Camera Based Barcode Localization and Decoding in Real-World Applications

In the last decades, many approaches were presented to localize and decode barcodes in images from off the shelf cameras. However, all proposed solutions usually only deal with one type of image artifacts, such as a poorly illuminated or noisy image, or an image that suffers from motion or out-of-focus blur. In this paper, we present a complete, fully automatic pipeline, which allows the localization and decoding of barcodes in real-world scenarios. Our method is capable of localization and decoding the correct barcode information even if the input image is noisy, poorly exposed, and blurred at the same time. We can also decode the correct information from barcode images whose resolution is actually too low, i.e., where the width of the smallest bar depicted is smaller than the width of a single pixel.