Improved Fisheye Image Target Detection Algorithm Based on YOLOv5 Network

doi:10.3778/j.issn.1002-8331.2210-0118

Abstract

Abstract: The existing target detection algorithms have the problems of low detection accuracy and poor real-time performance on the fisheye image dataset. Therefore, this paper proposes a target detection algorithm in fisheye images using improved network YOLOv5 in chemical scenes. Firstly, aiming at the problem that there is no public chemical industry fisheye image database at present, the dataset is transformed into the image with the same distortion effect of the fisheye image by using the coordinate relationship of pixels between different types of images. Then, inserting line scan algorithm in YOLOv5s data preprocessing stage can effectively eliminate the redundant information of fisheye image. Finally, the attention mechanism scSE and dilated convolution are used to improve the lightweight network ShuffleNetV2, and the improved lightweight network is used to replace the main feature extraction network in the original YOLOv5s to reduce the size of the model while ensuring algorithm detection accuracy. The experimental results show that under the same conditions of experimental settings, the improved algorithm reduces the model size from 27.4 MB to 14.2 MB, and the detection accuracy reaches 98.46% from 97.86%.

Key words: deep learning, target detection, fisheye image, ShuffleNetV2, scSE, YOLOv5

摘要： 针对目标检测算法应用在鱼眼图像数据集上检测精准率低、算法实时性差等问题，提出了在化工场景下利用改进网络YOLOv5进行鱼眼图像中的目标检测算法。由于无公开化工场景鱼眼图像数据集，提出了利用不同类型图像间像素点的坐标关系，将数据集转换为同鱼眼图像具有相同畸变效果的图像。为消除鱼眼图像中有效区域外的冗余信息，将线扫描算法应用到YOLOv5s数据预处理阶段。为在缩减模型的同时保证算法的检测精准率，提出了采用注意力机制scSE和空洞卷积来改进轻量级网络ShuffleNetV2，并利用改进后的轻量级网络代替原YOLOv5s中主特征提取网络。实验结果表明，在实验设置相同的条件下，改进后的算法在模型从27.4?MB缩减到14.2 MB的情况下，检测精准率从97.86%提高到98.46%。

关键词: 深度学习, 目标检测, 鱼眼图像, ShuffleNetV2, scSE, YOLOv5

LYU Xiaoling, YANG Shengyue, ZHANG Minglu, LIANG Ming, WANG Junchao. Improved Fisheye Image Target Detection Algorithm Based on YOLOv5 Network[J]. Computer Engineering and Applications, 2023, 59(6): 241-250.

吕晓玲, 杨胜月, 张明路, 梁明, 王俊超. 改进YOLOv5网络的鱼眼图像目标检测算法[J]. 计算机工程与应用, 2023, 59(6): 241-250.

