Research Article
Efficient Parcel Damage Detection via Faster R-CNN: A Deep Learning Approach for Logistical Parcels’ Automated Inspection
@INPROCEEDINGS{10.1007/978-3-031-63992-0_18, author={Zhi Chen and Cuifeng Du and Quanlong Guan and Yuyu Zhou and Vichen Hoo and Xiujie Huang and Zhefu Li and Shuanghuan Lv and Xiaofeng Wu and Xiaotian Zhuang}, title={Efficient Parcel Damage Detection via Faster R-CNN: A Deep Learning Approach for Logistical Parcels’ Automated Inspection}, proceedings={Mobile and Ubiquitous Systems: Computing, Networking and Services. 20th EAI International Conference, MobiQuitous 2023, Melbourne, VIC, Australia, November 14--17, 2023, Proceedings, Part II}, proceedings_a={MOBIQUITOUS PART 2}, year={2024}, month={7}, keywords={Faster R-CNN Parcel Damage Detection IoT}, doi={10.1007/978-3-031-63992-0_18} }- Zhi Chen
Cuifeng Du
Quanlong Guan
Yuyu Zhou
Vichen Hoo
Xiujie Huang
Zhefu Li
Shuanghuan Lv
Xiaofeng Wu
Xiaotian Zhuang
Year: 2024
Efficient Parcel Damage Detection via Faster R-CNN: A Deep Learning Approach for Logistical Parcels’ Automated Inspection
MOBIQUITOUS PART 2
Springer
DOI: 10.1007/978-3-031-63992-0_18
Abstract
Parcel damage detection poses a significant challenge in warehouse automation and conveyor belt transportation, with a particular focus on leveraging the potential of Internet of Things (IoT) technologies. With this purpose in mind, we introduce RPA R-CNN, an improved deep learning-based object detection method that seamlessly integrates IoT capabilities. Our proposed approach enhances the feature extraction network of the Faster R-CNN model and utilizes advanced post-processing techniques to further refine the extracted features. By employing a meticulously curated dataset specifically designed for parcel detection, we train our model and conduct rigorous experimental verification. Encouragingly, the outcomes convincingly demonstrate that the RPA R-CNN method surpasses the performance of the original model in various metrics in domestic or cross-border logistics parcel detection. These promising results not only offer a practical solution to the challenging problem of identifying parcel damage but also underscore the immense potential of integrating IoT principles to optimize automation and enhance operational efficiency within the domestic or cross-border logistics industry.
