CVB Foundation PKG
The complete easy-entry vision software suite
The easy, powerful, economic package
CVB offers Foundation Package to give a convenient entry to machine vision world for new users. This is a powerful package based on CVB image Manager with optimized image display which can control the various image processing hardware. Additionally, an abstract image access via user-defining algorithm is possible. All process can be complemented by optimized Edge, Blob, statical image analysis, filter and arithmetic. This comprehensive set provides users who deal with diverse machine vision tasks with the most important and widely used algorithm.
CVB Foundation Package also includes the core machine vision technology such as a new algorithm for 2D Calibration used in measurements or robot application. In addition, Foundation Package has dynamic thresholding used for images with the frequent brightness variation. other than that, special functions for Colour conversion and image filtering are contained.
Graphic program interface, iTuition is been added to this package. iTuition is the way to build image process application with a easy drag&drop method. Therefore, there is no need to buy a special compiler for production of application. Graphics will be very useful for all beginners who have no idea about programming or skilled programer too.
All high level language compilers such as Borland(C++Builder/Delphi), Microsoft(Visual Basic, Visual C++, C#.net) are usable and an enormous sample program sources are offered.
Function for filtering of images
high optimized filter algorithms’ comprehensive union for high speed preprocessing of images. The use of filters highlights or relieves the images to simplify or accelerate the readout of next level.
Function for binarization using dynamic thresholding
Dynamic Tresholding simplifies image process under the illumination changes. The images change into pure black&white through this progress.
Dynamic Tresholding use local threshold with auto-update fuction in case of changes in parts of lightings.(Constant Tresholding, Image Tresholding available)
Function for 2D-Calibration of image data
This revises the optic distortion by lens disportion or the image distortion by a non-vertical glancing angle. This algorithm produce the revised image date for other image readout after calibration. The easiest way of image change of rounded surface.
Functions for colour space conversion
Changing the original image into other forms of color system according to the image information or inspections tasks can be very useful in the next process. There are the optimized algorithms with different attitude to separate brightness, color, chroma of images.
Function for mathematical calculation
CVB Foundation Package includes groups of functions called “Math Helpers.” These functions offers easy calculation in 2D vector space such as angle, vector, line, and circle. User can easily use a normal, vector, intersection and regression.
Additional tools available include
CVB Foundation Package provides the group of comprehensive and optimized image process functions to users. This package enables to develop machine vision application of a far-reaching area with minimal investment.
These tools help users who have difficulty in use of machine vision application. Recognition of an object and reading letters are the special examples.
Now, CVB Minos, CVB Manto, CVB Shape Finder are the leading software tool in terms of recognition speed and stability. CVB Manto realize the application that had no industrial solution in the past.(recognition of hand written letters, punched letters on glittered surface, food industry etc.) CVB Barcode tool can be used to read 1D and 2D barcode. Almost all forms of barcode symbol and data matrix code can be recognized and location of the recognized codes will be provided. They focused on a rapid and stable code recognition even about the bad conditioned code of reality and industrial site. CVB Movie and CVB Sequence as recording tool can create perfect sequence recording system. On CVB’s hardware supports and modular concepts, users can have flexible and customizable recording solution.
CVB MANTO
Very Special pattern matching tool
- CVB’s special tool for vision application that can be solved by general machine vision tools
- application to the changing shape of objects
- application to the changing illumination of inspections objects due to unstable light
- application that needs to detect subtle color changes
- applying SVM(Support Vector Machine) algorithm to letter recognition(hand written), food inspection, surface inspection, face recognition
CVB MINOS
CVB Minos is new pattern matching tool. It creates classifier gathering shape-describing features from various images and then performs patter matching based the classifier. So, this is faster and has high reliability rather than the existing method to compare every single pixel. The algorithm is strong at the illuminations changes so that it’s not affected by noise interfere. One classifier can hold various patterns thereby they can be found at the same time. This is the precondition for efficient OCR and gives considerable flexibility to many different applications. Minos has a similar way to neural network logic. Classifier recognize Negative sample as well as positive sample from learning process. Negative sample is operated against positive one so that it remarkably reduces wrong recognition.
Minos provides special functions for OCR/OCV.
CVB SHAPE FINDER
Object searching tool
Shape Finder is a searching tool to find out the exact location of object rapidly regardless of turning orientation and size based on geometric attribute. One of the features is the high recognition rate about objects partially covered or with interfere such as noise and reflection. By creating many parameter, users can choose either extremely fast setting or much slower but accurate setting.
Algorithm
This tool basically uses Hough transform. Pattern is recognized in basis on image’s partial contrast. Contrast area focuses on the area where brightness changes, while the others are ignored. Shape Finder’s users can control the quality of efforts and results because they can define appropriate threshold value about contrast. Shape Finder also can find over one kind of patterns. In this case, only one searching is performed to find n patterns without n searching. Multidimensional result values will be calculated so that different pattern class can be recognized and the location can be determined. For this reason, Shape Finder can be used to OCR application as well.
Other important feature is the recognition without directions. This is a very important key point of many industrial application. Shape Finder’s algorithm structure has a very rapid searching velocity. In case of searching pattern with 30×30 pixel in 768×572 sized image, it only takes around 15ms.(depending processing power)
ShapeFinder features
- high recognition rate about overlapped pattern or noise
- available to register various pattern class
- recognition regardless of turning or size changes
- maximize either speed or accuracy on demand
- easy to use, simple solution to difficult problem
- high recognition rate about partially covered objects
- recognition regardless of turning or size changes
- fixity measurement about location, direction and scale
- available for created images rather than shooting images
- very fast processing by optimized searching technique
- easy to register
ShateFinder advantages
- registering one sample and searching location, direction, and scale precisely
- easy program interface
- attempting graphic interface for register, reliability and speed test
- simple button and options
- almost all applications can be used only by Default setting
- VB, NET source code
Application
- Pick & Place (Align) Application
- dequate in OCR