You Only Look Once–Aluminum: A Detection Model for Complex Aluminum Surface Defects Based on Improved YOLOv8 DOI Open Access

Jiashu Han,

H. Y. Chen,

Yongkun Ding

et al.

Symmetry, Journal Year: 2025, Volume and Issue: 17(5), P. 724 - 724

Published: May 9, 2025

Detecting aluminum defects in industrial environments presents significant challenges related to low-resolution images, subtle damage features, and an imbalance between easy difficult samples. The You Only Look Once–Aluminum (YOLO-AL) algorithm proposed this paper addresses these challenges. Firstly, enhance the model’s performance on images small object detection, as well improve its flexibility adaptability, C2f-US replaces first two CSP bottleneck with 2 Convolutions (C2f) layers original Backbone network. Secondly, boost multi-scale context capture strip defect a CPMSCA mechanism class-symmetric structure is integrated at end of Thirdly, efficiently both high-level semantics low-level spatial details, detection complex surface defects, ODE-RepGFPN introduced replace entire Neck Finally, address hard samples, Focaler-WIoU proposed. Extensive experiments conducted publicly available AliCloud dataset (APDDD) demonstrate that YOLO-AL achieves 86.5%, 77.8%, 81.5% for Precision, Recall, [email protected], respectively, surpassing baseline model other state-of-the-art methods. can be camera system automated inspection profiles production environment.

Language: Английский

Research on an Apple Recognition and Yield Estimation Model Based on the Fusion of Improved YOLOv11 and DeepSORT DOI Creative Commons
Zhibo Yan, Yu-Wei Wu, Wenbo Zhao

et al.

Agriculture, Journal Year: 2025, Volume and Issue: 15(7), P. 765 - 765

Published: April 2, 2025

Accurate apple yield estimation is essential for effective orchard management, market planning, and ensuring growers’ income. However, complex conditions, such as dense foliage occlusion overlapping fruits, present challenges to large-scale estimation. This study introduces APYOLO, an enhanced detection algorithm based on improved YOLOv11, integrated with the DeepSORT tracking improve both accuracy operational speed. APYOLO incorporates a multi-scale channel attention (MSCA) mechanism prior distribution intersection over union (EnMPDIoU) loss function enhance target localization recognition under environments. Experimental results demonstrate that outperforms original YOLOv11 by improving [email protected], [email protected]–0.95, accuracy, recall 2.2%, 2.1%, 0.8%, 2.3%, respectively. Additionally, combination of unique ID region line (ROL) strategy in further boosts 84.45%, surpassing performance method alone. provides more precise efficient system estimation, offering strong technical support intelligent refined management.

Language: Английский

Citations

2
Intelligent Vision System for Real-Time Pallet Detection, Counting and Efficient Warehouse Management DOI Creative Commons

Kunal G. Borase,

R Sowmiya,

Bharani Kumar Depuru

et al.

Published: April 1, 2025

Nowadays, keeping track of inventory accurately is a big challenge in warehouses, especially when dealing with large-scale pallet production. Traditional methods like manual counting can lead to errors, such as incorrect counts, causing inefficiencies and delays order fulfillment. These issues not only slow down operations but also increase costs impact the entire supply chain. This paper implements an high level solution deploying AI-powered system using deep learning computer vision count stacked pallets real-time, improving management operational efficiency develop reliable detection system, extensive data collection was carried out warehouse settings capturing images under different conditions dataset carefully labeled polygon-based annotations via open source annotation tool roboflow facilities for augmentations. To ensure precise object various augmentation techniques shear, exposure, noise, blur, grayscale, horizontal flip, saturation, rotation brightness adjustments were applied improve model robustness against real-world variations For real-time high-accuracy , YOLO (You Only Look Once) models YOLOv8, YOLOv9, YOLO11 used training optimization. offered fast inference speeds, ensuring low latency while maintaining precision. A comparative analysis versions provided insights into performance, accuracy, efficiency. The optimized implemented setting connected monitoring enabling automatic reducing effort errors. research highlights Switching from traditional presenting how enhance precision, minimize human mistakes, cut expenses. By implementing AI-driven Businesses optimize chain processes, implement scalable automation.

