetry is clarifying the goal before shooting. Once the purpose is clear, the necessary shooting methods, personnel, and check items naturally become apparent. If the goal is ambiguous, even careful subsequent steps are unlikely to produce satisfying results.

Precaution 2: Don’t underestimate on-site conditions and cause omissions or gaps

The next important on-site concern in cultural heritage photogrammetry is avoiding misreading on-site conditions. Photogrammetry reconstructs shape by overlapping photographs, so merely seeing the target is not enough. It is necessary to shoot from different angles with sufficient overlap under stable conditions. However, ideal shooting conditions are rarely present at cultural heritage sites, and success often depends on how well omissions are prevented within constraints.

A typical failure is photographing only the visible parts on site and later discovering insufficient reconstruction of rear or intricate areas during processing. Deeply shadowed areas of carvings, roofs and eaves, the underside of pedestals, recessed gaps in stone walls, back faces of structural members, and wall faces on narrow passage sides may be visually inspected on site, but if image overlap is insufficient, 3D reconstruction will be unstable. Cultural heritage shapes are often complex, and frontal visibility alone is frequently inadequate; a shooting route and viewpoint plan are essential.

Outdoors, weather and sunlight have a major impact. Strong direct sunlight can produce pronounced shadows that vary greatly between images and reduce processing stability. Conversely, too little light increases blur and noise, affecting fine-detail reproduction. Wet surfaces or highly reflective stone, metal, or lacquered surfaces can appear differently depending on angle, making it difficult to find matching points. Surface conditions of cultural heritage are not homogeneous, so relying only on general photography conventions can produce many images that are difficult to process.

Indoor cultural heritage sites also present challenges: insufficient light, uneven lighting color, confined spaces, restrictions on shooting positions, observation behind glass, and no-close zones. For display-case objects, reflections and glare can prevent satisfactory reproduction. Inside buildings, overlapping structural elements can create blind spots, and uniform floors or ceilings with few features can destabilize alignment. Rather than deciding that shooting seems possible just from a site visit, it is necessary to first identify where omissions are likely.

The most effective countermeasure is not to treat pre-shoot site checks as mere reconnaissance. Confirm access points, maximum shootable height, whether lighting conditions change, restrictions on traffic or public access, nearby obstructions, and whether auxiliary shooting routes can be secured. Then proactively include anticipated missing areas in your plan. Waiting until the day of shooting is too late, because these considerations affect shooting sequence and equipment needs; they should be organized during pre-shoot planning.

Also important is ensuring sufficient overlap. A high number of photos does not necessarily equal easy reconstruction. What you need is continuous viewpoint movement and stable overlap that links the target surface. Shooting rapidly with a wide-angle lens may leave insufficient detail for fine areas, and photographing only details can sever their relation to the whole. Design both a flow for linking the whole and a flow that captures priority areas.

On-site capacity for immediate verification is also indispensable to prevent omissions. For cultural heritage, it is often too late to notice missing parts after returning from the site, so enable at least minimal on-site checks: verify that the entire object is covered in a continuous loop, that likely blind spots have been addressed, that there are enough close-up images of important details, and that exposure and blur are acceptable. Having someone other than the shooter—one who understands the documentation purpose—check during operations reduces omissions.

Furthermore, consideration for the object must have top priority. Forced postures, excessive proximity, moving installed items, or intense lighting may be prohibited. Therefore, assume there will be parts you cannot photograph, decide in advance what you can reliably capture, and determine how to treat unphotographable parts in the record. Rather than pursuing perfect acquisition by risky means, the important practice in cultural heritage is to leave the most practical record within constraints.

If on-site conditions are underestimated, you may end up with many shots but unusable data. Conversely, if you anticipate areas prone to omission and build the plan assuming site constraints, you can achieve stable results even within limited time. In cultural heritage photogrammetry, the ability to perceive difficulty is required before photographic technique.

