Smart3DCapture takes as input a series of digital photographs of a static scene taken from distinct viewpoints, together with available camera calibration, position and orientation parameters, and automatically turns this dataset into a high resolution and textured manifold triangular mesh that can be exported in various 3D formats.
Smart3DCapture’s versatility seamlessly reconstructs subjects of varying scales, ranging from centimetres to kilometres, photographed from the ground or from the air, using any type of camera, on a decent Windows® based computer.
By exploiting the power of general purpose computation on graphics processing unitsSmart3DCapture can process 7 to 11 gigapixels per day. For aerial datasets with a ground resolution of 10-15cm and typical image overlap one Smart3DCapture engine can process 4-6 sq km per day.
Scalability & Grid Computing
Smart3DCapture’s highly scalable solutions allow distribution of the 3D generation process over multiple Smart3DCapture engines to dramatically reduce processing time.
Smart3DCapture pushes accuracy to the limits allowed by general photogrammetry principles. For photographs with correct overlap and sufficient baseline, the typical geometric accuracy is 1-2 times the projected pixel size.