||AgiSoft PhotoScan – allows you to automatically create high-quality 3D models of objects from digital photos. For the reconstruction of the object in PhotoScan enough upload photos, no additional information is required. Restored textured 3D-model can be saved in various formats – VRML, Wavefront OBJ, 3ds, Stanford PLY, COLLADA, PDF, U3D. PhotoScan capable of handling any photos taken with any digital camera, from any angle. The main thing is that each element of the reconstructed object was seen with at least two positions shooting. If the program enter at least one distance between the object and shooting positions, PhotoScan restores the scale of the entire model and allows you to define the distance between any two points of the object and calculate the area and volume of an object or its parts.
If the program to enter the coordinates of at least three points of an object or 3 positions shooting, PhotoScan automatically binds the model to a given coordinate system, allows us to calculate the projection of the model on a given surface (orthophoto), DEM with respect to a given surface (DEM), DEM and orthophoto display in various formats and coordinate systems.
Image processing is fully automated and does not require prior calibration or manual labeling of photos.
How does PhotoScan:
As a rule, the purpose of the users of the program is to restore PhotoScan textured 3D model of the object. This object is achieved by the passage of three processing steps:
– Definition of terms and parameters of external and internal orientation of the cameras. At this stage PhotoScan find common points of the photos and it defines all the parameters of the camera: position (up to scale), orientation, internal geometry (focal length, distortion parameters, etc.). The results of this phase are sparse cloud of points in common in 3D models and data on the position and orientation of the camera. In PhotoScan cloud point is not used in the later stages of processing, and serves only for the visual alignment quality evaluation images. However, the point cloud can be exported for further use in external programs. Same set of data on the position and orientation of the cameras used in the later stages of processing.
– Restoration of the geometry of the object. In the second phase builds a three-dimensional PhotoScan polygonal model that describes the shape of the object using one of the four algorithms provided: Quick, Tie, Smooth and the method of the height map. After building the model is sometimes necessary to edit it. Some changes, such
as optimization model, deleting a model component isolated, filling holes etc. can be made directly in PhotoScan. For more complex editing operations can be exported polygonal model being edited in an external editor, and then import the edited model is already back.
– Texturing object. The last step involves texturing and / or construction of orthophoto. In PhotoScan are several methods for constructing texture model in more detail are described in the appropriate section of the manual.
The main features
• Air and near triangulation
• Generation of point cloud (rare / tight)
• Polygonal generation model (plain / textured)
• Setting the coordinate system
• Generation of a digital elevation model (DEM)
• The actual generation of ortho – photo
• Linking with the flight log and / or reference points
• Multi – spectral image processing
• 4D reconstruction of dynamic scenes
• Support for Python Scripting