Comparing Cave Mapping Workflows: Manual Survey vs. Photogrammetry

1. Field Context: Where the Choice Matters

Every cave project starts with a question: what kind of map do we need, and how much time and gear can we afford to haul underground? The answer shapes whether you reach for a compass and tape or a camera rig. Manual survey and photogrammetry are not interchangeable—they suit different environments, team sizes, and end goals.

Manual survey has been the backbone of cave mapping for over a century. A team of two or three people runs a series of shots: distance, azimuth, inclination, and sometimes back sights for closure. The equipment is simple—a compass, clinometer, tape or laser rangefinder, and waterproof notebook. It works in tight crawls, vertical drops, and passages where setting up a tripod is impossible. The data comes out as a line plot, which can later be expanded into a 2D map or 3D model with software like Survex or Walls.

Photogrammetry, on the other hand, captures the cave geometry from overlapping photographs. You take hundreds or thousands of images through a passage, then process them in software (e.g., Metashape, RealityCapture) to produce a dense point cloud, mesh, and textured model. This method excels in large chambers, decorated passages, and areas where you need high-resolution detail for documentation or conservation. But it demands good lighting, stable camera settings, and significant post-processing time on a powerful computer.

The choice often comes down to the cave itself. A tight, muddy crawl with low visibility is a nightmare for photogrammetry but routine for manual survey. A large, dry chamber with clear walls is photogrammetry's sweet spot. Many teams now use a hybrid approach: manual survey for the overall framework, photogrammetry for detailed sections. Understanding the strengths and weaknesses of each workflow helps you plan efficiently and avoid wasted effort underground.

1.1 Typical Scenarios

Consider a project mapping a new section of a limestone cave. The entrance is a tight squeeze, then opens into a 20-meter-wide chamber with flowstone and stalactites. A manual survey team can push through the crawl quickly, taking shots every 2–3 meters. In the chamber, they could spend hours running a grid of shots to capture the shape, but the detail will still be coarse. Photogrammetry could capture the chamber in 30 minutes of photography, but the crawl section would be impossible to image properly. The best plan: manual survey the crawl, photogrammetry the chamber, and tie them together with control points.

1.2 Team and Time Constraints

Manual survey is more forgiving of small teams—two people can work effectively. Photogrammetry usually needs at least three: one photographer, one lighting assistant, and one to manage the camera settings and battery changes. The time underground is often shorter for photogrammetry in open passages, but the post-processing time can be 10–20 times the field time. Manual survey spreads the work more evenly: you spend hours underground and hours at the computer drafting. The trade-off is that manual survey data is easier to verify in the field—you can check closure errors and re-shoot bad shots immediately. Photogrammetry errors often only appear after processing, when you're back home.

2. Foundations Readers Confuse

Many cavers conflate accuracy with resolution, or assume that more data always means a better map. These misunderstandings lead to wasted effort and mismatched expectations. Let's clear up the most common confusions.

Accuracy vs. Precision. Accuracy is how close your measurement is to the true value; precision is how repeatable your measurements are. Manual survey can be very accurate if you use good instruments and follow proper procedures (e.g., calibrating compass for local declination, using back sights). Photogrammetry can produce high-precision models (mm-level point spacing) but accuracy depends on scale bars, control points, and camera calibration. A dense point cloud with poor control can be precisely wrong—every point is consistently offset by a meter. Always prioritize accuracy first; precision is a bonus.

Data Density vs. Usefulness. A photogrammetric model with millions of points looks impressive, but for most cave mapping needs—passage width, ceiling height, overall shape—a few hundred survey stations give you the same information. The extra detail is only useful if you need to document formations, monitor change over time, or create a virtual tour. For a basic survey to guide future trips, manual survey is often sufficient and faster overall.

Time Underground vs. Total Time. Beginners often think photogrammetry saves time because you spend less time in the cave. But total time includes setup, processing, and cleanup. A manual survey that takes 4 hours underground might take 2 hours at the computer. A photogrammetry project that takes 2 hours underground can take 20 hours of processing, plus another 5 hours for cleanup and control point refinement. The total time is usually higher for photogrammetry, unless you have automated workflows and a fast computer.

