Lightweight object detection model for food freezer warehouses
Abstract Warehouses are critical logistics nodes, with food freezer warehouses playing a key role in ensuring food quality while facing challenges such as high-density item distribution and extremely low temperatures required for occupational safety. Traditional management methods struggle to meet t...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-01-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-86662-z |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832585698725593088 |
|---|---|
| author | Jiayu Yang Zhihong Liang Mingming Qin Xingyu Tong Fei Xiong Hao An |
| author_facet | Jiayu Yang Zhihong Liang Mingming Qin Xingyu Tong Fei Xiong Hao An |
| author_sort | Jiayu Yang |
| collection | DOAJ |
| description | Abstract Warehouses are critical logistics nodes, with food freezer warehouses playing a key role in ensuring food quality while facing challenges such as high-density item distribution and extremely low temperatures required for occupational safety. Traditional management methods struggle to meet these demands, underscoring the need for intelligent and digital solutions to improve efficiency and mitigate safety risks. This study proposes the YOLOv8-RSS model, a lightweight and high-precision approach tailored for food freezer warehouse scenarios. The model incorporates the novel C2f_RDB module, which enhances detection accuracy while reducing parameter count and computational load. Additionally, the SimAM attention mechanism is applied to the Backbone’s final layer, enabling focus on critical image information without increasing parameters. Soft-NMS replaces the traditional NMS method, further improving detection accuracy. Experiments conducted on the food freezer warehouse dataset demonstrate that the YOLOv8-RSS model reduced the parameter count by 0.05 M, decreased FLOPs by 0.8G, increased mAP@0.5 by 1.4%, and improved mAP@0.5:0.95 by 3.9%. The YOLOv8-RSS is designed to meet the complex detection demands in food freezer warehouses, enabling precise and rapid detection of personnel and forklifts. It provides strong technical support for addressing various challenges in these environments and holds significant application value. |
| format | Article |
| id | doaj-art-3aaafda8ea8948918b988c9dd4977158 |
| institution | Kabale University |
| issn | 2045-2322 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj-art-3aaafda8ea8948918b988c9dd49771582025-01-26T12:33:32ZengNature PortfolioScientific Reports2045-23222025-01-0115111710.1038/s41598-025-86662-zLightweight object detection model for food freezer warehousesJiayu Yang0Zhihong Liang1Mingming Qin2Xingyu Tong3Fei Xiong4Hao An5Institute of Big Data and Artificial Intelligence, Southwest Forestry UniversityInstitute of Big Data and Artificial Intelligence, Southwest Forestry UniversityInstitute of Big Data and Artificial Intelligence, Southwest Forestry UniversityInstitute of Big Data and Artificial Intelligence, Southwest Forestry UniversityInstitute of Big Data and Artificial Intelligence, Southwest Forestry UniversityInstitute of Big Data and Artificial Intelligence, Southwest Forestry UniversityAbstract Warehouses are critical logistics nodes, with food freezer warehouses playing a key role in ensuring food quality while facing challenges such as high-density item distribution and extremely low temperatures required for occupational safety. Traditional management methods struggle to meet these demands, underscoring the need for intelligent and digital solutions to improve efficiency and mitigate safety risks. This study proposes the YOLOv8-RSS model, a lightweight and high-precision approach tailored for food freezer warehouse scenarios. The model incorporates the novel C2f_RDB module, which enhances detection accuracy while reducing parameter count and computational load. Additionally, the SimAM attention mechanism is applied to the Backbone’s final layer, enabling focus on critical image information without increasing parameters. Soft-NMS replaces the traditional NMS method, further improving detection accuracy. Experiments conducted on the food freezer warehouse dataset demonstrate that the YOLOv8-RSS model reduced the parameter count by 0.05 M, decreased FLOPs by 0.8G, increased mAP@0.5 by 1.4%, and improved mAP@0.5:0.95 by 3.9%. The YOLOv8-RSS is designed to meet the complex detection demands in food freezer warehouses, enabling precise and rapid detection of personnel and forklifts. It provides strong technical support for addressing various challenges in these environments and holds significant application value.https://doi.org/10.1038/s41598-025-86662-zDeep learningObject detectionWarehouse ManagementYou-only-look-once (YOLO) |
| spellingShingle | Jiayu Yang Zhihong Liang Mingming Qin Xingyu Tong Fei Xiong Hao An Lightweight object detection model for food freezer warehouses Scientific Reports Deep learning Object detection Warehouse Management You-only-look-once (YOLO) |
| title | Lightweight object detection model for food freezer warehouses |
| title_full | Lightweight object detection model for food freezer warehouses |
| title_fullStr | Lightweight object detection model for food freezer warehouses |
| title_full_unstemmed | Lightweight object detection model for food freezer warehouses |
| title_short | Lightweight object detection model for food freezer warehouses |
| title_sort | lightweight object detection model for food freezer warehouses |
| topic | Deep learning Object detection Warehouse Management You-only-look-once (YOLO) |
| url | https://doi.org/10.1038/s41598-025-86662-z |
| work_keys_str_mv | AT jiayuyang lightweightobjectdetectionmodelforfoodfreezerwarehouses AT zhihongliang lightweightobjectdetectionmodelforfoodfreezerwarehouses AT mingmingqin lightweightobjectdetectionmodelforfoodfreezerwarehouses AT xingyutong lightweightobjectdetectionmodelforfoodfreezerwarehouses AT feixiong lightweightobjectdetectionmodelforfoodfreezerwarehouses AT haoan lightweightobjectdetectionmodelforfoodfreezerwarehouses |