基于YOLOv8的破片穿靶效应检测

荣江伟; 曹红松; 褚文博; 王晴岳; 段明清; 张浩; 何晓萍

doi:10.12382/bgxb.2025.0646

您当前的位置：

首页 >

文章列表页 >

基于YOLOv8的破片穿靶效应检测

更新时间：2026-05-06

- 基于YOLOv8的破片穿靶效应检测
- Detection of Fragment Penetration Effect Based on YOLOv8
- 兵工学报 2026年47卷第4期页码：250646
- 作者机构：
  
  中北大学机电工程学院，山西太原 030051
- 作者简介：
  
  通信作者邮箱：chs@nuc.edu.cn
- 基金信息：
  
  国家自然科学基金项目(62205307)
- DOI：10.12382/bgxb.2025.0646
  中图分类号： TJ061;TJ413.1
- 收稿：2025-07-14，
  
  网络首发：2026-01-27，
  
  纸质出版：2026-04
- 稿件说明：
移动端阅览
荣江伟, 曹红松, 褚文博, 等. 基于YOLOv8的破片穿靶效应检测[J]. 兵工学报, 2026,47(4):250646.

RONG Jiangwei, CAO Hongsong, CHU Wenbo, et al. Detection of Fragment Penetration Effect Based on YOLOv8[J]. Acta Armamentarii, 2026, 47(4): 250646.
荣江伟, 曹红松, 褚文博, 等. 基于YOLOv8的破片穿靶效应检测[J]. 兵工学报, 2026,47(4):250646. DOI： 10.12382/bgxb.2025.0646.

RONG Jiangwei, CAO Hongsong, CHU Wenbo, et al. Detection of Fragment Penetration Effect Based on YOLOv8[J]. Acta Armamentarii, 2026, 47(4): 250646. DOI： 10.12382/bgxb.2025.0646.

摘要

针对靶场工作中人工采集数据效率低、精度差的问题，提出基于图像识别的自动检靶技术，以取代传统人工破片分布统计方法。破片靶孔存在形态多样、分布密集、背景复杂等特性，叠加靶板锈蚀和风化等噪声干扰，导致边缘轮廓分割精度受限，为此提出金属靶破片靶孔分割模型（YOLOv8-Target plate Fragment perforation Segmentation，YOLOv8-TFS）。在YOLOv8n-seg模型的基础上，增加微小靶孔检测层以提升模型对于不同尺寸靶孔的特征提取能力；优化跨阶段链接的特征融合网络结构；引入双通路自适应特征加权特征融合模块，强化特征表达并抑制背景噪声；设计多路径感受野注意力分割头，提高模型对靶孔特征的整合输出能力。试验结果表明，YOLOv8-TFS模型在自制数据集上的掩码精确率、召回率和mAP@0.5%分别达到85.2%、74.3%和73.8%，较原始模型分别提升11.6%、10.9%、9.0%，有效提高了靶孔分割的准确性。基于分割结果构建的自动检靶系统，通过空间矩计算与坐标转换实现靶孔面积及质心坐标的精确解算。与人工检靶相比，数量平均绝对偏差为1.07个，面积偏差控制在5%以内，验证了检靶方法的可靠性与高精度特性，为战斗部毁伤效能评估提供了高效的技术支撑。

Abstract

Aiming at the problems of low efficiency and poor accuracy of manual data collection in shooting range

an automatic target detection technology based on image recognition is proposed to replace the traditional manual statistical method of fragment distribution. The target holes have characteristics such as diverse shapes

dense distribution

and complex background

and are further influenced by the rust and weathering of target plate

which limits the accuracy of edge contour segmentation. Therefore

this paper proposes a target plate fragment perforation segmentation model (YOLOv8-Target plate Fragment perforation Segmentation

YOLOv8-TFS) .Based on YOLOv8 model

a micro-target hole detection layer is added to enhance the model's ability to extract the features of different-sized target holes. The feature fusion network structure of cross-stage connections is optimized

and a dual-path adaptive feature weighting feature fusion module is introduced to strengthen the feature expression and suppress the background noise. A multi-path receptive field attention segmentation head is designed to improve the model's ability to integrate and output the target hole features. Experimental results show that the mask precision

recall rate

and mAP@0.5% of YOLOv8-TFS model on the self-made dataset reach 85.2%

74.3%

and 73.8%

respectively

which are 11.6%

10.9%

and 9.0% higher than those of the original model

effectively improving the accuracy of target hole segmentation. The automatic target detection system constructed based on the segmentation results accurately calculates the area and centroid coordinates of the target holes through spatial moment calculation and coordinate transformation. Compared with manual target detection

the average absolute deviation in quantity is 1.07

and the area deviation is controlled within 5%.This verifies the reliability and high-precision characteristics of the target detection method

and providing efficient technical support for the assessment of warhead damage effectiveness.