2.1 Equipment Myths

Another common belief is that photogrammetry requires expensive cameras. While a DSLR with a wide-angle lens and external flash works well, many teams get good results with a GoPro or even a smartphone in good light. The key is consistent settings: fixed focus, fixed exposure, and overlapping images by at least 60%. Manual survey equipment is simpler and cheaper—a Suunto compass and clinometer, a fiberglass tape, and a notebook. Both methods have affordable entry points, but photogrammetry's hidden cost is the computer hardware and software licenses.

2.2 Skill Requirements

Manual survey is easy to learn but hard to master. The basic shot-taking takes an hour to teach, but consistent technique (keeping the tape straight, reading the compass correctly, avoiding metal interference) takes practice. Photogrammetry has a steeper learning curve: understanding camera settings, lighting, image overlap, and processing parameters. Many teams abandon photogrammetry after a few failed attempts because they didn't realize how critical good images are. We recommend starting with manual survey, then adding photogrammetry once you have solid field skills.

3. Patterns That Usually Work

Over years of reading trip reports and talking to survey teams, certain patterns emerge for successful cave mapping. These aren't rigid rules, but they increase your chances of getting usable data without rework.

Hybrid approach for complex caves. The most efficient pattern is to run a manual survey backbone—shots every 5–10 meters through the main passage—and then use photogrammetry for key features: large chambers, delicate formations, or areas with complex geometry. The manual survey provides the overall geometry and closure control; the photogrammetry adds detail where it matters. This pattern reduces the risk of losing orientation in the photogrammetry model and gives you a fallback if the images fail.

Standardize your manual survey workflow. Use a consistent station naming convention (e.g., A1, A2, B1…), record all data in a waterproof notebook, and take redundant measurements (back sights, duplicate shots on long legs). Many teams use a tablet with a survey app like TopoDroid or DistoX2 for digital data collection, which reduces transcription errors and speeds up processing. The key is to have a clear protocol that every team member follows.

Photogrammetry with scale bars and control points. Place physical scale bars (e.g., a 1-meter rod with marked ends) in the scene, and measure their positions with a tape or laser. Also, set up control points (e.g., reflective targets) whose coordinates you measure with the manual survey. This ties the photogrammetry model to real-world coordinates and ensures accuracy. Without control points, the model may be beautifully detailed but distorted in scale or orientation.

3.1 Recommended Workflow for Hybrid Projects

1. Plan the survey: decide which sections will be manual, which photogrammetry, and where control points will go.
2. Run the manual survey first through the entire cave, including control points at transition zones.
3. Photograph the photogrammetry sections, ensuring good overlap and consistent lighting.
4. Process the manual survey data to create a base map with known coordinates.
5. Process the photogrammetry images, using the control points to georeference the model.
6. Merge the two datasets in a 3D modeling or GIS environment.

3.2 Team Composition

For manual survey, a team of three works best: one person reads the compass/clino, one holds the tape and records, one takes notes and manages the station. For photogrammetry, a team of three to four: one photographer, one lighting person, one person placing scale bars and recording control points. Cross-train everyone so you can switch roles if someone gets tired or cold.

4. Anti-Patterns and Why Teams Revert

Some approaches look good on paper but fail in practice. Recognizing these anti-patterns can save you from wasted trips and frustrating data.

Full photogrammetry of a tight, muddy cave. We've seen teams try to photograph every inch of a narrow, muddy crawl. The images come out blurry, the lighting is uneven, and the processing software fails to align them. They end up with a fragmentary model and no manual survey to fall back on. The result: a wasted trip and no map. The anti-pattern is assuming photogrammetry works everywhere. It doesn't. If you can't get clear, well-lit images with good overlap, don't rely on it.

