Quad-Core Technology Speeds Up Machine Vision

The new generation of quad-core processors opens a new round of CPU architecture. They again drastically increase data throughput, provided that the software can make use of this luxury. HALCON 8.0, MVTec's new version of the leading machine vision software library, is already able to automatically use quad-core parallelization for a dramatically increased performance.

Since June 2007, HALCON 8.0 is on the market. This new version has been especially optimized for an improved speed-up by the automatic parallelization (Parallel HALCON) of image data. With regard to performance, this automatic parallelization by far can keep up with manual multi-threading. Even experts cannot manually implement parallelization much faster, which makes HALCON's automatic parallelization a good choice for every case.

This is particularly true for applications that must handle large amounts of image data (such as with color or other multi-channel images, 16 or more bit images, stereo images, Fourier transform, texture analysis, surface inspection, or blob analysis). Here, the splitting to several processors is indispensable. To split image data and then merge them again is a complex operation which is automatically optimized by HALCON 8.0. Especially for often used operators, such as filters for preprocessing, the automatic parallelization now has become significantly faster with HALCON 8.0.

About HALCON:
HALCON is the comprehensive standard software library with the IDE for machine vision. HALCON provides an outstanding performance for blob analysis, morphology, pattern matching, measuring, 3D object recognition, and binocular stereo vision, to name just a few. HALCON offers more than 1300 operators and provides interfaces to hundreds of industrial image acquisition devices. Furthermore, HALCON is platform-independent and compatible to Windows, Linux, and Solaris.

SOURCE: HALCON