Photogrammetry is a 3D “Reconstruction” tool. It allows you to build a model based on an object or area, rather than the other way around.
The basic principle of photogrammetry is the stitching of several images of the same object taken from multiple angles, in order to create a 3D wireframe model covered with photo realistic textures.
The purpose of the software and its algorithms is to follow the pixels and their relative position from one image to the next, to deduct a coordinate in a three-dimensional space. All these points when connected to each other allow for the tracing of the wire structure over which will be layered the photographic information that will make up the texture of the surfaces.
The quality of the final composite directly depends on the quality of the shots, and therefore, it is important to follow a strict methodology and to use the right tools.
For the camera, the two most important specifications are the resolution and the field of view. The higher the resolution, the greater the megapixel count, the more precise the final rendering will be. This will allow closer examination of the model without revealing too many rendering flaws. As for the field of view, a wide angle is to be preferred, provided that the optical distortions are minimal or fully corrected before the images are ingested into the compositing software.
Obviously, great hardware is not enough; The pictures must be taken following two basic rules.
First, it is important to shoot all the photos within the shortest possible time span, and in the brightest, most constant lighting conditions possible. The more homogeneous and even the lighting conditions are, the easier it will be for the software to organize and compare the pixels, as their brightness and color will vary very little from one photograph to the other.
Second, and perhaps most importantly, coverage must be exhaustive. Many shots must be taken from many different angles. Not only should a complete rotation around the object be divided into an acceptable number of shots, but there must be several rotations spread across the entire object’s length.
By diligently following this methodology, the analysis software will have all the information it needs to render a high-quality model.
Comparing photogrammetry and laser scanning is a bit like comparing aerial imagery by drone and helicopter: Each method has its advantages and disadvantages, and therefore, they are not mutually exclusive.
The main advantage of photogrammetry is its photographic rendering of textures. While laser scanning mainly sees in terms of volumes via distance calculations, photogrammetry sees in terms of colors and brightness and deduces the volume by algorithms that analyze variations from one image to the next, to detect corners and therefore each of the points of the wire structure over which the photographic texture will be layered.
Photogrammetry is particularly effective for modelling small objects or structures of moderate size. Since the necessary equipment costs much less, it quickly becomes much more profitable to choose this solution when the scale of the project fits and when photo realism is more important than extreme precision.
While the evolution of laser scanning is strictly hardware-based, the evolution of photogrammetry is mostly software-based. Photogrammetry equipment can thus be kept as is, while the processing programs continue to improve in precision and performance.
Photogrammetry is a modelling method that differs from traditional laser scanning. Photogrammetry will be preferred for mainly two reasons: When photorealism and cost are more important than the absolute accuracy of the model or when modeling moderately sized objects or areas.
For example, archaeologists prefer photogrammetric rendering for photo realistic reproductions of artifacts for a fraction of the cost of laser scanning. Likewise, the film and video game industries prefer this technology for rendering sufficiently credible environments in the context of a film or game.
Even in the mining industry, the accuracy of a photogrammetric model is often considered sufficient to assess volumes of ore, with a margin of error that is 2% less precise than the laser scanning one.
The Mavic Pro is one of the lightest, smallest and sturdiest drones on the market. In terms of image quality and maximum flight time, it is similar to the Phantom, but does not allow such complex or varied flight programs. For video capture, it's one of the best choices out there, but for more sophisticated applications like geomatics we will prefer the Phantom 4 Pro or the Phantom 4 RTK.
The little brother of the Inspire is a lighter drone, with better autonomy that retains a very high image quality. For all your needs of aerial imagery, the quality of the content produced by this drone is suitable for almost any commercial project.
The most advanced UAV in our fleet. It offers the best picture quality and unparalleled flexibility in terms of lenses you can use and the field of view they can provide. If you're looking for the sharpest images, the richest colors, the biggest flexibility in your composition and framing, and a decisively cinematic look, the Inspire Pro DJI will marvel you with the quality and accuracy of the images it produces.