[1]苏宇锋,边 锋,张玉堂.基于改进 YOLOv8s 算法的胸环靶弹孔检测技术[J].郑州大学学报(工学版),2024,45(05):16-22.[doi:10.13705/j.issn.1671-6833.2024.05.013]
　SU Yufeng,BIAN Feng,ZHANG Yutang.Bullet Hole Detection Technology of Chest Bitmap Based on Improved YOLOv8s Algorithm[J].Journal of Zhengzhou University (Engineering Science),2024,45(05):16-22.[doi:10.13705/j.issn.1671-6833.2024.05.013]
点击复制

基于改进 YOLOv8s 算法的胸环靶弹孔检测技术()
分享到：

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:
45
期数:
2024年05期
页码:
16-22
栏目:
出版日期:
2024-08-08

文章信息/Info

Title:
Bullet Hole Detection Technology of Chest Bitmap Based on Improved YOLOv8s Algorithm
文章编号:
1671-6833(2024)05-0016-07
作者:
苏宇锋1;  边 锋1;  张玉堂2
1. 郑州大学 机械与动力工程学院,河南 郑州 450001;2. 郑州纬达自动化科技有限公司,河南 郑州 450052
Author(s):
SU Yufeng1;  BIAN Feng1;  ZHANG Yutang2
1. School of Mechanical and Power Engineering, Zhengzhou University, Zhengzhou 450001, China; 2. Zhengzhou Weida AutomationTechnology Co. , Ltd. , Zhengzhou 450052, China
关键词:
YOLOv8s;  弹孔检测;  CBAM 注意力机制;  损失函数;  SPD-Conv
Keywords:
YOLOv8s;  bullet hole detection;  CBAM attention mechanism;  loss function;  SPD-Conv
分类号:
TP391
DOI:
10.13705/j.issn.1671-6833.2024.05.013
文献标志码:
A
摘要:
为了解决传统胸环靶弹孔检测技术在自然条件下易受光照强度、复杂背景影响的问题,设计了一种基于YOLOv8s 的改进算法。 首先,在数据集的制作过程中引入图形分割将背景与胸环靶分离,避免了复杂环境对弹孔识别精度的影响。 其次,为提升模型对弹孔的检测能力,在 C2f 中引入 CBAM 注意力机制,通过对空间和通道特征赋予不同的权值提高网络对弹孔目标的识别能力;增加检测尺度为 160× 160 的小目标输出层,减少了弹孔特征在下采样过程中的信息损失并降低弹孔漏检的概率;考虑到原有卷积层对小目标不敏感,采用 SPD-Conv 模块替换原有卷积层,提取更多的特征信息提升检测精度。 最后,将边界框损失函数更改为 WIoU 以减弱正负样本数量不均衡的影响,提高了预测框的回归精度。 在自制胸环靶数据集的实验结果表明:改进算法的准确率 P 为 96. 9%、召回率R 为 96. 4%、平均精度 mAP50 为 98. 0%,相较于原算法,分别提升 8. 8 百分点、25. 4 百分点、15. 3 百分点。 实验结果证明改进的 YOLOv8s 模型在复杂环境和密集弹孔的检测方面具有更好的性能。
Abstract:
Traditional chest bitmap bullet hole detection technology was easily affected by light intensity and complex background in natural conditions. In order to solve the proplem an improved algorithm based on YOLOv8s wasdesigned in this study. Firstly, in order to avoid the impact of complex environment on the accuracy of bullet holerecognition, graph segmentation was introduced in the process of data set production to separate the backgroundfrom the chest bitmap. Secondly, in order to improve the detection ability of the model to the bullet hole, CBAMattention mechanism was introduced into C2f, and the recognition ability of the network to the target bullet hole wasimproved by giving different weights to the spatial and channel characteristics. In order to reduce the informationloss of bullet hole characteristics in the down sampling process and reduce the probability of missing bullet holedetection, the detection scale was increased to 160×160 small target output layer. Considering that the original convolutional layer was not sensitive to small targets, the SPD-Conv module was used to replace the original convolutional layer to extract more feature information to improve detection accuracy. Finally, the loss function of the bounding box was changed to WIoU to weaken the influence of the unbalanced number of positive and negative samplesand improve the regression accuracy of the prediction box. The experimental results on the self-made chest bitmapdata set showed that the accuracy rate P of the improved algorithm was 96. 9%, the recall rate R was 96. 4%, andthe average accuracy mAP50 was 98. 0%, which were improved by 8. 8 percentage points, 25. 4 percentage points,and 15. 3 percentage points respectively, compared with the original algorithm. The experimental results showedthat the improved YOLOv8s model had better performance in the detection of complex environment and dense bulletholes.

参考文献/References:

[1] 杨子宁. 光电自动报靶系统的设计与实现[ J] . 科学技术与工程, 2007, 7(1) : 102-104, 111.

