Hello. Are there any approaches for fighting an optical irradiation on black-white barcodes when processing images? For example, i'm using circular barcode which circular rings have same sizes. But sizes of black & white rings on photography differ from reference barcode (white approx. 1.5-2x larger than black) on large distances.
Have you even faced this problem? Any papers maybe.
Hello Friend Mikhai Verkeenko!
It is not clear from your question description above what photographic technique you are using, but you should take into account that some changes in the images are eventual due to optical effects (known as Geometric and Chromatic aberrations). These are the defects in the optical systems.
If you are using digital camera/s to capture the images of these bar-codes or the concentric circles (as shown in the image) and applying digital image processing, then better pay attention to the camera optics and the illuminating light source as well.
Then, as you have stated, the sizes in the photography (I believe the images obtained) differ from the reference bar-code. This is likely to happen and you shall have to obtain a correction factor for the images and this will also depend upon your optical magnifications. You can have circles of known and standard sizes and then calibrate the number of pixels at four different positions separated by 90o angles. However, the more points you choose, the better will be your accuracy.
The wavelength of illuminating light source also plays a vital role in acquiring proper images. Try acquiring images with commonly available light sources such as RED, GREEN etc. to see that under which wavelength you get the best contrast of your images.
This is just a short explanation of the possible ways that might help you solving your problem. I shall be more than happy to help you further.
Thanks with best wishes!
Hello,
on the image processing side (processing a grey scale image) you could try to "erode" the bright (white) areas. See
https://en.wikipedia.org/wiki/Erosion_%28morphology%29#Grayscale_erosion
for some guidance (IMHO it's not too 'understandable' there). Better find literature about grey scale processing and search for 'erosion'.
The filter to be used depends on the "size" of your problem: 3x3 matrizes are usual, 5x5 also.
In principle 'erosion' is nothing more than replacing the current pixel by the minimum value in the neighborhood defined. This will increase dark areas and reduce white areas - the effect you are obviously needing.
- Badges
- Science topic