|
|||||
     |
|
||||
Get a GoStats hit counter
|
|||||||||||
|
Copyright © 2003, by Concepts in Computing Get a GoStats hit counter |
Concepts in Computing, Inc. is a leading supplier of high speed Optical Character Recognition software for the industrial environment. Optical Character Recognition (OCR) is used in a wide variety of industries for identicfication and / or classification of process components. These may include labels, integrated circuits, stamped engine blocks and other automotive components. Currently man readable codes are required as supplements to barcodes in food processing, pharmaceuticals, electronics, automotive, bulk mailing and many more. These applications use a common library of routines, EconoCR™
An example of an application for OCR in the automotive industry would be the reading of Vehicle Identification Numbers (VIN).
EReader™
Program EReader™
Help File Standard
Font Files Test
Images MuTech
MV™510 Windows
9x™ MuTech
MV™510 Windows
NT™/2000™/XP™ OCR A
- multiple sizes OCR B Semi
Stroke and Dot Matrix 16MB
Zip File Variety
of fonts and backgrounds with single and multiple linesDesigned in Microsoft Visual C++ 6.0™, EconoCR™ has a long list of features specifically tailored to industrial OCR. Compare the list below to other comparable tools for robustness, flexibility, and optimal control.
For a list of available features, click here.
Visit the application pages to view how these and other features are used to solve a variety of OCR problems.
Dot
matrix text Uneven
spacing Uneven
character width Auto
character split OCR A
font Use of
fielding Maximum
character width Different
Threshold types Searching
for characters Trained
as patterns High
threshold cutoff OCR A
font Contamination
and noise 3 Line
label Overlapping
dot matrix Dot
matrix text Uneven
spacing Uneven
character width Auto
character split OCR A
font Use of
fielding Maximum
character width Different
Threshold types Searching
for characters Trained
as patterns High
threshold cutoff PowerPC™ Analog
Devices 21XX™ TI C60
series™ Hitachi
SH4™EconoCR™ is becoming the "de facto" standard for industrial OCR applications. Framegrabber or smart camera, EconoCR™ provides the same robust, character reading solution. Check out our long list of features and then compare it to any other industrial OCR product.
Find out why industry leaders such as DVT, National Instruments, Coreco and PPT Vision have chosen EconoCR™ as their OCR solution!
Check out the latest demo. Gradient data detection now included as part of the utility library. The EReader demonstrates how gradients can be used for a variety of OCR applications. Download the "Gradient Data Extraction" white paper for an in-depth discussion of this new feature. (New demo 5/15/2003)
Now included is the EReader™ volume 1 & 2 user's guide. Volume 1 is a basic over-view of the features of the EReader™ program and EconoCR™ library. Volume 2 over-views a variety of applications from the downloadable test suite. (New 7/8/2003)
This application reads the date and lot code printed on both sides of a juice carton. As with many high speed dot matrix printing applications, spacing between characters is variable and character width may vary slightly. This is in addition to the normal degradation observed in an application of this type. Unless characters can be reliably separated, accurate reading is not possible.
Date and lot codes using standard thresholding. Notice how the slant of the characters and spacing cause incorrect segmentation.
Applying the maximum character width feature created the correct bounding boxes. In this case, the character widths vary from 21-25 pixels. The OCR software determined the correct split.
Metal cases with stamped letters can create problems for any OCR reader. In this application it was further compounded by streaks in the finish. OCR A was used as the font style. This application required all three lines to be read. However, streaking and variable shading created difficulty in positioning. EconoCR™ returns the bounding box positions which can be used to position other reader windows. This approach was used for this application. The month was chosen as the reference point since it suffered the least number of image problems.
As the summary line shows, all three lines were correctly read. (The time was 8.7 milliseconds.) The teal color indicates computed, or single threshold for the reader window. The pink color indicates non-linear thresholding which is a seperate threshold value for each character.
