We aim to provide high quality 3D models of real objects as a cost efective alternative to laser
range scanners and manually constructed models. The reconstructed models contain not only
accurate geometric data but also reflectance data which allows the customer/user to view and
re-illuminate the digitised objects in generic Virtual Reality applications.
Our software creates accurate surface models and high quality segmentation from volumetric data
- Sub-grid accuracy
- No need for help from expert users
- Resulting surface data can be loaded/converted into standard visualisation packages
- Shapes can be reproduced (e.g. using manufacturing techniques such as stereo lithography)
- Measurements (e.g. area, volume, distance) are readily available.
Our sophisticated software creates accurate Computer Graphics models from image data. Features of the new technology include:
- INPUT: multiple digital camera images of the scene/object (no specialised equipment required)
- OUTPUT: Computer Graphics formats such as VRML, STL, MPEG, etc. Output mode/format and level of detail can be
chosen to suit the particular application.
- Complete geometric model (can `walk around' object)
- Complete photometric colour model (user can change lighting)
- Performs well also on curved objects (sculptures etc.) including those with concavities/grooves
- Objects can be of arbitrary size
- Can handle multiple objects
- Output can be modified by the user to change shape and/or texture of the object or to
animate the obtained model.
- Current limitations: complete modelling of glass and semi-transparent objects
© Cambridge InnoVision Limited