改进FSSD的差速器壳体表面缺陷实时检测算法

doi:10.3778/j.issn.1002-8331.2201-0218

摘要/Abstract

摘要： 针对传统检测方法对于汽车差速器壳体表面小目标缺陷的误检和漏检问题，提出了一种改进的FSSD_MobileNet缺陷检测模型。该模型将FSSD（feature fusion single shot multibox detector）算法的基础骨干网络VGG16替换成轻量级MobileNet网络，构建了一种高效的特征融合结构并调整了默认框的尺寸，进一步提升对小目标缺陷的检测能力。同时使用RMSProp（root mean square propagate）梯度下降算法来优化损失函数，加快了模型的收敛速度。实验结果表明，改进后的FSSD_MobileNet模型的mAP为96.7%，相比于改进前提升了16.2个百分点。在保持较高检测精度的同时，检测速度达到了191?FPS，高于目前单阶段算法中速度较快的YOLOv5s网络，相较于传统的SSD（single shot multibox detector）和FSSD分别提升了94?FPS和102?FPS，同时模型较为精简，能够更好地满足实际生产中对准确性和实时性的综合要求。

关键词: 目标检测, 特征融合, FSSD, 深度可分离卷积, 小目标缺陷

Abstract: Aiming at the traditional detection methods for the problems of false detection and missed detection of small target defects on the surface of automobile differential housing, an improved FSSD_MobileNet defect detection model is proposed. The model replaces the basic backbone network VGG16 of the FSSD（feature fusion single shot multibox detector） algorithm with the lightweight MobileNet network, constructs an efficient feature fusion structure. The model also adjusts the size of the default box, which further improves the detection ability of small target defects. Meanwhile, the RMSProp（root mean square propagate） gradient descent algorithm is used to optimize the loss function, which speeds up the convergence speed of the model. The experimental results show that the mAP of the improved FSSD_MobileNet model is 96.7%, which is 16.2 percentage points higher than that before the improvement. While maintaining high detection accuracy, the detection speed has reached 191 FPS, which is higher than the faster YOLOv5s network in the current one-stage algorithm. Compared with the traditional SSD（single shot multibox detector） and FSSD, the detection speed has increased by 94 FPS and 102 FPS respectively. At the same time, the model is relatively simplified, which can better meet accuracy and real-time requirements comprehensively in actual production.

Key words: target detection, feature fusion, FSSD, depthwise separable convolution, small target defects

王国东, 唐金亮, 陈特欢, 崔杰. 改进FSSD的差速器壳体表面缺陷实时检测算法[J]. 计算机工程与应用, 2022, 58(22): 262-270.

WANG Guodong, TANG Jinliang, CHEN Tehuan, CUI Jie. Real-Time Detection Algorithm for Differential Housing Surface Defects Based on Improved FSSD[J]. Computer Engineering and Applications, 2022, 58(22): 262-270.

