When those working on excavations and record preservation want to preserve archaeological features accurately as photographs as well as drawings, orthoimages become important. Orthoimages, which allow you to check the planform shape, extent, intersections, subtle soil-color differences, and the spatial relationships of artifacts with minimal distortion, greatly influence the quality of the record. However, in practice it is often misunderstood that simply taking a photo from directly above is enough, and after shooting people frequently face problems such as “dimensions don’t match,” “edges are degraded,” “parts are missing,” or “cannot be overlaid on drawings.”

Archaeological features differ from clear-outlined subjects like building exteriors: their surface relief is small, and judgments rely on soil texture and color differences. Therefore, even slight disruptions in shooting conditions can create large differences between readable and unreadable images. The feature surface is also sensitive to sunlight, shadows, puddles, footprints, progress of excavation, and reflections of temporary materials, and sometimes photography becomes a race against time. This article explains the mindset needed to stably produce orthoimages of archaeological features in practice, and divides the fail-safe creation procedure into six steps. It is useful not only for those who will be photographing on site, but also for those who want to clarify conditions when outsourcing work or assess the quality of deliverables.

Table of Contents

• What is an orthoimage of archaeological features?

• Step 1 Decide the purpose and deliverable conditions first

• Step 2 Arrange on-site reference points and shooting range

• Step 3 Shoot with attention to being directly overhead and overlap

• Step 4 Perform additional shooting to cover elevation differences and blind spots

• Step 5 Carefully perform georeferencing and orthorectification

• Step 6 Check scale, distortion, and missing parts and prepare for delivery

• Common failures in creating orthoimages of archaeological features

• Summary

What is an orthoimage of archaeological features?

An orthoimage is an image in which tilt and perspective distortions that occur in photographs are corrected so that the image can be handled like a plan. A regular photograph will show the same-length object differently at the center and the edge of the frame, and a slight change in shooting position alters the sense of dimension. Orthoimages, by contrast, are corrected based on positional and shape information, so lengths and areas are easier to handle on a plane and they are suitable for overlaying with drawings or other survey products.

Orthoimages are useful in recording archaeological features because they become materials that can be re-read later, not merely commemorative photos. On excavation sites, once digging progresses the original state cannot be restored. Therefore, accurately preserving the feature surface at a given stage has great significance. With an orthoimage, it is easier to share information with stakeholders who are not on site, to check plans, organize note positions, produce illustrations for reports, and provide a baseline for longitudinal comparison.

However, an orthoimage is not “a single sheet made by stitching photos together.” Even when a surface appears flat, in reality there are cuts, upstands, steps, surviving walls, fill boundaries, and these affect the assumptions for correction. If there is insufficient overlap during shooting, reference points are vague, angles are not sufficiently overhead, or shadows are too strong, the result may look clean at first glance but be unusable in practice. In short, the quality of an orthoimage of a feature is determined by the preparation before shooting as much as by the shooting itself.

Step 1 Decide the purpose and deliverable conditions first

The first step to avoid failure is to clarify in advance why you are creating the orthoimage. If you start shooting without this clarity, you will either produce images that are coarser than necessary or expend excessive effort, and waste will inevitably occur downstream. First, organize the intended use of the deliverable. Resolution, positional accuracy, and color fidelity required vary depending on whether it is for report publication, to aid in making measured drawings, for stakeholder sharing, or for long-term archival re-examination. Determining from the outset whether you only need the general overall shape or whether you need to read subtle stratigraphic boundaries and feature contours will determine shooting height, field of view, necessary number of images, and on-site verification methods.

Next decide the target extent. Whether you will crop only the feature, include surrounding reference lines and artifact layouts, show the entire trench in one image, or separate by feature will drastically change the shooting plan. In practice, people frequently later say “I wish I had included a bit more outside,” so set the shooting range with margins beyond the necessary area. Shooting right to the feature’s edge may leave insufficient usable area if edge quality degrades during processing.

Also crucial is which coordinate system or reference will be used. Every site has different premises—site-specific references, grid lines of the survey area, consistency with existing survey results. If you plan to overlay the orthoimage on a plan later, unify how references are taken at this stage. Trying to force alignment later using only images can produce apparent agreement while leaving detailed mismatches, undermining the reliability of the material.

