When Photogrammetry Is the Wrong Tool

Photogrammetry has many use cases and is being used more and more across different industries. However, like all things, the photogrammetric process is not all-powerful and does have limitations. In this blog, I usually explore new areas where photogrammetry might be useful and explain why that is the case. This time, however, I will share my experience with situations where photogrammetry is downright impossible, and where it may not be the preferred method for 3D scanning.

Moving and Unstable Surfaces – The Fundamental Limitation

Photogrammetry fundamentally assumes that the scene does not change between overlapping images. Once something moves, even slightly, the reconstruction starts to break. Only the camera can, and must, move. This includes obvious cases like people, cars, or vegetation in the wind, but also less obvious ones such as reflections on water or glass. Reflections change with every angle, which makes them behave like moving geometry and therefore unsuitable for scanning.

I tried scanning a person who was attempting to remain completely still in ideal conditions, and even that produced subpar results.

Read more: Human Photogrammetry with Single Camera

At the extreme, this is why you cannot realistically scan an ocean surface or a mirror. Even if it looks stable to the eye, it is not stable to the algorithm. In practical terms, this affects use cases such as coastal mapping, urban scenes with glass buildings, or construction sites with moving machinery. Some parts of the scene may reconstruct, but unstable areas will introduce noise or completely incorrect geometry. A commonly overlooked example is treetops. Trees may seem stable, but leaves create movement even in barely noticeable wind. This is why forest photogrammetry can be difficult.

Practical Limitations and Edge Cases

Low Texture Surfaces

Photogrammetry relies on identifying and matching visual features. Surfaces that lack texture, such as white walls, new asphalt, snow, or uniform roofs, do not provide enough information for reliable matching. Even with high overlap, the software struggles to anchor points in space.

This becomes a problem in use cases such as building facades with clean finishes, warehouse interiors, or agricultural fields at certain stages. The result is often incomplete or unstable reconstruction, especially in large uniform areas. This is likely the most common limitation after the fundamental one.

Thin and Complex Geometry

Very thin objects and highly detailed structures are difficult to reconstruct accurately. Wires, fences, antennas, and small branches often disappear or become distorted. This is due to both image resolution limits and smoothing during reconstruction.

This limitation is important in infrastructure inspection, power line mapping, or vegetation analysis. Even if these elements are visible in images, they are rarely reliable in the final model. These types of objects can also be problematic for other 3D scanning methods such as LiDAR, although usually to a lesser extent.

Lighting and Dynamic Range

Lighting conditions directly affect image quality and feature detection. Low light introduces noise, while strong shadows and high contrast reduce usable detail. Overexposed or underexposed areas lose information entirely. Using modern high-quality cameras, such as the Sony A7V, can help mitigate this issue for terrestrial scanning, while drones like the DJI Mavic 4 Pro offer strong aerial image quality.

The requirement for high-quality input data impacts indoor scanning without proper lighting, night-time capture, or scenes with harsh sunlight and deep shadows. Even with good planning, results can be inconsistent. When lighting can be controlled, it adds complexity, but well-designed lighting setups can resolve many of these issues.

Scale and Accuracy Requirements

Photogrammetry can produce visually convincing models that are slightly incorrect in scale or geometry. Without proper reference data, such as control points or known distances, small errors can propagate across the entire model.

Read more: Relative vs Absolute Accuracy in Photogrammetry

This becomes critical in surveying, engineering, or any workflow where measurements must be precise. A model that looks correct is not necessarily accurate enough for decision making. With proper referencing, accuracy can be improved significantly, but this introduces additional setup and complexity.

When Photogrammetry Becomes Inconvenient

Interiors and Tight Spaces

Photogrammetry can work indoors, but it quickly becomes inefficient. Limited space, lack of texture, and complex lighting conditions make capture more difficult. You often need many images, careful movement, and additional lighting to achieve usable results.

In these scenarios, handheld scanning or laser-based methods are often more straightforward and faster to deploy. Industrial environments and factory spaces are good examples, especially when BIM models are required and surfaces may be uniform or reflective.

Repetitive or Time-Critical Work

Photogrammetry requires planning, capture, and processing time. If the same type of object or environment needs to be scanned repeatedly, or if results are required quickly, this workflow can become a bottleneck.

Examples include routine inspections, industrial workflows, or rapid documentation tasks. In such cases, faster but less flexible methods may be more practical. If a quick decision is required, photogrammetry may not be the most efficient option. Dedicated 3D scanning equipment can provide near real-time results in these scenarios.

Quick High Precision Workflows

When very high precision is required, photogrammetry often needs additional setup, such as ground control points or careful camera and lens calibration. This increases complexity and reduces speed.

In surveying or engineering projects, laser scanners or total stations can provide more direct and predictable accuracy with less iteration. They also offer real-time feedback during capture. This comes at a higher equipment cost, but for large-scale or high-value work, the trade-off is often justified.

Conclusion

Photogrammetry works best in stable, well-textured, and well-lit environments. Outside of those conditions, results can become less reliable or the workflow less efficient. Understanding these limitations is not about avoiding photogrammetry, but about knowing when it will perform well and when another tool will produce better results.

Photogrammetry remains one of the most accessible and cost-effective 3D scanning methods. When used in the right conditions and with the correct workflow, it can deliver highly competitive results without requiring specialized equipment.