Molded numbers in the bottom of a plastic bottle create a very challenging OCR application. Lighting and lenses are critical for success. Fortunately, in this application, the three numbers coming down the line were always to be the same. If a different number was encountered, the line was to stop. The key here was to not misread. Another key was the fact that the numbers remained in a relatively small quadrant of the bottle. Since the quality of the characters was quite variable, the three digits were taught as a single pattern. Several samples were taught representing slightly different angles or representations.
With searching enabled and three samples taught, the time required to search and return the string was 1.8 milliseconds.
IC Substrates create unique and difficult problems for OCR. Two major issues are contamination and gaussian noise. The sequence of images below contains both.
 |
 |
| Text on Substrate with "noise" |  |
 |
 |
| Text on Substrate contaminated |  |
The font file is the standard OCR A file. Fielding is used to define the character type at each position. The inter-character spacing is set to "step over" noise between characters.
Flexibility and control over the library are key features of EconoCR™.
As with many label applications, more than one line of text must be read. Normally, this is done using one "reader" for each line of text. For three lines of text, this triples the work required to setup and properly align each reader. In many cases, alignment can never be achieved due to small variations in printing and rotation.
This example clearly demonstrates the use of a single reader in a multiline application. As an added benefit, different character fielding parameters were set for each line. Thus, improving both read accuracy and performance.
One of the results of high speed ink-jet or laser printing is slanted text. When the slant is severe, it will cause one character to overlap the next character. Thus, individual character segmentation is impossible.
As can be seen in the second line, the first five characters overlap one another. The only solutions are either limit the maximum width for each character when segmenting or straighten or de-slant the text.
By applying a -14 degree slant angle, the text in the second line is effectively straightened without distorting the characters' aspect ratio. Once de-slanted, each character can be correctly segmented and trained.
The EReader™ demonstration program combines rotation and de-slanting of text for reading, training, and automatic reader configuration. Train a row of text rotated 90, 180, or 270 degrees, and / or flipped, and / or reversed and then read another image in any of these orientations.
Concepts in Computing, Inc. has released the latest enhancement to the utility library used in conjunction with the EconoCR™ Optical Character Recognition Library. Gradient Data extraction has been included to augment the threshold previously provided. Gradient Data extraction is useful for applications where thresholding is difficult or impossible to use.
The latest EReader™ demonstration program contains example implementations of these features.
Concepts in Computing, Inc. supplies OCR solutions for the industrial vision market. EconoCR™ is currently used by DVT, National Instruments, Coreco Imaging, PPT Vision, and a number of integrators and end users world wide.
Concepts in Computing, Inc. is a leading supplier of high speed Optical Character Recognition software, EconoCR™, for the industrial environment. Optical Character Recognition (OCR) is used in a wide variety of industries for identicfication and / or classification of process components. These may include labels, integrated circuits, stamped engine blocks and other automotive components. Currently man readable codes are required as supplements to barcodes in food processing, pharmaceuticals, electronics, automotive, bulk mailing and many more.
An example of an application for OCR in the electronics industry would be the reading of serial and lot numbers on printed circuit boards.
Concepts in Computing, Inc. is a leading supplier of high speed Optical Character Recognition software, EconoCR™, for the industrial environment. Optical Character Recognition (OCR) is used in a wide variety of industries for identicfication and / or classification of process components. These may include labels, integrated circuits, stamped engine blocks and other automotive components. Currently man readable codes are required as supplements to barcodes in food processing, pharmaceuticals, electronics, automotive, bulk mailing and many more.
An example of an application for OCR in the food processing industry would be the reading of date and manufacture site codes on the bottom of canned goods. The software has the capability of angular orientation of items as they come down a conveyor line.
Concepts in Computing, Inc. is a leading supplier of high speed Optical Character Recognition software, EconoCR™, for the industrial environment. Optical Character Recognition (OCR) is used in a wide variety of industries for identicfication and / or classification of process components. These may include labels, integrated circuits, stamped engine blocks and other automotive components. Currently man readable codes are required as supplements to barcodes in food processing, pharmaceuticals, electronics, automotive, bulk mailing and many more.
