Bar code on how to improve the recognition rate, here are a few key points:
Barcode picture quality is very important. Typically, barcode recognition processing in the pixel area counted to determine the position and width of the barcode type specified in the article. Therefore, if the resolution is not high, this process will be affected.
Typically, to achieve a resolution of at least 200dpi, and the higher the resolution the better. Most developers want to save storage space by minimizing the resolution. Just be sure not to store barcode recognition efficiency at the expense of efficiency. Remember, the lower the resolution, the more likely a lack of pixel density required for identification.
Generally, for a one-dimensional barcode, each bar code bars and spaces minimum of at least three pixels. For two-dimensional bar code, it is usually about 5 pixels. Sometimes, even though the resolution is not low but that is no good pixel density, which may be many reasons - poor quality labels or printer, and so on.
Processing the barcode unreadable or corrupted
For low quality or damaged barcodes, required to determine the content of three, to determine whether the barcode image enhancement by preprocessing:
Expansion ( Dilation ) - adding pixels to the boundaries of objects in the image, and the expansion will be deleted pixels on object boundaries.
Binarization ( Binarization ) - convert the image to a monochrome image can more easily distinguish the image edge.
Remove dust ( Despeckle ) - Consider a region pixel average value, to attempt to detect whether the pixel is a color or heteroaryl generated in said image acquisition process.
Tilt correction
If the barcode is inclined, depending on the degree of tilt angles and the pitch of the bar code may be thinner or thicker than normal. At the same time, reduce image quality correction bar code, which can lead to decreased accuracy of barcode recognition. You can use the SDK pre-processing functions of the bar code automatic skew correction to maximize recognition accuracy. But consume some time correction and reduce the speed of the barcode recognition.
Scanning distance
For the one-dimensional bar codes, scanning distance is important because it may affect bar code recognition, because it may not need to be scanned pixels. For example, the number on the page "1", whether it is this part of Chinese sentence, or part of the bar code type, based on its location and scanning distance, could be mistaken for a bar code. Accordingly, it is desirable to set the scanning distance five pixels or more, to get the best accuracy and speed of decoding. Typically, the greater the distance the scan settings (more than 5), the faster the scanning speed.
In general, compared to smart phones, laser scanners perform better at reading distance more than two feet. They usually perform better in low-light scenes. Preprocessing function by means of the built-in software, can be corrected quickly damaged or difficult to read the bar code during the scan in order to improve the recognition rate.
Attached: Dynamsoft Barcode Reader free online demo
This article is a " best practice bar code reading technology " Chapter IV series.