Precaution 3: Don’t postpone accuracy management and alignment

A common misunderstanding in cultural heritage photogrammetry is to equate the existence of a three-dimensional model with having a reliable record. Even if the model looks well finished, if dimensions are unstable, positional relationships are shifted, or the basis for scale is weak, its practical value drops significantly. Cultural heritage often requires future comparisons, investigative deliberations, drafting, and partial measurements, so accuracy management must be integrated from the planning stage rather than considered at the end of processing.

A frequent mistake is to build the model first and then try to adjust dimensions afterward. If the original images lack sufficient conditions, such adjustments will be impossible. For example, when length references are scarce, not evenly distributed across the object, shooting directions are biased, or image sets for parts are isolated, you may see the overall shape but fail to obtain correct scale or stable geometry. In cultural heritage records, small deviations can affect later interpretation, so an approach different from appearance-focused 3D modeling is necessary.

It is especially important to distinguish what kind of accuracy is needed. Whether it is acceptable that overall dimensions are approximately correct, whether you need to confirm local undulations, or whether you want to use data for multi-temporal comparisons changes the required management approach. For display viewing models, rough shape consistency may suffice, while for monitoring deterioration or comparing cross-sections, reproducibility of positional relationships is critical. Rather than using the vague term “accuracy,” start by deciding what you want to trust and to what degree.

As a countermeasure, first clarify the notion of references. According to the object size and site conditions, plan known lengths for scale checks, control points for alignment, and reference information that can be common across multiple acquisitions. Since attaching or placing references on cultural heritage objects may be restricted, devise non-invasive ways to secure references. The important thing is to leave the rationale for the achieved accuracy explainable: if it is unclear which information was used to determine scale or align positions, third-party trust in the data decreases.

Also ensure that shooting supports the whole evenly. Even if some areas have high-density images, a lack of images that link the whole can make the overall shape unstable despite local detail. For large cultural properties or complex structures, prepare images at multiple levels—overview, intermediate, and detail—and design them so each level connects. This reduces separation of whole and parts and makes later alignment and checks easier.

During processing, do not judge solely by visual appearance. Excessive hole-filling or smoothing can hide areas that were not actually captured, making the result look good but misleading. For cultural heritage records, it is more important to know which parts were reliably captured and which are uncertain. Using data without understanding gaps or uncertain areas risks incorrect interpretation. Emphasize the integrity of the record as well as aesthetic finish.

Furthermore, when planning multi-temporal comparisons, design from the first acquisition to facilitate re-acquisition. Tracking changes over time can be important for cultural heritage, but without baseline information for references and shooting conditions, future comparisons are difficult. Recording and preserving where, which area, and under what conditions you shot will improve reproducibility for re-shoots years later. Whether you can manage accuracy with future comparisons in mind greatly affects the long-term value of the data.

In cultural heritage photogrammetry, accuracy management is not only the concern of specialists. Commissioners, managers, and users alike need to share what they want to preserve as trustworthy data. Do not be content that the shape is visible in 3D; be aware of what can be measured, what can be compared, and what the limitations are—this awareness is key to avoiding failure.

Precaution 4: Don’t produce data without considering post-delivery usage

The last commonly overlooked precaution in cultural heritage photogrammetry is post-delivery operations. After shooting and processing, when the 3D data are complete, it may feel as if the job is done, but in cultural heritage practice the real stage of utilization begins there. Data created without considering how they will be used afterward tend to be difficult to revisit, share, and reuse. It is not uncommon for carefully acquired data to become unusable within months.

One common failure is having too many types of deliverables that are disorganized. If original photos, processing data, 3D models, surface images, verification images, and report materials are mixed and it is unclear which are the official deliverables, finding the needed file later becomes a burden. Cultural heritage staff are not always 3D data experts, so distinguish between viewing and archival files, and between public and record files. Ambiguous file names or folder structures significantly increase operational load.

Be wary of data that can be viewed but not used. For example, a model may be rotatable on screen but not suitable for dimensional checks; orthophotos may not be exportable; correspondence with original images may be unknown; or the photographed range may be unclear—these states limit a dataset’s value as cultural heritage documentation. Conversely, even if viewing convenience is limited, if acquisition range, date, object name, positional relationships, scale rationale, and usage conditions are well organized, the dataset will be an asset in the long term. In cultural heritage, practical, reusable structure often proves more helpful than flashy display.

