Image Processing Bar Codes - Selecting The Right Bar Code With Image Binarization
Below is a MRR and PLR article in category Computers Technology -> subcategory Web Development.
Image Processing & Bar Codes: Choosing the Right Bar Code with Image Binarization
Summary
When documents are scanned for image processing or storage, they are typically captured as grayscale images with 8 bits per pixel, representing 256 shades of gray. While essential for capturing photos and graphics, this detail isn't typically needed for text, which is why scanned images are often binarized. Binarization selects a threshold to convert pixel values into black or white, reducing the image size from 3.6 megabytes to 450 kilobytes for a standard document. While this process is effective for text and fax transmission, it can alter bar codes enough to complicate decoding. This article explores two key considerations to improve the readability of binarized bar codes: sample density and symbology type.Sample Density
Sample density, the number of samples per inch (DPI), ranges typically from 200 to 300 for document processing. For bar code decoding, the critical measurement is samples per module, which includes both the device's sample density and the bar code's size. To reduce issues caused by variations from binarization, a typical value of 4 pixels per module is recommended for bar codes, with a minimum module size of 20 mils for documents scanned at 200 DPI. Increasing samples per module is advisable for documents subjected to multiple binarizations, such as repeated fax transmissions.Symbology Type
The type of symbology also significantly impacts readability after binarization. Linear bar codes can be divided into wide/narrow codes or multiple width codes. Wide/narrow codes, like Code 39 and Interleaved 2 of 5, use two possible widths and perform better post-binarization because they are easier to decode. Multiple width codes, such as UPC and Code 128, allow for more width variations and can encounter more decoding challenges. Although wide/narrow symbologies have lower data density, they are favorable when binarization is involved.2D Symbologies
Two-dimensional matrix codes, like Data Matrix and QR Code, rely on a grid to encode data. These codes offer higher data density and error correction, making them robust against binarization errors with sufficient samples per module. A sample density of 5 pixels per module ensures good read performance. Although PDF417 is often regarded as a 2D symbology, it functions as a stacked linear code and may struggle with binarization at lower sample densities.Conclusion
For scanned documents with bar codes to be binarized, using linear symbologies with wide/narrow choices and at least 4 pixels per module enhances read performance. Two-dimensional matrix codes, with a minimum of 5 pixels per module, further increase data density and reliability, making them an excellent choice for binarized images.You can find the original non-AI version of this article here: Image Processing Bar Codes - Selecting The Right Bar Code With Image Binarization.
You can browse and read all the articles for free. If you want to use them and get PLR and MRR rights, you need to buy the pack. Learn more about this pack of over 100 000 MRR and PLR articles.