Decide shooting timing according to purpose as well. Is it a mid-excavation record or the cleaned final state? Do you want to avoid post-rain dampness or morning low-angle light? These considerations may require adjustments to site workflows. Treating photography as independent of excavation flow can cause more footprints, leftover materials, light changes, or surface drying and color change. In short, the first step to creating orthoimages is not choosing equipment. It is concretely specifying what you want to read, by what reference, and which state you want to record. Once this is set, the subsequent steps all become rational.

Step 2 Arrange on-site reference points and shooting range

The next step is to prepare the foundation for the orthoimage on site. Orthoimages of features are not completed solely by the photos; they stabilize only when reference points or lines and a shootable site condition are all in place.

First and foremost, clearly identify positional reference points on site. It may be tempting to use only the four corners or a few representative points, but when the area is large or the shape complex, local distortions easily occur, so place references evenly over the whole area. References are useful not only for image processing but also for verifying deliverables and aligning with other materials. Install them in a way that anyone on site can see, that they will appear in photos, and that they do not impede work.

References should not be placed thoughtlessly. It is important that they do not move, are not lost, and are effective not only at the edges but also within the area. If references are biased to one part of the feature surface, shifts may be hard to detect in the center or the opposite side. Especially in excavation areas, visibility changes easily due to scaffolding, spoil, or temporary materials, so check on site whether they are clearly visible from the intended shooting positions.

At the same time, pre-shoot cleaning and leveling are indispensable. Unwanted items left on the feature surface, scattered tools, temporarily placed spoil, footprints, and water creating strong reflections are not only obstacles to image processing but also hamper interpretation. Of course, items that should be retained as part of the excavation record must remain, but at least remove anything unnecessary for the shooting purpose. Since orthoimages capture a large amount of information, they also equally capture unnecessary information. That is why pre-shoot site tidying directly affects quality.

Also ensure sufficient perimeter outside the shooting range. Shooting exactly to the feature’s outline can result in necessary parts being cropped during trimming or peripheral distortion correction. Leaving some margin increases image stability and makes the deliverable easier to handle. While you can later cut away unnecessary parts, you cannot add missing peripheral area afterwards.

Another often-forgotten check is lighting. While direct sunlight on a clear day may look good, depending on surface relief and soil texture strong shadows can make boundary reading difficult. Conversely, an evenly bright sky may better capture surface color differences uniformly. The important point is not to judge times or weather absolutely good or bad, but to choose the conditions on the day that allow the most readable information. In this step, adopt the mindset of preparing the site into a state suitable for the deliverable before shooting. Thinking “post-processing will fix it” is risky for orthoimages of features.

Step 3 Shoot with attention to being directly overhead and overlap

Once site conditions are set, begin shooting. The most important thing here is to maintain as stable an angle and consistent conditions as possible relative to the target surface while ensuring sufficient overlap. The quality of an orthoimage is determined not by simply taking high-resolution photos but by assembling a set of photos that are easy to process.

First, pay attention to being directly overhead. For orthoimages of features, the basic rule is to shoot as close to perpendicular to the target surface as possible. If many images are taken at oblique angles, edges may be stretched or the appearance at steps may change, reducing consistency after correction. Complete overhead shooting may be difficult due to site constraints, but keeping angle variation minimal and shooting under consistent conditions is important.

Next, image overlap is crucial. If adjacent photos have little common area, georeferencing becomes unstable and causes seams and local distortion. Conversely, if you maintain systematic overlap while shooting, processing stability increases greatly. In practice, rather than shooting randomly, plan rows or bands, decide an order, and design movement paths to avoid gaps.

Also avoid large changes in shooting height or distance mid-process. Variations in height change the per-image scale and make forced corrections likely. In wide survey areas, widening shooting intervals to save effort can lead to loss of resolution or consistency in parts of the area, degrading overall uniformity. What is needed is to record the whole area at an average quality. A central area of high quality with coarse surroundings is impractical.

Pay attention to blur and exposure fluctuations. With feature surfaces often monotonous, even slight blur destabilizes feature-point extraction. Strong backlighting, localized overexposure, deep shadows, and concentrated reflections also reduce processing accuracy. Bright-looking photos are not always good; what matters is whether uniform information across the surface is obtained.

Make it a habit to check images on site. If you leave thinking you’ll review them all later, you may find gaps or blur when reshooting is no longer possible. Feature surfaces change as work proceeds, so the same state may not be reproducible. Check a global overview image, enlarged views of representative parts, edge capture, and reference-point visibility on site, and immediately supplement any deficiencies. In this step, the focus should be on collecting material that can be correctly integrated later, not on getting a single perfect photo. Reducing the number of shots is less important than eliminating gaps and instability.

