 |
 |
IDE with Improved GUI for
Best Usability |
|
|
 |
HALCON's unique IDE (integrated development environment) for
machine vision, HDevelop and HDevEngine, has been radically
improved in design and functionality. The HDevelop GUI has been
modernized and now is available in many common languages.
Furthermore, HDevelop provides new assistants to greatly
simplify the generation of frequently used
operator sequences.
HDevelop now with syntax highlighting,
integrated online help, and many more new features.
|
|
|
|
3D Object Recognition |
|
|
 |
|
You need no specialized imaging software to enter the world of 3D,
e.g., for bin picking and other robot tasks. With HALCON 8.0,
3D matching now can be done with a machine vision standard
software: 3D matching finds objects in arbitrary 3D poses based
on a CAD model with the same power and robustness as HALCON's
well-known shape-based matching. To identify an object from its
CAD model, only one image taken by a single camera is necessary.
CAD models can be read in DXF format.
|
|
|
|
Shape-based Matching in
Color Images, Correlation-based Matching |
|
|
 |
|
HALCON's robust methods for matching have been further improved.
Shape-based matching now also handles objects with different
scaling in row and column as well as objects that are partially
located outside of the image. Further, it can process images
with 8 or 16 bits and handles also color or multi-channel images.
Correlation-based matching is particularly robust against
defocus, shape deformation, and texture.
|
|
|
|
A Library Faster than the
State of the Art |
|
|
 |
|
The speed of the entire HALCON software library has been
significantly enhanced, on average by 25 %, single operators
by up to 500% for typical applications. This speedup is achieved
by an advanced memory management as well as an extended use of
MMX and SSE2, and many more improvements of the HALCON core.
Also the performance on multi-core processors has been
substantially improved by optimizing Parallel HALCON 8.0.
|
|
|
|
Easy-to-Use OCR with
Auto-Correction |
|
|
 |
HALCON's powerful classifiers have been enhanced and optimized
for usability. They lead to highest recognition rates
"out of the box", such as 0.65 % error rate on
the image set of MNIST standard. Additionally, HALCON provides
a syntactic and lexicon-based auto-correction.
Make OCR more robust by regular
expressions, e.g., by forcing five numbers and one letter
in a row.
|
|
|
|
Further New Features |
|
|
|
- Native HALCON/.NET interface
(also for Mono under Linux/UNIX)
- HDevEngine: native support of .NET
- 3D rectangle pose with one camera
- Data code print quality inspection
- New classifiers: Support Vector Machines and
Gaussian Mixture Models
- Easy text segmentation and text character
selection for OCR
- Extended and improved documentation, now also
with new Solution Guide
- Full-text search in the complete HALCON
documentation
- Rectangle measurement
- Enhanced variation model
- Improved optical flow
- And many more...
|
|
|
|
