包装与食品机械2025，Vol.43Issue(3)：62-70,9.DOI:10.3969/j.issn.1005-1295.2025.03.007
弱光环境下自动装箱系统的视觉识别方法
Visual recognition method of automatic case loading system in low-light environment
摘要
Abstract
To address the low recognition accuracy of boxes in automatic case loading systems under low-light conditions inside vans,this study proposes a visual recognition method based on an optimized Kindling the Darkness(KinD)network and You Only Look Once 11(YOLO 11)object detection.Structural optimizations were applied to the KinD network by introducing bilateral filtering denoising and Sobel edge enhancement modules,effectively improving image brightness and detail.Corrugated boxes were detected using the YOLO 11 algorithm for 3D localization and pose estimation.Results show that after enhancement by the optimized KinD network,the average peak signal-to-noise ratio reached 15.04 dB,the structural similarity index was 0.72,and image processing time was 0.338 s.The YOLO 11 algorithm achieved 26.7%full matching and 42.7%high matching for enhanced images,with an overall normalized mean position error of 0.014 3.This research provides technical support for automated loading/unloading in logistics.关键词
自动装箱系统/弱光环境/图像增强/目标检测/KinD网络/YOLO 11Key words
automatic case loading system/low-light environment/image enhancement/object detection/KinD network/YOLO 11分类
通用工业技术引用本文复制引用
丁孟孟,岳晓丽..弱光环境下自动装箱系统的视觉识别方法[J].包装与食品机械,2025,43(3):62-70,9.基金项目
国家重点研发计划项目(2017YFB0309701) （2017YFB0309701）
