Research And Implementation Of PDF417 Recognition Based On Embedded System

With the development of information and computer technology, barcodes emerge and develop well heretofore. Further more, with the popularization of IMM and 3G, mobile 2D barcodes are influencing our life deeply. Obtaining the information of business cards, automatically linking the network, using electron notes through mobile barcodes are being more and more applied. The purpose of this paper is to detect and recognize PDF417 barcode through J2ME MIDlets.First, this paper studies pre-process technology of PDF417 barcode images. Barcode images may have high light districts, image noise or illegibility because of equipments'weak capability or the impact of ambient light. This paper presents a binary algorithm which calculates different binarization thresholds for uneven light regions and the other region. Results show that this binary algorithm can overcome the uneven light of the barcode image effectively and fulfill the real-time requirement.Following, the barcode image may be of complicated background and misshapen. Aimed at this phenomenon, this paper presents a barcode detection algorithm which optimizes the detection algorithm based on sub-regions'multiple features to eliminate most background of the image, then detects PDF417 barcodes based on the encoding features. The detection algorithm uses a variance threshold to find regions that are similar to PDF417 barcodes'start pattern or stop pattern which is propitious to detect distorted barcodes. Results show that this algorithm can detect barcodes more effectively.Then, this paper presents an improved fast distortion correction algorithm based on studies of the classical algorithm. The result shows that this algorithm can correct distortions of barcodes more quickly.Finally, we designed a barcode decoding system based on J2ME platform, and resolved the technical characters of J2ME application development, such as image capture, file orientation, multi-thread, memory optimization, etc. Results show that this system can decode PDF417 with a good performance.