徐州工程学院学报（自然科学版）2026，Vol.41Issue(1)：39-45,7.
图像二值化分割在物流包装物标识图像识别中的应用
Application of Image Binarization Segmentation in the Recognition of Logistics Packaging Identification Images
摘要
Abstract
In response to the issues of poor segmentation effect and high identification error rates in existing recognition methods for identification images,this study proposes a logistics packaging identification image recognition method based on image binarization segmentation.The original logistics packaging identification image is preprocessed through grayscale conversion and Gaussian filtering to select the optimal threshold and reduce background interference.The anchor box position and identification area are determined via identification localization and loss function computation.The binary segmentation algorithm is further improved by integrating global and local information.Using this enhanced algorithm,the packaging identification image is segmented to extract the target identification region.This target-region image is then input into an improved deep convolutional neural network model,where the identification is successively localized,recognized,and decoded.Test results demonstrate that the logistics packaging label images segmented by the proposed method retain complete label information,free from interference such as noise points or light spots.The minimum misrecognition rate of packaging labels reaches 0.05%.关键词
物流包装物标识/灰度化处理/条形码识别/图像二值化分割/标识区域/误识率Key words
logistics packaging identification/grayscale processing/barcode recognition/image binarization segmentation/identification area/misrecognition rate分类
信息技术与安全科学引用本文复制引用
夏希鼎,万飞..图像二值化分割在物流包装物标识图像识别中的应用[J].徐州工程学院学报（自然科学版）,2026,41(1):39-45,7.基金项目
安徽省高校自然科学研究项目(自然科学类)(2025AHGXZK30914) （自然科学类）
