Express Barcode Image Extraction based on Machine Vision

Due to creases, blurring, scratches, and reflections on the surface of express delivery labels, it is not possible to quickly extract the barcodes, leading to the accumulation of packages and resulting in losses. This paper optimizes the images by using image restoration with generative adversarial networks (GANs) and adaptive adjustment enhancement algorithms. Secondly, after initially locating the barcode section using color and shape features, three major segmentation methods are employed to separate it out. Furthermore, this paper introduces a feature extraction method based on machine vision, which combines image processing techniques and machine learning algorithms. This method effectively extracts the features of express delivery barcodes and outperforms traditional methods.