As a countermeasure, consider in advance who will use the delivered files and for what. Determine what managers want to see, which data researchers need, which formats are easy for public distribution, and what should be retained for future reprocessing. Then decide the minimum information to include. For example, manage as a package: a full set of original photos, the deliverable itself, viewing-friendly data, documentation that shows the acquisition range, shooting dates and object information, accuracy and scale rationale, and notes on caveats. This makes future use stable.

Do not underestimate the importance of original images. In photogrammetry, the source images are as important a record as the final 3D model. If you later want to reprocess, create outputs for another purpose, or partially verify results, without the original images you cannot respond. Keeping only the 3D model while failing to preserve and organize originals greatly narrows reusability. For cultural heritage, where re-acquisition is difficult, preserving original images is especially important.

Also assume staff turnover when planning. Cultural heritage management spans long periods, so the staff who handled acquisition may not remain the same. If a new person sees the data years later and cannot identify the object name, shooting range, acquisition purpose, usage instructions, or caveats, the data will be effectively unusable. Therefore, leave explanations that are comprehensible even without specialist knowledge. They need not be advanced technical manuals; it is crucial that a third party can navigate the dataset without confusion.

When public distribution is intended, separate presentation and recordability. Lightweight, user-friendly data are convenient for public use, but alone they may not retain the original recordability. Public-facing formats and archival, verifiable originals serve different roles. If you combine everything into a single format without distinction, you may have data that are suitable for display but unsuitable for future research or reuse. Cultural heritage data should be designed not only for immediate usability but also for long-term preservation.

Designing for post-delivery utilization greatly affects the success of cultural heritage photogrammetry. By paying a little attention during shooting and processing, you can significantly extend the lifespan of the data. Cultural heritage is not finished being recorded once it has been captured; it will be referenced long into the future. With that premise, planning how to preserve and present the data is a crucial countermeasure to avoid failure.

How to approach cultural heritage photogrammetry for success

From the four precautions discussed above, it becomes clear that failures in cultural heritage photogrammetry are more often due to insufficient preparation than to a lack of technical skill. Avoid proceeding with unclear shooting objectives, do not underestimate on-site conditions, do not postpone accuracy management, and design with post-delivery utilization in mind. Even adhering to these four points will greatly improve the stability of outcomes.

For practitioners, it is important not to expect photogrammetry to be a panacea. While 3D reconstruction from photographs is a highly effective method, it has strengths and weaknesses depending on the object and conditions. That is why it is crucial to pre-organize what you want to preserve, how far you want to go, and where the limitations lie, and to design your approach according to on-site conditions. Rather than forcing a single method to solve everything, combining the means necessary for the record’s purpose will reduce failures.

Also, in cultural heritage recording, how to retain positional information and on-site connections is important. If only the 3D shape is retained, but the location of the cultural property and its relationship with the surroundings are unknown, utility is limited. For outdoor sites, stone monuments, and surroundings of buildings, including local coordinates and positional relationships makes it easier to compare later or overlay with other materials. In practice, linking shape data with positional information increases the value of the record.

In that sense, to make photogrammetry more practical in the field, do not treat 3D data as a standalone deliverable but consider linking it with positioning and on-site records. For example, for outdoor cultural heritage, ruins, and site records, being able to operate while capturing photographic and object positions at high accuracy facilitates comparison during revisits and correspondence with surrounding environments. The perspective of how to support understanding of the entire site as well as the object will become increasingly important.

If you want to more reliably advance cultural heritage documentation on site, preparing an environment that connects photogrammetric outputs with local coordinates and positioning information is a practical strength. LRTK, as an iPhone-mounted GNSS high-precision positioning device, provides mechanisms that make on-site position recognition and record tasks more efficient. It is an appropriate option for sites that want to seamlessly connect photographic records with positioning workflows, helping to raise cultural heritage photogrammetry from mere 3D digitization to on-site records usable in the future. Considering such positioning linkage in operations is a sound first step to preventing failures.