Improved algorithm of GDT-YOLOV 3 image target detection

Yue TANG; Ge WU; Yan PIAO

doi:10.37188/YJYXS20203508.0852

您当前的位置：

首页 >

文章列表页 >

Improved algorithm of GDT-YOLOV 3 image target detection

Image Processing | 更新时间：2020-08-11

- Improved algorithm of GDT-YOLOV 3 image target detection
- Chinese Journal of Liquid Crystals and Displays Vol. 35, Issue 8, Pages: 852-860(2020)
- 作者机构：
  
  长春理工大学电子信息工程学院, 吉林长春 130022
- 作者简介：
- 基金信息：
- DOI：10.37188/YJYXS20203508.0852
  CLC： TP394.1;TH691.9
- Received：09 November 2019，
  
  Revised：26 February 2020，
  
  Accepted：26 February 2020，
  
  Published：2020-08
- 稿件说明：
移动端阅览
Yue TANG, Ge WU, Yan PIAO. Improved algorithm of GDT-YOLOV 3 image target detection[J]. Chinese journal of liquid crystals and displays, 2020, 35(8): 852-860.
DOI：

Yue TANG, Ge WU, Yan PIAO. Improved algorithm of GDT-YOLOV 3 image target detection[J]. Chinese journal of liquid crystals and displays, 2020, 35(8): 852-860. DOI： 10.37188/YJYXS20203508.0852.

摘要

为了解决视频图像中目标检测准确率低、速度慢等问题，本文提出了一种基于YOLOV3改进的目标检测方法。通过引入GIOU Loss，可解决原IOU无法直接优化的非重叠部分问题，在借鉴了密集连接网络的思想之后，将YOLOV3中的3个残差块更换为3个密集块，并结合Max Pooling加强密集连接块之间的特征传递，重新替换IOU和原网络的连接结构，检测设计出新的网络结构后，减少了参数量，增强了特征的复用与融合，最终实现了优于原方法的效果。实验结果表明：改进的GDT-YOLOV3算法与原有的算法相比，无论是在简单还是复杂交通场景中都有较优秀的检测效果，本文所提出的算法平均检出准确率高达92.77%，速度达到25.3 f/s，基本满足了实时性。此外在检测精度上，改进的GDT-YOLOV3算法要优于SSD512、YOLOV2与YOLOV3算法。

Abstract

In order to solve the problems of low accuracy and slow speed of target detection in video images

an improved YOLOV3-based object detection method is proposed. By introducing GIOU Loss

the non-overlapping part problem that the original IOU cannot directly optimize can be solved. After drawing on the idea of densely connected networks

the three residual blocks in YOLOV3 are replaced with three dense blocks

and the denseness is combined with Max Pooling to strengthen the denseness. After transferring the features between the connected blocks

and replacing the IOU and the original network to detect the connection structure

a new network structure is designed. The number of parameters is reduced

the feature reuse and fusion are enhanced

and the effect is better than the original method. The experimental results show that comparing with the original algorithm

the improved GDT-YOLOV3 algorithm has excellent results in both simple and complex traffic scenarios. The average detection accuracy of the algorithm proposed in this paper is up to 92.77%

and the speed has reached 25.3 f/s

which basically meets the real-time performance. In addition

in terms of detection accuracy

the improved GDT-YOLOV3 performs better than SSD512

YOLOV2

and YOLOV3.

关键词

Keywords

references

刘博艺 , 程杰仁 , 唐湘滟 , 等 . 复杂动态环境下运动车辆的识别方法 . 计算机科学与探索 , 2017 . 11 ( 1 ): 134 - 143 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jsjkxyts201701013 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jsjkxyts201701013 .

B Y LIU , J R CHENG , X Y TANG , 等 . Moving vehicles recognition in complex dynamic environment . Journal of Frontiers of Computer Science & Technology , 2017 . 11 ( 1 ): 134 - 143 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jsjkxyts201701013 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jsjkxyts201701013 .

H LI , K FU , M L YAN , 等 . Vehicle detection in remote sensing images using denoizing-based convolutional neural networks . Remote Sensing Letters , 2017 . 8 ( 3 ): 262 - 270 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=10.1080/2150704X.2016.1258127 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=10.1080/2150704X.2016.1258127 .

吴言枫 , 王延杰 , 孙海江 , 等 . 复杂动背景下的"低小慢"目标检测技术 . 中国光学 , 2019 . 12 ( 4 ): 853 - 865 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zggxyyygxwz201904015 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zggxyyygxwz201904015 .

Y F WU , Y J WANG , H J SUN , 等 . LSS-target detection in complex sky backgrounds . Chinese Optics , 2019 . 12 ( 4 ): 853 - 865 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zggxyyygxwz201904015 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zggxyyygxwz201904015 .

高文 , 朱明 , 贺柏根 , 等 . 目标跟踪技术综述 . 中国光学 , 2014 . 7 ( 3 ): 365 - 375 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zggxyyygxwz201403002 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zggxyyygxwz201403002 .

W GAO , M ZHU , B G HE , 等 . Overview of target tracking technology . Chinese Optics , 2014 . 7 ( 3 ): 365 - 375 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zggxyyygxwz201403002 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zggxyyygxwz201403002 .

赵永强 , 饶元 , 董世鹏 , 等 . 深度学习目标检测方法综述 . 中国图象图形学报 , 2020 . 25 ( 4 ): 629 - 654 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zgtxtxxb-a202004001 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zgtxtxxb-a202004001 .

Y Q ZHAO , Y RAO , S P DONG , 等 . Survey on deep learning object detection . Journal of Image and Graphics , 2020 . 25 ( 1 ): 629 - 654 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zgtxtxxb-a202004001 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zgtxtxxb-a202004001 .

