Research Of Barcode Detection Algorithm

In recent years,with the popularity of mobile camera equipment,more and more people began to use mobile phones and other devices to scan barcode to obtain commodity information,payment of goods and so on.This makes barcode detection and identification become more difficult.So it is very important to detect and locate the barcode accurately in the whole detection and recognition.The biggest challenge of barcode detection and positioning is how to detect the barcode quickly and accurately in a variety of complex backgrounds.Therefore,this paper mainly studies how to detect the barcode in a variety of complex backgrounds,in order to facilitate the subsequent identification work.The main contents of this paper are as follows:(1)Aiming at the problem that the tilt barcode is not easy to locate accurately,we propose using a region detection algorithm.The first step is detect the image with MSER(Maximally Stablly Stable Extremal Regions)algorithm,and then through the main direction analysis to get the angle of barcode.Finally,correcting the direction of image.Compared with the traditional methos,it is very fast and accurate.(2)It is difficult to detect the barcode angle for the fuzzy image using the MSER method.We enhance the fuzzy image by Adaptive Manifolds(AM)filter and then detecting the angle.This method improves the robustness of the MSER detection result.(3)Aiming at the problem that is difficult to detect barcode in complex background.We propose the method based on CNN to detect barcode.This method makes full use of the advantages of CNN in feature learning,and has the advantages of high detection accuracy and good robustness.After the experiment in the Muenster and ArTe-Lab barcode datasets,the proposed method's detection accuracy is greatly improved compared to previous method in a variety of complex scenes and has a good detection robustness.