The scientific objectives of the network are to investigate:
– methods for construction of finite and unbounded (ie. Without the natural closure of individual objects) CAD models of built environments including both interiors and exteriors.
– methods for extraction and incremental fusion of typical environmental features (eg. pipes, walls, floors).
– how to fuse sparse, but continuously tracked shape, obtained from projective geometrical analysis of intensity images with dense shape obtained from single range images.
– how to plan for incremental sensing for extending and filling out models.
– how to exploit constraints typical of built environments (eg. systematic rectangularity and parallel features) or typical architectural structures (eg. doorways) to improve model accuracy.
– how to link locally accurate models (eg. of individual rooms) to form accurate global descriptions of the environment (eg. by exploiting constraints arising from two rooms being adjacent across a wall).
– how modules implementing the research results might be integrated together in a usable system.

The training objectives of the project are:
1 – develop cross-disciplinary skills in computer vision, computer graphics, computer-aided design, sensor navigation and fusion and photogrammetry.
2 – provide post-doctoral researchers with broader project perspectives, trans-national experience and experience in a commercial organisation.
3 – provide a forum for idea exchange and accelerated progress in a new and rapidly developing research area. Within CAMERA, ISR/IST main interests are on the areas of range data acquisition, 3D reconstruction based on range and/or visual data, integration of multisensory range information, perception planning and execution, control of data acquisition, camera tracking, reverse engineering.