姜文涛 , 张驰 , 张晟翀 , 等 . 多尺度特征图融合的目标检测 . 中国图象图形学报 , 2019 . 24 ( 11 ): 1918 - 1931 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zgtxtxxb-a201911007 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zgtxtxxb-a201911007 .

W T JIANG , C ZHANG , S C ZHANG , 等 . Multiscale feature map fusion algorithm for target detection . Journal of Image and Graphics , 2019 . 24 ( 11 ): 1918 - 1931 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zgtxtxxb-a201911007 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zgtxtxxb-a201911007 .

杨丹 , 戴芳 . 运动目标检测的ViBe算法改进 . 中国图象图形学报 , 2018 . 23 ( 12 ): 1813 - 1828 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zgtxtxxb-a201812004 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zgtxtxxb-a201812004 .

D YANG , F DAI . Improved ViBe algorithm for detection of moving objects . Journal of Image and Graphics , 2018 . ( 12 ): 1813 - 1828 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zgtxtxxb-a201812004 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zgtxtxxb-a201812004 .

董潇潇 , 何小海 , 吴晓红 , 等 . 基于注意力掩模融合的目标检测算法 . 液晶与显示 , 2019 . 34 ( 8 ): 825 - 833 . http://www.opticsjournal.net/Articles/abstract?aid=OJ191012000165OlRnUq http://www.opticsjournal.net/Articles/abstract?aid=OJ191012000165OlRnUq .

X X DONG , X H HE , X H WU , 等 . Object detection algorithm based on attention mask fusion . Chinese Journal of Liquid Crystals and Displays , 2019 . 34 ( 8 ): 825 - 833 . http://www.opticsjournal.net/Articles/abstract?aid=OJ191012000165OlRnUq http://www.opticsjournal.net/Articles/abstract?aid=OJ191012000165OlRnUq .

葛雯 , 史正伟 . 改进YOLOV3算法在行人识别中的应用 . 计算机工程与应用 , 2019 . 55 ( 20 ): 128 - 133 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jsjgcyyy201920018 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jsjgcyyy201920018 .

W GE , Z W SHI . Application of improved YoLOV3 algorithm in pedestrian identification . Computer Engineering and Applications , 2019 . 55 ( 20 ): 128 - 133 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jsjgcyyy201920018 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jsjgcyyy201920018 .

梁华 , 宋玉龙 , 钱锋 , 等 . 基于深度学习的航空对地小目标检测 . 液晶与显示 , 2018 . 33 ( 9 ): 793 - 800 . http://www.opticsjournal.net/Articles/abstract?aid=OJ181218000033OkRnUq http://www.opticsjournal.net/Articles/abstract?aid=OJ181218000033OkRnUq .

H LIANG , Y L SONG , F QIAN , 等 . Detection of small target in aerial photography based on deep learning . Chinese Journal of Liquid Crystals and Displays , 2018 . 33 ( 9 ): 793 - 800 . http://www.opticsjournal.net/Articles/abstract?aid=OJ181218000033OkRnUq http://www.opticsjournal.net/Articles/abstract?aid=OJ181218000033OkRnUq .

李云鹏 , 侯凌燕 , 王超 . 基于YOLOv3的自动驾驶中运动目标检测 . 计算机工程与设计 , 2019 . 40 ( 4 ): 1139 - 1144 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jsjgcysj201904038 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jsjgcysj201904038 .

Y P LI , L Y HOU , C WANG . Moving objects detection in automatic driving based on YOLOv3 . Computer Engineering and Design , 2019 . 40 ( 4 ): 1139 - 1144 . http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jsjgcysj201904038 http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=jsjgcysj201904038 .

REDMON J, DIVVALA S, GIRSHICK R, et al . You only look once: unified, real-time object detection[C]//Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE, 2016: 779-788.

REDMON J, FARHADI A. YOLO9000: better, faster, stronger[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu: IEEE, 2017: 6517-6525.

DAI J F, LI Y, HE K M, et al . R-FCN: object detection via region-based fully convolutional networks[C]//Proceedings of the 30th International Conference on Neural Information Processing Systems. Barcelona: ACM, 2016: 379-387.

HUANG G, LIU Z, VAN DER MAATEN L, et al . Densely connected convolutional networks[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu: IEEE, 2017: 2261-2269.

HE K M, ZHANG X Y, REN S Q, et al . Deep residual learning for image recognition[C]//Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE, 2016: 770-778.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

3D hand pose estimation algorithm based on cascaded features and graph convolution

Traffic sign recognition algorithm based on depthwise separable convolutions

Design of heterogeneous FPGA hardware accelerator based on CNN

Lightweight image super-resolution combining residual learning and layer attention

Related Author

Yi-lin LIN

Shan-ling LIN

Zhi-xian LIN

Jin-sheng YANG

Yan-nan YANG

Tian-jiao LI

林依林

林珊玲

Related Institution

College of Physics and Information Engineering， Fuzhou University

Fujian Science & Technology Innovation Laboratory for Optoelectronic Information of China

School of Advanced Manufacturing， Fuzhou University

School of Microelectronics, Tianjin University

Changchun Institute of Optics， Fine Mechanics and Physics， Chinese Academy of Sciences

AI问答

Address：No.3888 Dong Nanhu Road, Changchun, Jilin, China 130033 Postal code：130033
Tel：0431-86176059 Email：yjxs@ciomp.ac.cn
Technical support is provided by Beijing Founder electronics co., LTD 吉ICP备11002662号-17 京公网安备11010802024621
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