Over-reliance on automatic processing. Some teams take thousands of images, dump them into software, and expect a perfect model. They skip scale bars, don't check image quality, and ignore processing errors. The result is a model with holes, distortions, and unknown scale. They then spend hours trying to fix it in post-processing, often giving up and reverting to manual survey. The lesson: photogrammetry requires careful field work and manual quality checks at every stage.

Manual survey without closure checks. Another common mistake is taking shots without back sights or loop closures. The data accumulates errors, and by the end of the trip, the map is off by meters. Teams that skip closure checks often have to re-survey sections, which is demoralizing. Always close loops and check your error budget. A closure error of 1% is acceptable for most exploratory mapping; more than that requires re-shooting.

4.1 Why Teams Revert to Manual Survey

Many teams try photogrammetry, hit these problems, and go back to manual survey for the next project. The reasons are practical: manual survey gives you immediate feedback, doesn't require a computer, and works in any passage size. The data is easier to share and combine with other surveys. Photogrammetry is powerful but fragile—it fails if the conditions aren't perfect. For most cave mapping, manual survey is the reliable workhorse.

4.2 When Hybrid Fails

Even the hybrid approach can fail if the control points aren't measured accurately. If your manual survey has a 2% error and you tie the photogrammetry to it, the photogrammetry inherits that error. The solution is to run the manual survey with high precision in the control areas, using multiple redundant shots and a laser rangefinder if possible.

5. Maintenance, Drift, or Long-Term Costs

Cave maps are often used for years—for trip planning, scientific study, or conservation. The long-term costs of each workflow differ significantly.

Manual survey data is easy to maintain. The raw data (shot list) is a simple text file or spreadsheet. It can be re-processed with updated software, combined with new surveys, and corrected if errors are found. The map can be updated by adding new stations. The long-term cost is low: just store the data in a standard format (e.g., Survex .svx or Compass .dat) and keep backup copies.

Photogrammetry data is bulky and software-dependent. The raw images, point clouds, and meshes can take gigabytes per project. Processing software changes every few years, and old files may not open in new versions. The model may drift if you don't have stable control points—over time, the coordinate system can shift if you re-process with different settings. Maintaining a photogrammetric model requires active management: archive the raw images, the processing project file, and the exported model in a standard format (e.g., .obj or .ply). Even then, you may need to keep an old computer running the original software.

Drift in manual survey data. Manual survey data can drift if the original measurements were not corrected for magnetic declination or if the tape was not tensioned consistently. Over decades, these small errors accumulate. However, they can be corrected by re-surveying key sections or by comparing with GPS data at the entrance. The data is transparent—you can trace every shot and see where errors come from.

5.1 Cost Comparison

Equipment costs: manual survey gear (compass, clinometer, tape, notebook) costs around $200–$500. Photogrammetry gear (camera, lenses, flashes, tripod, scale bars) costs $1,000–$5,000, plus software licenses ($180–$3,500/year). Computer hardware for photogrammetry processing adds another $1,500–$4,000. The total cost of ownership for photogrammetry is 5–10 times higher than manual survey.

Time costs: a typical manual survey of a 1 km cave passage takes 10–20 hours underground and 5–10 hours at the computer. Photogrammetry of the same passage (if conditions allow) might take 5–10 hours underground but 50–100 hours of processing. The total time is often 3–5 times higher for photogrammetry.

6. When Not to Use This Approach

Every workflow has its limits. Knowing when to avoid a method is as important as knowing when to use it.

Don't use manual survey when you need high-resolution 3D data. If the goal is to document a rare formation, monitor erosion, or create a virtual tour, manual survey won't give you the detail. The station spacing (typically 1–10 meters) is too coarse. Use photogrammetry or laser scanning instead.

Don't use photogrammetry in wet or muddy conditions. Water droplets on the lens, mud on the walls, and fog from your breath ruin images. Even with careful cleaning, the results are often poor. If the cave is actively dripping or has high humidity, stick to manual survey.

Don't use photogrammetry if you can't carry the gear. A full photogrammetry kit (camera, lenses, flashes, tripod, scale bars, batteries) can weigh 5–10 kg. In a tight crawl or a long approach, that weight is exhausting. Manual survey gear weighs less than 1 kg. If the cave is physically demanding, choose the lighter option.