References

[1] 王殿伟，赵梦影，刘颖，等.改进的R-SSD全景视频图像车辆检测算法[J].计算机工程与应用，2021，57（3）：189-195.
WANG D W，ZHAO M Y，LIU Y，et al.Improved R-SSD panoramic video image vehicle detection algorithm[J].Computer Engineering and Applications，2021，57（3）：189-195.
[2] 陈斌，张漫，徐弘祯，等.基于改进YOLO v3-tiny的全景图像农田障碍物检测[J].农业机械学报，2021，52（S1）：58-65.
CHEN B，ZHANG M，XU H Z，et al.Farmland obstacle detection in panoramic image based on improved YOLOv3-tiny[J].Transactions of the Chinese Society for Agricultural Machinery，2021，52（S1）：58-65.
[3] CHEN L，ZOU Q，PAN Z Y，et al.Surrounding vehicle detection using an FPGA panoramic camera and deep CNNs[J].IEEE Transactions on Intelligent Transportation Systems，2019，21（12）：5110-5122.
[4] AKITA T，YAMAUCHI Y，FUJIYOSH H.Machine learning-based stereo vision algorithm for surround view fisheye cameras[C]//2018 21st International Conference on Intelligent Transportation Systems（ITSC）.Maui，Hawaii，USA：IEEE，2018：1103-1108.
[5] WANG T，HSIEN Y，WONG F，et al.Mask-RCNN based people detection using a top-view fisheye camera[C]//2019 International Conference on Technologies and Applications of Artificial Intelligence（TAAI），Qingdao，China，2019：1-4.
[6] KAI T，LU H，KAMIYA T.Object recognition from spherical camera images based on YOLOv3[C]//2020 20th International Conference on Control，Automation and Systems（ICCAS），Hangzhou，Zhejiang，China，2020：419-422.
[7] HOWARD A，SANDLER M，CHU G，et al.Searching for mobilenetv3[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision.South Korea：IEEE，2019：1314-1324.
[8] ZHANG X Y，ZHOU X Y，LIN M X，et al.Shufflenet：an extremely efficient convolutional neural network for mobile devices[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.New York，USA：IEEE，2018：6848-6856.
[9] MA N N，ZHANG X Y，ZHENG H T，et al.Shufflenet v2：practical guidelines for efficient cnn architecture design[C]//Proceedings of the European Conference on Computer Vision（ECCV）.Cambrige，UK：Springer，2018：116-131.
[10] LIU W，HASAN I，LIAO S C.Center and scale prediction：a box-free approach for pedestrian and face detection[EB/OL].[2019-04-05].https：//arxiv.org/pdf/1904.02948.pdf.
[11] IOANNOU Y，ROBERTSON D，CIPOLLA R，et al.Deep roots：improving CNN efficiency with hierarchical filter groups[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.New York，USA：IEEE，2017：1231-1240.
[12] ROY A G，NAVAB N，WACHINGER C.Concurrent spatial and channel ‘squeeze & excitation’ in fully convolutional networks[C]//International Conference on Medical Image Computing and Computer-Assisted Intervention.Granada，Spain：Springer，2018：421-429.
[13] 王大宇，崔汉国，陈军.鱼眼图像轮廓提取及校正研究[J].计算机工程与设计，2007，28（12）：3.
WANG D Y，CUI H G，CHEN J.Improvement of fisheye image edge extraction algorithm[J].Computer Engineering and Design，2007，28（12）：3.
[14] YU F，KOLTUN V.Multi-scale context aggregation by dilated convolutions[EB/OL].[2016-04-30].https：//arxiv.org/pdf/1511.07122.pdf.
[15] 董文轩，梁宏涛，刘国柱，等.深度卷积应用于目标检测算法综述[J].计算机科学与探索，2022，16（5）：1025-1042.
DONG W X，LIANG H T，LIU G Z，et al.Review of deep convolution applied to target detecyion algorithms[J].Journal of Frontiers of Computer Science and Technology，2022，16（5）：1025-1042.
[16] 罗安能，万海斌，司志巍，等.基于改进YOLOv5s的可回收垃圾的检测算法[J/OL].激光与光电子学进展：1-15[2022-09-12].http：//kns.cnki.net/kcms/detail/31.1690.tn.20220713.
1957.657.html.
LUO A N，WANG H B，SI Z W，et al.Detection algorithm of recyclable garbage based on improved YOLOv5s[J/OL].Laser & Optoelectronics Progress：1-15[2022-09-12].http：//kns.cnki.net/kcms/detail/31.1690.tn.20220713.1957.657.html.
[17] HAN K，WANG Y H，TIAN Q，et al.Ghostnet：more features from cheap operations[C]//2020 Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Seattle，WA，USA：IEEE，2020：1580-1589.
[18] YOU S，CHEN Z Y，LIU S D，et al.YoLite+：a lightweight multi-object detection approach in traffic scenarios[J].Procedia Computer Science，2022，199：346-353.
[19] WEN C M，WEN J，LI J H，et al.Lightweight silkworm recognition based on multi-scale feature fusion[J].Computers and Electronics in Agriculture，2022，200：107234.
[20] GUO J J，HE G H，DENG H，et al.Pigeon cleaning behavior detection algorithm based on light-weight network[J].Computers and Electronics in Agriculture，2022，199：107032.
[21] YANG L J，ZHANG P，HUANG L，et al.Sea-sky-line detection based on improved YOLOv5 algorithm[C]//2021 IEEE 2nd International Conference on Information Technology，Big Data and Artificial Intelligence（ICIBA），Nanchang，China，2021：688-694.
[22] LI A，YANG P，ZHANG J，et al.Designated mobile target detection method based on mobileNet-V3-you only look once-v5 useful in various fields e.g.pedestrian and vehicle recognition on road，product defect recognition in manufacturing process，etc，involves obtaining detection result：CN114005020-A[P].2022-02-01.
[23] DONG X D，SHUAI Y，DUAN C Q.A lightweight vehicles detection network model based on YOLOv5[J].Engineering Applications of Artificial Intelligence，2022，113：104914.
[24] LIU C，ZHU H J，GUO W，et al.EFDet：an efficient detection method for cucumber disease under natural complex environments[J].Computers and Electronics in Agriculture，2021，189：1063978.