Empowering photovoltaic power generation with edge computing: A recognition and location approach for hot spot

The presented paper proposes a photovoltaic module hot spot detection algorithm based on YOLOv8‐BCB. The algorithm addresses issues such as component efficiency degradation and poor contact in long‐term operation of PV systems, with a focus on the hot spot effect’. To achieve rapid detection and loc...

Full description

Saved in:
Bibliographic Details
Published in:Electronics letters 2023-12, Vol.59 (24), p.n/a
Main Authors: Cao, Shenglei, Wang, Yashun, Wang, Zun, Guan, Rongqiang, Ding, Liyuan, Lv, Yuze
Format: Article
Language:English
Subjects:
Accuracy
Algorithms
Alternative energy sources
Conflicts of interest
Datasets
Deep learning
Edge computing
Energy consumption
flat panel displays
infrared imaging
Kitchenware
Localization
Photovoltaic cells
photovoltaic effects
Sequences
Solar energy
Target detection
Unmanned aerial vehicles
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The presented paper proposes a photovoltaic module hot spot detection algorithm based on YOLOv8‐BCB. The algorithm addresses issues such as component efficiency degradation and poor contact in long‐term operation of PV systems, with a focus on the hot spot effect’. To achieve rapid detection and localization of small targets in complex scenes, the algorithm incorporates the Weighted Bi‐directional Feature Pyramid Network (BiFPN) into YOLOv8. Additionally, a lightweight upsampling operator called Content‐Aware ReAssembly of FEatures (CARAFE) is used to reduce the detection load on the unmanned aerial vehicle (UAV) detection system. The BiFormer attention mechanism is also integrated into the C2f module to better capture dependencies between sequences and reduce background interference. To highlight the superiority of the authors’ algorithm compared to other high‐quality algorithms, the authors conducted comparative experiments on the same dataset. The YOLOv8‐BCB algorithm surpasses the SSD, faster‐cnn, and RetinaNet algorithms in both accuracy and detection speed. It achieves a precision of 97.1% and a recall rate of 94.9%, with an FPS of 110. The result is a target detection algorithm that is both fast and accurate. 1. A detection model of hot spot for photovoltaic (PV) panel based on YOLOv8‐BCB is established; 2. A small target detection algorithm for unmanned aerial vehicle is proposed; 3. A thermal spot detection test of PV power station was carried out.
ISSN:0013-5194
1350-911X
DOI:10.1049/ell2.13056