Barcode Localization And Recognition In Complex Scenes

The widely used barcode has promoted the development of economic globalization, it provides common language for the trades between different countries and regions, it also brings great convenience for commodity management. The dedicated barcode reader is the most commonly used device for barcode recognition, it converts the barcode image into the electrical signals and obtains the information after shaping and decoding. However, the barcode reader relies on manually operation to locate the barcode. Hence, the barcode reader is quite inefficiency. With the development and application of image processing technology, the barcode recognition based on image processing has become a research focus in recent years.In order to improve the efficiency of barcode processing, this paper presents a method for localizing and recognizing multiple barcodes in complex scenes. The method is composed of two parts, barcode localization and barcode recognition. We construct a new feature parameter, the ratio of horizontal gradient to vertical gradient of each sub-block of the image, which can reflect the distinctive feature of barcode area from other parts of the image. We then use this feature to detect the positions of each barcode with an adaptive threshold extracted from the image. Since the barcode information is encoded in the width of each bar, the key procedure of barcode recognition is determining the width of each bar. However, the resolution of barcode image that located in the complex scenes is quite low. We propose to process the detected barcode images by sub-pixel interpolation algorithms and reconstruction algorithms based on the projection curve, respectively. The barcode can be decoded by the similar edge distance algorithm after obtaining the width of each bar.The experiments have been conducted on a variety of commodity images, which contain multiple barcodes and other commodity images. The experiment results illustrate that 97% of barcodes are located with our localization algorithm and 85% of barcodes can be recognized with our reconstruction algorithm based on the projection curve. These results verify the effectiveness of the barcode localization and recognition algorithm that proposed in the paper.