Step 4 Perform additional shooting to cover elevation differences and blind spots

A commonly overlooked point in creating orthoimages of features is how to supplement information that standard overhead shooting cannot capture. If the feature surface were perfectly flat, a simple shooting plan would suffice, but in reality there are shallow depressions, upstands, excavation walls, protruding stones, residual surface elevation differences, and blind spots caused by machinery or temporary structures. Simple planar processing is often insufficient.

Pay special attention to areas with steps or depth. Orthoimages are strong for planar interpretation but require care in handling three-dimensional changes. If you force the entire area into a single plane despite elevation differences, unnatural stretching or misalignment may occur at boundaries. In such places, supplement shape information by additional shooting and consider separating the target into different parts if necessary.

Also consider areas that are hard to see from directly overhead. Gaps in stonework, the shoulders of channels, upstands, and areas in shadow under covers are often deficient when viewed only from above. The important point is not to add oblique photos indiscriminately, but to understand which parts require reproduction and add oblique or auxiliary shots accordingly. Maintain overall overheadness while reinforcing problem areas.

Be mindful of time differences too. If additional shooting is performed later, changes in light direction or surface state can make the same area look quite different. Soil drying, lengthening shadows, and human traffic can alter conditions in a short time. Therefore, perform additional shots as close to the main shooting flow as possible and keep conditions similar. Shooting on a different day after noticing deficiencies may produce processable but hard-to-interpret images.

The key in this step is not to try to finish everything in a single uniform pass. Orthoimages of features must balance plan readability with actual shape information. Anticipate problematic areas and include necessary supplemental shooting in the plan to greatly improve final reliability.

Step 5 Carefully perform georeferencing and orthorectification

Post-shoot processing is a critical stage that determines the quality of the orthoimage. No matter how carefully you shot on site, a sloppy processing approach can produce a deliverable that is hard to use in practice. Conversely, in processing you must not merely merge images into one sheet; you must align them while preserving the on-site references and interpretability.

The base task is aligning the images. At this stage, use photo overlap and feature points along with the reference information secured on site to arrange the overall relationships. Do not blindly trust automatic processing results. Even when they look neat in software, some areas may be stretched or boundaries misaligned. Especially for feature surfaces with repetitive texture, partial misrecognition can occur without notice—be vigilant.

When orthorectifying, decide which surface to treat as the target plane. If the feature is generally flat, planar correction works well, but when local elevation differences are large, consider how to handle their impact. Uniform processing can make the image look tidy but diverge from the actual shape. Depending on the situation, separate target areas, refer to auxiliary information, or treat problem areas separately—make judgments according to the nature of the feature.

Also be careful in adjusting color and brightness. Large brightness differences between photos make seams conspicuous and make it hard to tell whether soil-color differences are real or due to shooting conditions. However, excessive cosmetic correction to improve appearance can erase subtle soil-color differences and discoloration information. For orthoimages of features, legibility outweighs prettiness. Tone correction is necessary for homogenization but should be limited so as not to remove on-site information.

Be cautious when removing unwanted objects. Remaining shadows, human silhouettes, equipment parts, or adjacent disorder reduce visual quality and can cause misreading. But forcibly removing such elements during processing can undermine the record’s reliability. Clearly define what to keep and what to remove in a practical record, and avoid unnatural edits that alter the original state.

A commonly overlooked aspect in processing is exporting with the final use in mind. Even with high-resolution source data, deliverables can become hard to use if they are over-compressed, have severed positional information, or are in formats difficult to overlay on plans. Organize files by intended use—viewing, overlaying on plans, archival—to make them easier for both the site team and those handling the deliverables. The key in this step is not to leave processing to machines. For orthoimages of features, trustworthiness as a record is more important than visual appeal. Therefore, in processing carefully balance appearance and measurability.

Step 6 Check scale, distortion, and missing parts and prepare for delivery

Creating an orthoimage is not complete the moment it is made. If you skip the final checks, the deliverable may become unusable later. The final step to avoid failure is to validate the completed image as a document and prepare it in a form usable in practice. First check scale consistency. Compare distances between reference points and known dimensions to verify there are no large discrepancies. Even if the whole seems correct, parts may have been stretched, so check multiple locations. Confirming the feature perimeter, center, and edges helps detect local distortion.