Don't use photogrammetry if you need results immediately. Manual survey gives you a line plot within hours of leaving the cave. Photogrammetry takes days or weeks. If you need a map for the next trip, manual survey is the only practical choice.

6.1 Scenarios Where Both Fail

In extremely large chambers (over 100 meters across), both methods struggle. Manual survey requires long tapes or laser rangefinders, and the shots become less accurate over distance. Photogrammetry requires multiple overlapping images from different positions, which is difficult to set up without a team of photographers. In these cases, a combination of laser scanning (LiDAR) and photogrammetry is often used, but that's a different workflow beyond the scope of this guide.

6.2 Safety Considerations

This guide provides general information only. Cave mapping involves inherent risks, including falling, hypothermia, and getting lost. Always follow established safety protocols, carry adequate gear, and never map alone. Consult with experienced cavers or your local caving organization for training and best practices.

7. Open Questions / FAQ

We often hear the same questions from cavers trying to decide between methods. Here are our honest answers, based on common experiences.

Q: Can I use photogrammetry with just a smartphone? A: Yes, but results vary. Modern smartphones with good cameras (e.g., recent iPhones or Pixels) can produce usable models in good light. The main limitations are battery life, lens quality, and lack of manual control over focus and exposure. For serious mapping, a dedicated camera is still better.

Q: How accurate is manual survey compared to photogrammetry? A: With careful technique, manual survey can achieve 1–2% closure error over a loop. Photogrammetry with good control points can achieve centimeter-level accuracy over tens of meters. But accuracy depends more on the surveyor than the method. A sloppy manual survey is worse than a careful photogrammetry project, and vice versa.

Q: Do I need to learn both methods? A: Not necessarily. If you only map small, tight caves, manual survey is sufficient. If you work in large, open caves and need detailed models, photogrammetry is valuable. Many teams learn manual survey first, then add photogrammetry for specific projects. We recommend starting with manual survey—it teaches you the fundamentals of cave geometry and measurement.

Q: How do I combine data from both methods? A: Use common control points. Place physical targets or natural features that you measure with both methods. Then import the manual survey data into your photogrammetry software as ground control points, or export the photogrammetry model and align it to the manual survey in a 3D program. Tools like CloudCompare or Meshlab can help with alignment.

Q: What's the biggest mistake beginners make with photogrammetry? A: Taking too few images with poor overlap. You need at least 60% overlap between adjacent images, and 80% is better. Beginners often take one image every few meters, then wonder why the software can't align them. Take more images than you think you need, and always check the alignment before leaving the cave.

8. Summary + Next Experiments

Choosing between manual survey and photogrammetry is about matching the method to the cave, the team, and the goal. Manual survey is reliable, simple, and works in almost any passage. Photogrammetry offers high detail but demands good conditions, heavy gear, and significant processing time. The hybrid approach often gives the best of both worlds, but requires careful planning and control points.

If you're new to cave mapping, start with manual survey. Learn to take consistent shots, close loops, and draft a map. Once you're comfortable, try photogrammetry on a simple, dry chamber with good lighting. Compare the results with your manual survey to see where each method shines. Over time, you'll develop a sense for which workflow fits each project.

Here are three specific next moves:

1. Run a test survey in a local cave you know well. Map a 50-meter section with manual survey, then map the same section with photogrammetry. Compare the time, effort, and final map quality. This will teach you more than any guide.
2. Join a survey trip with an experienced team. Watch how they set up stations, handle errors, and process data. Most cavers are happy to share knowledge—ask questions and take notes.
3. Experiment with hybrid control points on your next project. Place a few reflective targets or marked rocks, measure their positions with a manual survey, and use them to georeference a photogrammetry model. See how the alignment works and note any issues.

Cave mapping is a skill that improves with practice. Both manual survey and photogrammetry have their place, and the best mappers know when to use each. Start with what you have, learn from each trip, and don't be afraid to try new methods. The cave will always be there for another attempt.