关键词

Keywords

references

李付刚，梁民族，李翔宇，等.杀伤爆破战斗部杀伤威力的多目标优化[J].兵工学报，2021，42（增刊1）：11-21.

LI F G, LIANG M Z, LI X Y, et al. Multi-objective optimization of lethal power of blast-fragmentation warhead [J]. Acta Armamentarii, 2021, 42 (S1): 11-21. (in Chinese)

武江鹏，乔明军，闫振纲，等.战斗部破片场参数测试技术发展综述[J].兵器装备工程学报，2019，40（5）：105-109.

WU J P, QIAO M J, YAN Z G, et al. Overview of testing technology for warhead fragments characterization [J]. Journal of Ordnance Equipment Engineering, 2019, 40 (5): 105-109. (in Chinese)

王林，刘永付，李晓辉，等.大当量杀伤战斗部破片飞散特性试验方法研究[J].弹箭与制导学报，2012，32（6）：68-70.

WANG L, LIU Y F, LI X H, et al. Study on the scatter of high capacity head fragments [J]. Journal of Projectiles, Rockets, Missiles and Guidance, 2012, 32 (6): 68-70. (in Chinese)

武锦辉，刘吉.战斗部静爆场破片参数测试技术发展现状[J].兵器装备工程学报，2019，40（10）：104-110.

WU J H, LIU J.Development status of testingthe parameters of warhead fragment in the static explosive field [J]. Journal of Ordnance Equipment Engineering, 2019, 40 (10): 104-110. (in Chinese)

李明富，何宸，凌艳，等.图像识别技术在战斗部静爆试验分析中的应用[J].四川兵工学报，2014，35（6）：101-104.

LI M F, HE C, LING Y, et al. Application of image recognition technology in the warhead explosion test [J]. Sichuan Ordnance Journal, 2014, 35 (6): 101-104. (in Chinese)

任杰，蒋海燕，姬建荣，等.战斗部破片场飞散分布测试技术研究综述[J].兵器装备工程学报，2024，45（12）：122-131.

REN J, JIANG H Y, JI J R, et al. Research summary of test technologies for sacttering distribution of warhead fragment field [J]. Journal of Ordnance Equipment Engineering, 2024, 45 (12): 122-131. (in Chinese)

张斌，李佳潞，赵冬娥，等.基于小波滤波及相关分析的激光光幕破片测速信号数据处理[J].兵工学报，2016，37（3）：489-495.

ZHANG B, LI J L, ZHAO D E, et al. Signal processing of laser screen fragments velocity measurement based on wavelet transform and correlation analysis [J]. Acta Armamentarii, 2016, 37 (3): 489-495. (in Chinese)

陈瑞，倪晋平.基于正交试验的光电立靶光幕阵列结构参数优化方法[J].兵工学报，2017，38（11）：2234-2239.

CHEN R, NI J P.Optimization method of light-screen-array structure parameters of photoelectric target based on orthogonal test [J]. Acta Armamentarii, 2017, 38 (11): 2234 - 2239. (in Chinese)

黄国伟.基于声学定位的激波报靶系统设计与优化[D].杭州：浙江大学，2020.

HUANG G W.Design and optimization of shock wave scoring system based on acoustic localization [D]. Hangzhou: Zhejiang University, 2020. (in Chinese)

于正亮，肖曦.开放式声电定位报靶系统[J].四川兵工学报，2015，36（2）：15-17，40.

YU Z L, XIAO X.Research on open acoustic scoring system [J]. Sichuan Ordnance Journal, 2015, 36 (2): 15 - 17, 40. (in Chinese)

张泉斌.基于机器视觉的自动报靶系统设计与关键技术研究[D].兰州：兰州大学，2020.

