Community Research and Development Information Service - CORDIS

Advanced algorithm for CAD/CAM vectorisations

A Russian organisation has developed an advanced algorithm that can support present and future requirements within the field of software drawing technology.
Advanced algorithm for CAD/CAM vectorisations
Raster-to-vector is whereby an intense magnitude of parallel lines (that combine together to form an image), are converted into a software data quantity. The magnitude and direction of this data enables software drawing programs to automatically recognise the image in a vector format. In essence this means that an image such as technical, architectural, mechanical and other drawings like maps, can be scanned and imported into software drawing programs like CAD.

Much like many other forms of software there have been many approaches to optimisation of software vectorisation. With the evolution of the computer supported industries of the last 20 years, the demands and capabilities imposed on vectorisation applications have begun to separate. As such it is expected that only a few programs will be able to conform to future requirements; and one such application has been developed by a Russian organisation.

The concept behind this particular raster-to-vector application is based on precision reconstruction, topological consistency with built-in connectivity information and the allowance for compact storage. To facilitate this, the designers had to develop an efficient algorithm that would be able to scan and analyse every single line of the image in just one single scan.

Unlike conventional formats where image area is critical, the developed algorithm is able to provide constant memory for internal structures that are dependent on image width. It also enables scanning without whole images being stored in the memory, and can exploit integer arithmetic only. In addition and depending on its specific application, this engine can gather further essential information and filter out minor details.

For efficiency and correctional purposes, this raster-to-vector conversion engine has been thoroughly tested on numerous images. As a result it has been shown to exhibit excellent efficiency in a multitude of applications including CAD, CAM, CIM OCR and GIS systems. The engine is implemented is a software library for windows, is available in C and its technology is core to raster-to-vector conversion engines. The Russian organisation responsible for this innovative advancement, are currently looking for licence agreements.

Related information

Record Number: 80930 / Last updated on: 2005-09-18
Domain: IT, Telecommunications