该文基于无人机光伏组件可见光图像采集,提出一种基于迁移学习的光伏组件鸟粪检测方法。方法首先基于掩膜区域卷积神经网络（Mask-RCNN）对光伏组件边界进行框选,再利用迁移学习策略,构建光伏组件鸟粪缺陷检测模型,实现鸟粪智能检测。利用上述方法,实现光伏组件鸟粪覆盖检测准确率为96.75%。
Abstract
In this paper, based on the visible light image acquisition of photovoltaic（PV） panel by using unmanned aerial vehicle, a bird droppings detection method of photovoltaic panel based on transfer learning is proposed. Firstly, the boundary of PV panel is selected based on Mask-RCNN, and then the bird droppings defect detection model of PV panel was constructed by using transfer learning strategy to realize intelligent detection of bird droppings. Using the above method, the detection accuracy of bird droppings coverage of PV panel is 96.75%.
关键词
迁移学习 /
光伏组件 /
掩膜区域卷积神经网络 /
缺陷检测
Key words
transfer learning /
photovoltaic modules /
Mask-RCNN /
defect detection
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] NICCOLAI A, GANDELLI A, GRIMACCIA F, et al.Overview on photovoltaic inspections procedure by means of unmanned aerial vehicles[C]//2019 IEEE Power Tech, Milan, 2019.
[2] 徐晶, 韩军, 童志刚, 等. 一种无人机图像的铁塔上鸟巢检测方法[J]. 计算机工程与应用, 2017, 53(6): 231-235.
XU J, HAN J, TONG Z G, et al.A bird nest detection method based on UAV image[J]. Computer engineering and application, 2017, 53(6): 231-235.
[3] 宋伟, 左丹, 邓邦飞, 等. 高压输电线防震锤锈蚀缺陷检测[J]. 仪器仪表学报, 2016, 37(S1): 113-117.
SONG W, ZUO D, ZHANG H B, et al.Corrosion defect detection of earthquake hammer for high voltage transmission line[J]. Chinese journal of scientific instrument, 2016, 37(S1): 113-117.
[4] WAN G W, WANG Y F, HAN J, et al.Recognition and drop-off detection of insulator based on aerial image[C]//International Symposium on Computational Intelligence and Design, IEEE, Hangzhou, China, 2017: 162-167.
[5] 王亚丽. 基于红外热成像的太阳能板缺陷检测[D]. 杭州: 中国计量学院, 2015.
WANG Y L.Solar panel defect detection based on infrared thermal imaging[D]. Hangzhou: Institute of Metrology, 2015.
[6] GIRSHICK R, DONAHUE J, DARRELL T, et al.Region-based convolutional networks for accurate object detection[J]. IEEE transactions on pattern analysis and machine intelligence, 2016, 38(1): 142-158.
[7] HE K M, ZHANG X Y, REN S Q.Spatial pyramid pooling in deep convolutional networks for visual recognition[J]. IEEE transactions on pattern analysis and machine intelligences, 2015, 37(9): 1904-1916.
[8] REN S, HE K M, GIRSHICK R, et al.Faster RCNN: towards real-time object detection with region proposal networks[J]. IEEE transactions on pattern analysis and machine intelligence, 2015, 39(6): 1137-1149.
[9] 田晓杰, 程耀瑜, 常国立. 基于深度学习优化SSD算法的硅片隐裂检测识别[J]. 机床与液压, 2019, 47(1): 36-40, 60.
TIAN X J, CHENG Y Y, CHANG G L.Silicon chip stealth detection and identification based on deep learning optimization SSD algorithm[J]. Machine tools and hydraulics, 2019, 47(1): 36-40, 60.
[10] 王宪保, 李洁, 姚明海, 等. 基于深度学习的太阳能电池片表面缺陷检测方法[J]. 模式识别与人工智能, 2014, 27(6): 517-523.
WANG X B, LI J, YAO H M, et al.Deep learning-based solar cell surface defect detection methods for pattern recognition and artificial intelligence[J]. Pattern recognition and artificial intelligence, 2014, 27(6): 517-523
[11] 车曦. 基于红外图像识别的光伏组件热斑故障检测方法研究[D]. 重庆: 重庆大学, 2015.
CHE X.Study on the detection method of thermal spot failure of photovoltaic modules based on infrared image recognition[D]. Chongqing: Chongqing University, 2015.
[12] 郭梦浩, 徐红伟. 基于Faster RCNN的红外热图像热斑缺陷检测研究[J]. 计算机系统应用, 2019, 28(11): 265-270.
GUO M H, XU H W.Research on thermal spot defect detection in infrared thermal images based on Fast RCNN[J]. Computer systems & applications, 2019, 28(11): 265-270.
基金
江西省科技厅重点研发项目（20192BBE50057）
PDF(1756 KB)