An example of an application for OCR in the pharmaceutical industry would be the reading of drug content and prescription information found on the container.
Mutech is a privately held Massachusettes Corporation founded in 1988 to distribute high technology products internationally. In 1992 the company was re-focused to developing plug-in boards to IBM PC™ compatible computers that digitize images and display them on the computer's monitor.
Applications include:
Industrial Vision Machine Vision Security Surveillance Image Analysis Forensics, Fingerprint Enchancement Image Database Image Processing Research Optical Character RecognitionA division of Coreco Imaging, Inc., ipd, specializes in providing machine vision solutions for the factory floor. A general-purpose vision package, Sherlock™, is used for these applications, one of which is OCRPlus™. Sherlock™ is a Windows™-based "point and click" software environment specifically designed to simplify development and deployment of machine vision solutions. Providing a suite of robust and proven tools inside a flexible, yet easy-to-use interface, Sherlock™ has become the general-purpose product of choice for system integrators and end-users.
Data Translation, Inc. is a world leader in Data Acquisition, Image Analysis and Machine Vision hardware and software products. They have designed and built the highest quality leading edge technology printed circuit boards for the past 25 years. Low to mid volume assembly manufacturing and quick turn prototype assembly is their specialty. Turnkey and consigned jobs can be quickly quoted and processed.
The EconoCR™ library is compatible with DT Vision Foundry.
DVT offers a line of smart sensors for support in the areas of quality control inspections, coordinate information for motion controllers, statistical process control, OCR and 2-D code verification. Industrial applications include automotive, electronics, food/beverage, health care, metal processing, packaging, plastics and semiconductors.
National Instruments offers a wide assortment of tools, bundles and services to help build the best solution for your application needs.
These include:
Scale quickly from prototyping environment to programming environment Integrate machine vision frame grabbers easily with NI data acquisition and motion Deploy embedded real-time machine vision systems on the PXI™ platform using Use best-of-class third-party machine vision components Perform complex image processing tasks with lower production costs Choose from hundreds of machine vision functions including OCRNEOpt is a vision integrator and distribution company that supports OCR. For more information contact:
NEOpt CorporationOmnitron is a German company that offers a single beam reading head S100™ that is designed for online reading of Pharmacodes printed on packaging material such as labels, cartons, leaflets and other OCR applications. In the automated pharmaceutical manufacturing process Pharmacode is used for 100% online integrity checking of the various packing material. Up to date packaging machines process the goods at very high speeds. In case of intermittent motion peak processing speeds of 2 m/sec could easily be reached.
Since their founding in 1982, PPT has specialized in industrial applications where accuracy, repeatability, high speed and flexibility are important requirements. Their products are sold throughout the Americas, Europe and Asia either directly to end users, or through a network of systems integrators, machine builders, and full service international distributors.
Machine vision-based inspection systems enable manufacturers to realize significant economic paybacks by increasing the quality of manufactured parts and improving the productivity of manufacturing processes. PPT's 2D™ machine vision product line, which includes OCR, is sold on a global basis to end-users, system integrators, and original equipment manufacturers primarily in the electronic and semiconductor component, automotive, medical device, pharmaceutical and packaged goods industries.
Vision Components is a German-based company that boasts the following:
Smallest smart cameras (60 X 60 X 10mm) Fastest smart cameras (1200 Mips) Best smart camera hardware variety (19 models) Best software variety (>100 suppliers) Lowest smart camera prices (100+ << 500 Euro)This line of low-cost high-speed smart cameras is ideal for an OCR application.
Western Vision Software allows the user to create their own custom machine vision applications. HLIamge++™ is an object-orientated image processing and machine vision software product that runs under Windows98™, NT™ and 2000™. The software is ideal for scientific imaging applications requiring measuring, counting, edge tracking, calibration, data logging, size distribution analysis and OCR.