ZHANG Q B.Design and key technologies research of automatic scoring system based on machine vision [D]. Lanzhou: Lanzhou University, 2020. (in Chinese)

罗杰，张之明，付凯城，等.基于融合技术的小波变换的弹孔检测算法[J].电光与控制，2016，23（9）：19-23.

LUO J, ZHANG Z M, FU K C, et al. Bullet hole detection algorithm based on wavelet transform using fusion rules [J]. Electronics Optics & Control, 2016, 23 (9): 19-23. (in Chinese)

张旭光，孙巍，韩广良，等.一种弹孔自动识别算法的研究[J].光学精密工程，2005，13（6）：747-753.

ZHANG X G, SUN W, HAN G L, et al. An auto matic recognition algorithm for bullet holes [J]. Optics and Precision Engineering, 2005, 13 (6): 747-753. (in Chinese)

杜红棉，葛竹，孟晓山.基于图像分割的效应靶自动报靶系统[J].科学技术与工程，2018，18（7）：191-196.

DU H M, GE Z, MENG X S.Automatic target system of effect targets based on image segmentation [J]. Science Technology and Engineering, 2018, 18 (7): 191-196. (in Chinese)

何淼，王浩，黄勇，等.基于图像处理的野外金属破片靶测量研究[J].弹道学报，2021，33（4）：83-90.

HE M, WANG H, HUANG Y, et al. Research on field metal fragment target measurement based on image processing [J]. Journal of Ballistics, 2021, 33 (4): 83-90. (in Chinese)

BI Y, XUE B, MESEJO P, et al. A survey on evolutionary computation for computer vision and image analysis:past, present, and future trends [J]. IEEE Transactions on Evolutionary Computation, 2023, 27 (1): 5-25.

GU W C, BAI S, KONG L X. A review on 2D instance segmentation based on deep neural networks [J]. Image and Vision Computing, 2022, 120:104401.

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

YI H, LIU B, ZHAO B, et al. Small object detection algorithm based on improved YOLOv8 for remote sensing [J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2024, 17:1734-1747.

赵海丽，包大泱，张从豪，等.基于改进SDU-YOLOv8的军事飞机目标检测算法[J].兵工学报，2026，47（1）：250294.

ZHAO H L, BAO D Y, ZHANG C H, et al. Military aircraft object detection algorithm based on improved SDU-YOLOv8 [J]. Acta Armamentarii, 2026, 47 (1): 250294. (in Chinese)

HUSSAIN M. YOLOv1 to YOLOv8, the rise of YOLO and its complementary nature toward digital manufacturing and industrial defect detection [J]. Machines, 2023, 11 (7): 677.

LÜ Z L, ZHAO Z Q, XIA K W, et al. Steel surface defect detection based on MobileViTv2 and YOLOv8 [J]. The Journal of Supercomputing, 2024, 80 (13): 18919-18941.

SUN S Z, MO B, XU J W, et al. Multi-YOLOv8: an infrared moving small object detection model based on YOLOv8 for air vehicle [J]. Neurocomputing, 2024, 588:127685.

ZHANG J H, JIAO H, LI G Y, et al. YIP:optimized YOLOv11seg for accurate intramuscular fat percentage prediction in pigs using non-invasive ultrasound imaging [J]. Science China Life Sciences, 2026, 69:683-686.

刘婷婷，宋家友.融合多模态特征与改进YOLOv12的光伏电池板组件缺陷检测[J/OL].计算机工程与应用，2025（2025-10-30）[2025-11-03].https:∥link.cnki.net/urlid/11.2127.tp.20251030.1414.004.

LIU T T, SONG J Y.Defect detection of photovoltaic panel components integrating multimodal features and improved YOLOv12 [J/OL].Computer Engineering and Applications, (2025-10-30)[2025-11-03].https: ∥link. cnki. net/urlid/11.2127.tp.20251030.1414.004. (in Chinese)

何性顺，姬建荣，魏巍，等.一种基于激光扫描的获取靶板穿孔密度的方法：CN107657609[P].2018-02-02.

HE X S, JI J R, WEI W, et al. A method for obtaining perforation density of target plate based on laser scanning: CN107657609 [P].2018-02-02. (in Chinese)

浏览量

下载量

CNKI被引量

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于实例分割与光流的动态环境SLAM