应俊,刘迅,曾学仁,方亮,田楠
YING Jun,LIU Xun,ZENG Xueren,FANG Liang,TIAN Nan

• 1:上海扩博智能技术有限公司
• 2:中国三峡新能源(集团)股份有限公司西北分公司

摘要(Abstract)：

风机叶片在运行过程中,由于环境和高速旋转等原因会产生各种缺陷,这些缺陷直接影响风机发电量,严重者甚至导致叶片产生不可逆的损伤,带来巨大的经济损失。针对风机叶片的缺陷识别和检测,采用优化后的SSD(单步多框目标检测)算法,该算法从采集的缺陷样本中自主学习叶片缺陷特征,实现风机叶片缺陷的自动检测、定位和分类。最终在测试数据集上达成mAP(平均精度均值)为82.1%,召回率为90.3%。该算法已应用于企业级项目中,实践证明深度学习算法在企业级项目中具有很好的鲁棒性和商业价值。
Due to environment and high-speed rotation, turbine blades are prone to various defects that impair generation capacity, or even lead to irreversible damage to the turbine blade, resulting in huge economic losses. Aimed for turbine blade defect identification and detection, the paper adopts an optimized SSD(single shot multibox detector) algorithm, which can learn the features of defects independently and automatically detect, locate and classify blade defects. The model mAP(average mean precision) reaches 82.1% and the recall rate 90.3%. This algorithm has been applied to enterprise projects, and it shows that the deep learning algorithm has good robustness and commercial value in enterprise projects.

关键词(KeyWords)： 风机叶片;缺陷检测;SSD算法优化;超大分辨率
turbine blade;defect detection;SSD algorithm optimization;ultra-high resolution

Abstract:

Keywords:

基金项目(Foundation):

作者(Author): 应俊,刘迅,曾学仁,方亮,田楠
YING Jun,LIU Xun,ZENG Xueren,FANG Liang,TIAN Nan

DOI: 10.19585/j.zjdl.202108007

参考文献(References)：

• [1]RUSSAKOVSKY O,DENG J,SU H,et al.Image Net large scale visual recognition challenge[J].International Journal of Computer Vision,2015,115(3):211-252.
• [2]KRIZHEVSKY A,SUTSKEVER I,HINTON G E.Image Net classification with deep convolutional neural networks[J].Communications of the ACM,2017,60(6):84-90.
• [3]EVERINGHAM M,GOOL L,WILLIAMS C K I,et al.The pascal visual object classes (VOC) challenge[J].International Journal of Computer Vision,2010,88(2):303-338.
• [4]LIN T Y,MAIRE M,BELONGIE S,et al.Microsoft coco:Common objects in context[C]//European Conference on Computer Vision.Cham:Springer,2014:740-755.
• [5]REN S Q,HE K M,GIRSHICK R,et al.Faster R-CNN:Towards real-time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(6):1137-1149.
• [6]LIU W,ANGUELOV D,ERHAN D,et al.SSD:Single shot multi Box detector[C]//European Conference on Computer Vision.Cham:Springer,2016.
• [7]REDMON J,DIVVALA S,GIRSHICK R,et al.You only look once:unified,real-time object detection[C]//2016 IEEEConference on Computer Vision and Pattern Recognition(CVPR).Las Vegas,NV,USA:IEEE,2016:779-788.
• [8]REDMON J,FARHADI A.YOLO9000:Better,faster,stronger[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).Honolulu,HI,USA:IEEE,2017:6517-6525.
• [9]LAW H,DENG J.Corner Net:Detecting objects as paired keypoints[J].International Journal of Computer Vision,2020,128(3):642-656.
• [10]LIN T Y,GOYAL P,GIRSHICK R,et al.Focal loss for dense object detection[C]//2017 IEEE International Conference on Computer Vision (ICCV).Venice,Italy:IEEE,2017:2999-3007.
• [11]彭大芹,刘恒,许国良.基于多源域深度迁移学习的液晶面板缺陷检测算法[J].广东通信技术,2019,39(3):57-62.
• [12]卢颖颖,孙育.基于机器视觉的电镀件表面缺陷检测系统[J].电镀与环保,2019,39(2):59-61.
• [13]彭赶,张平,潘奕创.基于机器视觉的手机屏幕缺陷检测系统研究[J].自动化技术与应用,2018,37(9):104-107.
• [14]马丽萍,崔明,朱培逸.基于机器视觉的汽车安全带的表面缺陷检测系统[J].机床与液压,2018,46(2):134-138.
• [15]赵欣洋,蔡超鹏,王思,等.基于深度学习的不规则特征识别检测技术[J].轻工机械,2019,37(3):60-65.
• [16]胡嘉成,王向阳,刘晗.基于深度学习的连铸坯表面缺陷检测[J].上海大学学报(自然科学版),2019,25(4):445-452.
• [17]SUN R Q,ZHU X G,WU C R,et al.Not all areas are equal:Transfer learning for semantic segmentation via hierarchical region selection[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Long Beach,CA,USA:IEEE,2019:4355-4364.
• [18]金巨波,葛雷,徐秀丽.计算机数据挖掘技术的开发及其应用[J].浙江水利水电学院学报,2019,31(6):68-72.