Language: Английский

Citations

0
DEEP LEARNING FRAMEWORK FOR FRUIT COUNTING AND YIELD MAPPING IN TART CHERRY USING YOLOv8 and YOLO11 DOI Creative Commons
Anderson Luiz dos Santos Safre, Alfonso F. Torres‐Rua, Brent Black

et al.

Smart Agricultural Technology, Journal Year: 2025, Volume and Issue: unknown, P. 100948 - 100948

Published: April 1, 2025

Language: Английский

Citations

0
YOLOv11n for precision agriculture: lightweight and efficient detection of guava defects across diverse conditions DOI
G. Madasamy Raja,

P. Selvaraju,

P. Pathmanaban

et al.

Journal of the Science of Food and Agriculture, Journal Year: 2025, Volume and Issue: unknown

Published: May 5, 2025

Abstract BACKGROUND Automated fruit defect detection plays a critical role in improving postharvest quality assessment and supporting decision‐making agricultural supply chains. Guava presents specific challenges because of diverse disease types, varying maturity levels inconsistent environmental conditions. Although existing you only look once (YOLO)‐based models have shown promise tasks, they often face limitations balancing accuracy, inference speed computational efficiency, particularly resource‐constrained settings. This study addresses this gap by evaluating four YOLO (YOLOv8s, YOLOv5s, YOLOv9s YOLOv11n) for detecting defective guava fruits across five diseases (scab, canker, chilling injury, mechanical damage rot), three (mature, half‐mature immature) healthy fruits. RESULTS Diverse datasets facilitated robust training evaluation. YOLOv11n achieved the highest mAP50‐95 (98.0%) exhibited bounding box loss (0.0565), classification (0.2787), time (3.9 milliseconds) (255 FPS). YOLOv5s had precision (94.9%), while excelled recall (96.2%). YOLOv8s offered balanced performance metrics. outperformed all with lightweight architecture (2.6 million parameters) low cost (6.3 giga floating‐point operations per second), making it suitable applications. CONCLUSION These results highlight YOLOv11n's potential applications, such as automated control, which require high accuracy real‐time analysis provides insights into deploying to enhance efficiency reliability management. © 2025 Society Chemical Industry.

Language: Английский

Citations

0
You Only Look Once–Aluminum: A Detection Model for Complex Aluminum Surface Defects Based on Improved YOLOv8 DOI Open Access

Jiashu Han,

H. Y. Chen,

Yongkun Ding

et al.

Symmetry, Journal Year: 2025, Volume and Issue: 17(5), P. 724 - 724

Published: May 9, 2025

Detecting aluminum defects in industrial environments presents significant challenges related to low-resolution images, subtle damage features, and an imbalance between easy difficult samples. The You Only Look Once–Aluminum (YOLO-AL) algorithm proposed this paper addresses these challenges. Firstly, enhance the model’s performance on images small object detection, as well improve its flexibility adaptability, C2f-US replaces first two CSP bottleneck with 2 Convolutions (C2f) layers original Backbone network. Secondly, boost multi-scale context capture strip defect a CPMSCA mechanism class-symmetric structure is integrated at end of Thirdly, efficiently both high-level semantics low-level spatial details, detection complex surface defects, ODE-RepGFPN introduced replace entire Neck Finally, address hard samples, Focaler-WIoU proposed. Extensive experiments conducted publicly available AliCloud dataset (APDDD) demonstrate that YOLO-AL achieves 86.5%, 77.8%, 81.5% for Precision, Recall, [email protected], respectively, surpassing baseline model other state-of-the-art methods. can be camera system automated inspection profiles production environment.

Language: Английский

Citations

0