Next, inspect for image distortion and seams. If the feature outline unnaturally wavers, linear references bend, soil boundaries do not connect, or stone shapes collapse mid-way, the processing may have been forced. People familiar with the site often notice visual inconsistencies more readily, so have both the processor and the site team check the result.

Also confirm there are no missing parts. Even when the whole looks complete, some areas may have thin information, extreme drops in resolution, or parts obscured by shadow and hard to read. Pay special attention to edges, elevation-change areas, and places that required additional shooting. If required information cannot be read, the image is incomplete as a record even if it appears finished overall.

If the image will be annotated or overlaid with other materials, test the actual workflow during verification. For example, check whether overlaying plan lines feels natural, whether feature numbers and notes can be placed easily, and whether necessary information is retained at high zoom. An image file that is “complete” in itself has little value if it does not fit the practical workflow.

When preparing deliverables, balance viewability and archival robustness. Lightweight viewing copies may not withstand careful analysis, and extremely high-resolution masters can be inconvenient for site sharing. Therefore, organizing several versions by use reduces downstream confusion. Since feature records may be revisited years later, manage files so that people other than the original author can understand reference information, target extent, and the conditions at the time of creation. This final check is modest but extremely important. An orthoimage of a feature is not finished by shooting or by processing alone. It is only complete when future users can trust and use it.

Common failures in creating orthoimages of archaeological features

We have covered six steps, but several failures repeatedly occur in actual sites. Knowing these makes on-site judgment much easier.

First: thinking you shot directly overhead when in fact the angle was unstable. It is often hard to notice during work, but later you may find one side stretched or edge shapes distorted. Causes include variability in shooting position, changes in how the camera is held, and awkward postures to avoid obstacles. Being directly overhead must be ensured by following consistent conditions, not by feel.

Second: insufficient overlap. To reduce shots or save time, widening intervals leads to unstable alignment in processing. Soil surfaces are particularly monotonous, and without sufficient common information alignment often fails. The problem is lack of required overlap, not merely a large number of photos.

Third: weak reference points. Even if image processing can visually align things, overlays with existing maps or survey results may reveal mismatches. In feature recording, positional reliability outweighs appearance, so do not delay securing references.

Fourth: underrating shadows and wetness. Soil-color differences are critical for interpretation. But strong shadows or reflections from dampness make boundaries hard to read. A photo that looks good is not necessarily one that is easy to read. Prioritize the recording purpose on site and choose conditions that present surface information uniformly.

Fifth: ignoring elevation differences and applying planar processing. Even shallow steps can produce boundary distortion over a wide area. Especially for features with cuts or upstands, determine at the outset whether planar treatment is appropriate. If necessary, treat the target in parts for better readability.

Sixth: skipping the final checks. When short on time people often become complacent at the end of processing, but failing to verify scale, distortion, missing parts, and overlay capability leads to rework later. Since site conditions often change and reshooting may be impossible, verification is essential. What these failures have in common is treating orthoimages as mere photo organization. In reality you need both a measurement perspective and a record-material perspective. Simply adopting this mindset will significantly change on-site decisions.

Summary

To create reliable orthoimages of archaeological features, treat preparation, reference setting, processing, and verification as a continuous workflow, not just a matter of shooting technique. Concretely: first organize the purpose and deliverable conditions; next prepare on-site reference points and shooting environment; then shoot with attention to being directly overhead and ensuring overlap; perform additional shooting to cover elevation differences and blind spots; carefully carry out georeferencing and orthorectification; and finally check scale, distortion, and missing parts to prepare a deliverable usable in practice. Following these six steps brings orthoimages closer to being trustworthy records, not just visually pleasing images.

For practitioners searching for “遺構 オルソ,” what matters is not producing a beautiful single image but obtaining a deliverable that can be read later, overlaid with drawings, and easily shared among stakeholders. Because feature conditions change as excavation progresses, the value lies in how accurately you can preserve the information at that moment. Orthoimages are a very effective means for this purpose.

At the same time, creating on-site reference systems and handling coordinates can be a burden. If you want to link orthoimages smoothly to plans and other survey products, consider not only shooting but also how to capture positional information. Methods to combine on-site photo acquisition and high-precision positional capture have become more accessible in recent years. For example, using an iPhone-mounted GNSS high-precision positioning device such as LRTK makes it easier to organize site photos and point relationships while recording, facilitating orthoimage creation and subsequent plan integration. If you want to improve the efficiency and reusability of feature records, consider not only how to create orthoimages but also how to obtain high-precision positions on site.