How can you capture high-quality photogrammetry data in the field?

Nowadays, achieving this is possible with a phone equipped with a good camera. Some phones, starting from 2020, can be really effective for this purpose. Consider getting a turntable, which you can find on Amazon for around 25 bucks; it can be very helpful to start. Additionally, using a tripod is beneficial. To ensure quality, avoid shaking and take small steps after every photo. For the first pass, capture shots from the outside by taking a photo, then a small step, repeating until you complete a circle. Afterward, follow the same process for capturing more detailed shots.

Software like Pix4d can even use 360 photos to produce models bringing everyday high quality camera photogrammetry into the BIM world on control.

Not only a good camera but a camera with lidar is key to good reliable models. I'm not an "iPhone guy" by any means but the newer iPhone have lidar and that makes it waaaay more accessible for people to get accurate models made up with next to no work.

LiDAR on iPhone provides an excellent way to capture the 3D structure of the scene for objects in a room at a distance of 2-3 meters in a single frame, which I have captured, and they offer a good resolution. However, the resolution is still insufficient for closer objects like human faces. An iPhone equipped with a polarizer could work in this case, but I have yet to experiment on this one.

Use a stable platform for capturing images, whether it's a tripod for ground-based photography or a well-calibrated gimbal on a drone. Stability minimizes blur and ensures clear, sharp images. Be mindful of weather conditions. Overcast skies can provide diffuse lighting, reducing harsh shadows, while avoiding extreme brightness or glare is crucial for maintaining image quality. Remove unnecessary objects or clutter from the scene that could obstruct the view or create confusing details in the model. Keep the area as clean and unobstructed as possible.

Winter can provide helpful diffuse light. Avoid going to extremely exposed or overly dark environments, and steer clear of harsh highlights and shadows. When shooting, look for objects you can navigate around to capture better details of your model and to avoid distortion.

For photographing the object, it is necessary to count on a time frame in which the weather conditions will not change. Cloudy weather with good visibility and not marked by the direction of lighting is suitable.

Planning your shooting strategy is vital for successful photogrammetry. Determine the required number of photos, consider optimal angles and distances, and address lighting and weather challenges. Thoroughly scout the location for potential obstacles, reflections, or shadows that could impact photo quality. A recommended approach involves capturing a minimum of 60 photos per object or scene, ensuring a 70% overlap between photos, and varying camera height and angles to enhance data accuracy and reconstruction.

It helps to have the plan going to take any pictures of any equipment or items to kind of have a step-by-step storyboard plan of what you're going to be doing what you're trying to cash her and the reasoning for cashing it it's not just the angle of the cat of the camera there's more to it you have to make some choices you have to look at what you're doing to come up with the best possible outcome sometimes that'll mean going back to the office and actually looking through the pictures and making some adjustments cleaning them up changing the angle maybe clipping things of that nature oh no you can't come up with a very good and worthwhile instead of photos for the object or subject that you are trying to capture

Take the necessary time to capture the photos, and look for a well-lit scenario; it can save you a lot of time in the future. When scanning objects, choose a time when it's not too busy, especially if it's near a public place. For natural elements, try to avoid shadows created by surrounding objects.

A consistent exposure focusing to the object, check your photos as you go and make sure that they are clear, well-aligned and take a shot.

Iterate on your taken photos; take the time to try multiple shots, if possible, in different scenarios. This will help in testing which lighting suits your model best. Make sure not to move the object while capturing the pictures.

The exposure setting should be manual and constant, adapted to the average lighting value. Due to the sharpness of the photos, the depth of field should be deep (small aperture) and photos that are taken handheld, without image stabilizer, should have shorter exposures, more than the angle of the lens used to take the photo. For example, when using a 50mm full frame equivalent lens, it is desirable that the exposure is not slower than 1/60s.

Substance Sampler provides excellent services for converting models, along with extensive documentation. In comparison to pricey applications, Substance Sampler offers licenses for learning purposes, making it a cost-effective option. Export it depending on what you want to do. the pipeline can vary depending on your constrains.

3D scannings should be cleaned by removing any loose parts, filling all the remaining holes and smoothing any of the irregular surfaces caused by error that occurred during the scanning process, the model must be rotatet until it achieves the desired orientation, then scaled to match the required size. When the model is clean and correctly oriented, then it is ready to be modeled, textured, illuminated, etc. If necessary, it is possible to change the model polygon count to achieve the file size and model quality for every specific application.

Photogrammetry without LIDAR could have issue with it's on point accuracy, especially in measurement. I wouldn't recommend any normal photogrammetry, especially if it is just raster images to do accurate BIM modelling, since photogrammetry for 3D with raster images normally involves only matching each vertexes to respective "Similar" raster point in the images.