参考文献

[1] 樊智涛，刘洋.基于WorkBench的差速器壳体有限元分析及轻量化设计[J].军事交通学院学报，2020，22（2）：91-95.
FAN Z T，LIU Y.Finite element analysis and lightweight design of differential housing based on WorkBench[J].Journal of the Academy of Military Transportation，2020，22（2）：91-95.
[2] 李明山，韩清鹏，张天宇，等.改进SSD的安全帽检测方法[J].计算机工程与应用，2021，57（8）：192-197.
LI M S，HAN Q P，ZHANG T Y，et al.Safety helmet detection method of improved SSD[J].Computer Engineering and Applications，2021，57（8）：192-197.
[3] GIRSHICK R，DONAHUE J，DARRELLAND T，et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]//2014 IEEE International Conference Computer Vision and Pattern Recognition，Columbus，Jun 23-28，2014：580-587.
[4] GIRSHICK R，Fast R-CNN[C]//2015 IEEE International Conference on Computer Vision，Santiago，Dec 7-13，2015：1440-1448.
[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] REDMON J，DIVVALA S，GIRSHICK R，et al.You only look once：unified，real-time object detection[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition，Las Vegas，Jun 27-30，2016：779-788.
[7] LIU W，ANGUELOV D，ERHAN D，et al.SSD：single shot multibox detector[J/OL].（2016-12-29）[2020-06-07].https：//arxiv.org/abs/1512.02325.
[8] 辛鹏，许悦雷，唐红，等.全卷积网络多层特征融合的飞机快速检测[J].光学学报，2018，38（3）：0315003.
XIN P，XU Y L，TANG H，et al.Fast aircraft detection based on multi-layer feature fusion of full convolutional network[J].Acta Optica Sinica，2018，38（3）：0315003.
[9] 马露，裴伟，朱永英，等.基于深度学习的跌倒行为识别[J].计算机科学，2019，46（9）：106-112.
MA L，PEI W，ZHU Y Y，et al.Fall behavior recognition based on deep learning[J].Computer Science，2019，46（9）：106-112.
[10] FU C Y，LIU W，RANGA A，et al.DSSD：deconvolutional single shot detector[J/OL].（2017-01-23）[2020-06-07].https：//arxiv.org/abs/1701.06659.
[11] 王永祥.基于机器视觉的差速器壳体识别技术研究[D].济南：山东大学，2020.
WANG Y X.Research on differential case recognition technology based on machine vision[D].Jinan：Shandong University，2020.
[12] 陈辉，朱易坤，周林娟，等.汽轮机铸件汽缸疏松缺陷超声波检测研究[J].机械，2018，45（10）：77-80.
CHEN H，ZHU Y K，ZHOU L J，et al.Research on ultrasonic inspection of turbine cylinder porosity defects[J].Machinery，2018，45（10）：77-80.
[13] FERGUSON M，AK R，LEE Y T T，et al.Automatic localization of casting defects with convolutional neural networks[C]//2017 IEEE International Conference on Big Data，Boston，Dec 11-14，2017：1726-1735.
[14] SIMONYAN K，ZISSERMAN A.Very deep convolutional networks for large-scale image recognition[J/OL].（2015-04-10）[2019-12-12].https：//arxiv.org/abs/1409.1556.
[15] 周凡，朴燕，秦晓伟.基于密集模块与特征融合的SSD目标检测算法[J].计算机工程与应用，2020，56（16）：105-111.
ZHOU F，PIAO Y，QIN X W.SSD target detection algorithm based on dense module and feature fusion[J].Computer Engineering and Applications，2020，56（16）：105-111.
[16] LI Z X，ZHOU F Q.FSSD：feature fusion single shot multibox detector[J/OL].（2018-05-17）[2020-06-07].https：//arxiv.org/abs/1712.00960.
[17] LIN T Y，DOLLAR P，GIRSHICK R，et al.Feature pyramid networks for object detection[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition，Honolulu，Jul 21-26，2017：936-944.
[18] 高洁，王战红，刘纲.基于FSSD的微光烟雾检测方法[J].电子测量技术，2021，44（5）：123-128.
GAO J，WANG Z H，LIU G.Low-light smoke detection method based on FSSD[J].Electronic Measurement Technology，2021，44（5）：123-128.
[19] HOWARD A，ZHU M L，CHEN B，et al.MobileNets：efficient convolutional neural networks for mobile vision applications[J/OL].（2019-04-17）[2020-06-07].https：//arxiv.org/abs/1704.04861.
[20] CAESAR H，UIJLINGS J，FERRARI V.COCO-stuff：thing and stuff classes in context[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition，Salt Lake City，Jun 18-23，2018：1209-1218.
[21] EVERINGHAM M，GOOL L，WILLIAMS C K I，et al.The pascal visual object classes（VOC） challenge[J].International Journal of Computer Vision，2020，88（2）：303-338.
[22] TIELEMAN T，HINTON G.Lecture 6.5-RMSProp：divide the gradient by a running average of its recent magnitude[J].COURSERA：Neural Networks for Machine Learning，2012，4（2）：26-31.