YANG Z N. Design and realization of the automatic target-scoring system [ J ] . Science Technology and Engineering, 2007, 7(1) : 102-104, 111.
[2] 冯斌, 石秀华, 康智强, 等. 声阵列与光探测组合测试弹着 点 坐 标 的 研 究 [ J ] . 压 电 与 声 光, 2012, 34(2) : 177-179.
FENG B, SHI X H, KANG Z Q, et al. Study on measurement of impact point coordinates with the combinationof acoustic array and optical detection[ J] . Piezoelectrics& Acoustooptics, 2012, 34(2) : 177-179.
[3] SUN J H, WANG S S, ZHANG F M, et al. Research onkey technologies in automatic target-reporting[ C]∥International Conference on Automatic Control and ArtificialIntelligence (ACAI 2012) . London: Institution of Engineering and Technology, 2012: 1732-1735.
[4] 黄应清, 陈晓明, 谢志宏, 等. 一种基于计算机视觉的胸环 靶 弹 孔 检 测 方 法 [ J] . 兵 器 装 备 工 程 学 报,2021, 42(11) : 157-162, 221.
HUANG Y Q, CHEN X M, XIE Z H, et al. Bullet holedetection method of chest ring target based on computervision technology [ J ] . Journal of Ordnance EquipmentEngineering, 2021, 42(11) : 157-162, 221.
[5] 宋雨, 王亚林, 杜博军, 等. 基于改进最大类间方差法的靶 板 重 孔 检 测 [ J] . 兵 工 学 报, 2022, 43 ( 4) :924-930.
SONG Y, WANG Y L, DU B J, et al. Detection of overlapped bullet holes based on improved otsu′s thresholdingmethod[J]. Acta Armamentarii, 2022, 43(4): 924-930.
[6] REDMON J, DIVVALA S, GIRSHICK R, et al. You onlylook once: unified, real-time object detection [ C]∥2016IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway: IEEE, 2016: 779-788.
[7] 院老虎, 常玉坤, 刘家夫. 基于改进 YOLOv5s 的雾天场景车辆 检 测 方 法 [ J] . 郑 州 大 学 学 报 ( 工 学 版) ,2023, 44(3) : 35-41.
YUAN L H, CHANG Y K, LIU J F. Vehicle detectionmethod based on improved YOLOv5s in foggy scene[ J] .Journal of Zhengzhou University ( Engineering Science) ,2023, 44(3) : 35-41.
[8] 薛均晓, 武雪程, 王世豪, 等. 基于改进 YOLOv4 的自然人群口罩佩戴检测方法[ J] . 郑州大学学报( 工学版) , 2022, 43(4) : 16-22.
XUE J X, WU X C, WANG S H, et al. A method onmask wearing detection of natural population based on improved YOLOv4 [ J ] . Journal of Zhengzhou University(Engineering Science) , 2022, 43(4) : 16-22.
[9] WANG C Y, BOCHKOVSKIY A, LIAO H Y M.YOLOv7: trainable bag-of-freebies sets new state-of-theart for real-time object detectors [ C] ∥2023 IEEE / CVFConference on Computer Vision and Pattern Recognition(CVPR) . Piscataway: IEEE, 2023: 7464-7475.
[10] WOO S, PARK J, LEE J Y, et al. CBAM: convolutionalblock attention module[EB / OL]. (2018-07-17) [2023-12-11]. https:∥doi. org / 10. 48550 / arXiv. 1807. 06521.
[11] SUNKARA R, LUO T. No more strided convolutions orpooling: a new CNN building block for low-resolution images and small objects [ M ] ∥ Machine Learning andKnowledge Discovery in Databases. Cham: Springer,2023: 443-459.
[12] TONG Z, CHEN Y, XU Z, et al. Wise-IoU: boundingbox regression loss with dynamic focusing mechanism[EB / OL] . (2023-01-24) [2023-12-11] . https:∥doi.org / 10. 48550 / arXiv. 2301. 10051.
[13] MNIH V, HEESS N, GRAVES A, et al. Recurrent models of visual attention[EB / OL] . (2014-01-24) [ 2023-12-11] . https:∥doi. org / 10. 48550 / arXiv. 1406. 6247.
[14] HOU Q B, ZHOU D Q, FENG J S. Coordinate attentionfor efficient mobile network design[C]∥2021 IEEE / CVFConference on Computer Vision and Pattern Recognition(CVPR) . Piscataway: IEEE, 2021: 13708-13717.
[15] HU J, SHEN L, ALBANIE S, et al. Squeeze-and-excitation networks[ EB / OL ] . ( 2017 - 09 - 05 ) [ 2023 - 12 -11] . https:∥doi. org / 10. 48550 / arXiv. 1709. 01507.
[16] ZHENG Z H, WANG P, LIU W, et al. Distance-IoUloss: faster and better learning for bounding box regression[EB / OL] . (2019-11-19) [2023-12-11] . https:∥doi. org / 10. 48550 / arXiv. 1911. 08287.
[17] ZHANG Y F, REN W Q, ZHANG Z, et al. Focal andefficient IOU loss for accurate bounding box regression[ J] . Neurocomputing,2022,506(28) :146-157.
[18] Ultralytics. YOLOv5[EB / OL] . (2020- 06- 26) [ 2023-12-11] . https:∥github. com / ultralytics/ yolov5.
[19] LIU W, ANGUELOV D, ERHAN D, et al. SSD: singleshot multibox detector [ C] ∥ECCV 2016 14th EuropeanConference. Cham: Springer,2016:21-37.

更新日期/Last Update: 2024-09-02