[1]王烨奎,曹铁勇,王 杨,等.CAMOU-YOLO:一种迷彩伪装目标检测模型[J].计算机技术与发展,2022,32(12):29-36.[doi:10. 3969 / j. issn. 1673-629X. 2022. 12. 005]
　WANG Ye-kui,CAO Tie-yong,WANG Yang,et al.CAMOU-YOLO:A Camouflaged Object Detection Model[J].,2022,32(12):29-36.[doi:10. 3969 / j. issn. 1673-629X. 2022. 12. 005]
点击复制

CAMOU-YOLO:一种迷彩伪装目标检测模型()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
32
期数:
2022年12期
页码:
29-36
栏目:
大数据与云计算
出版日期:
2022-12-10

文章信息/Info

Title:
CAMOU-YOLO:A Camouflaged Object Detection Model
文章编号:
1673-629X(2022)12-0029-08
作者:
王烨奎1; 2 ; 曹铁勇1 ; 王 杨1 ; 方 正1 ; 刘亚九2 ; 郑云飞1; 3; 4 ; 付炳阳1
1. 陆军工程大学 指挥控制工程学院,江苏 南京 210007;
2. 31401 部队,吉林 长春 130000;
3. 陆军炮兵防空兵学院,江苏 南京 211100;
4. 偏振光成像探测技术安徽省重点实验室,安徽 合肥 230031
Author(s):
WANG Ye-kui1; 2 ; CAO Tie-yong1 ; WANG Yang1 ; FANG Zheng1 ; LIU Ya-jiu2 ; ZHENG Yun-fei1; 3; 4 ; FU Bing-yang1
1. School of Command and Control Engineering,Army Engineering University of PLA,Nanjing 210007,China;
2. Unit 31401,Changchun 130000,China;
3. The Army Artillery and Defense Academy of PLA,Nanjing 211100,China;
4. Key Laboratory of Polarization Imaging Detection Technology,Hefei 230031,China
关键词:
迷彩伪装; 目标检测; YOLO; 深度可分离卷积; 动态注意力
Keywords:
camouflage; object detection; YOLO; depth separable convolution; dynamic attention mechanism
分类号:
TP751
DOI:
10. 3969 / j. issn. 1673-629X. 2022. 12. 005
摘要:
由于迷彩伪装目标与所处背景高度融合,现有深度目标检测模型在此类目标上的检测效果并不出众。 为提升对迷彩伪装目标的检测精度,以 YOLOv5s 模型为基础,提出了 CAMOU-YOLO———一种结合深度可分离卷积和动态注意力的迷彩伪装目标检测模型。 针对迷彩伪装目标特征提取难的问题,结合深度可分离卷积与残差结构设计了新的特征提取模块,并对原有骨干网络进行改进,在增强提取能力的同时,减小了模型的参数量;针对迷彩伪装目标定位难度大的问题,在聚合网络中引入动态注意力机制,强化了模型的空间感知能力,使模型对迷彩伪装目标的定位更加精准。 在一种公开的迷彩数据集上进行实验,CAMOU-YOLO 的 mAP@ 0. 5、mAP@ 0. 75 和 mAP@ 0. 5:0. 95 指标较原始模型提高了 3. 2% 、5. 1% 、2. 3% ,在大、中、小目标上的召回率分别提高了 4. 1% 、2. 7% 、1. 2% ,且参数量降低了 9. 7% ;对比其他 7 种检测算法,CAMOU-YOLO 在检测精度上亦具有优势,验证了所提模型对迷彩伪装目标检测任务的有效性。
Abstract:
Since the camouflaged object being highly fused with the background, the existing object detection models based on deeplearning have poor detection performance on such objects. Based on the YOLOv5s model, a camouflage object detection modelcombining depth separable convolution and dynamic attention is proposed to improve the detection accuracy. Aiming at the difficulty offeature extraction of camouflaged object,combined with the structure of depth separable convolution and residual,a new feature extractionmodule is designed,and the original backbone network is improved,which not only enhances the extraction ability but also reduces thenumber of parameters of the model. Aiming at the difficulty of positioning camouflaged object, the dynamic attention mechanism isintroduced into the aggregation network to strengthen the spatial perception ability of the model and make the positioning of thecamouflage object more accurate. Experimenting on a public camouflage dataset,CAMOU-YOLO’s mAP@ 0. 5,mAP@ 0. 75 and mAP@ 0. 5:0. 95 are 3. 2% ,5. 1% ,2. 3% higher than the original model,and recall rates on large,medium,and small objects are 4. 1% ,2. 7% ,1. 2% higher, respectively. At the same time, the parameter quantity is reduced by 9. 7% . Compared with the other sevendetection algorithms,CAMOU-YOLO also has advantages in detection accuracy,which verifies the validity of the proposed model forcamouflage target detection tasks.

相似文献/References:

[1]刘晓明 李毓蕙 高燕 郑华强.基于目标区域清晰显示的H.264编码策略[J].计算机技术与发展,2010,(06):29.
　LIU Xiao-ming,LI Yu-hui,GAO Yan,et al.A Coding Strategy of H.264 Based on High-definition Display of Target Region[J].,2010,(12):29.
[2]刘翔 吴谨 祝愿博 康晓晶.基于视频序列的目标检测与跟踪技术研究[J].计算机技术与发展,2009,(11):179.
　LIU Xiang,WU Jin,ZHU Yuan-bo,et al.A Study of Object Detecting and Tracking Based on Video Sequences[J].,2009,(12):179.
[3]曙光 张超 蔡则苏.基于改进的混合高斯模型的目标检测方法[J].计算机技术与发展,2012,(07):60.
　SHU Guang,ZHANG Chao,CAI Ze-su.Target Detection Method Based on Improved Gaussian Mixture Model[J].,2012,(12):60.
[4]刘洁,李目,周少武.一种混沌混合粒子群优化RBF神经网络算法[J].计算机技术与发展,2013,(08):181.
　LIU Jie[],LI Mu[],ZHOU Shao-wu[].An Algorithm of Chaotic Hybrid Particle Swarm Optimization Based on RBF Neural Network[J].,2013,(12):181.
[5]蒋翠清,孙富亮,吴艿芯. 基于相对欧氏距离的背景差值法视频目标检测[J].计算机技术与发展,2015,25(01):37.
　JIANG Cui-qing,SUN Fu-liang,WU Nai-xin. Video Object Detection of Background Subtraction Method Based on Relative Euclidean Distance[J].,2015,25(12):37.
[6]卢官明,衣美佳. 步态识别关键技术研究[J].计算机技术与发展,2015,25(07):100.
　LU Guan-ming,YI Mei-jia. Research on Critical Techniques in Gait Recognition[J].,2015,25(12):100.
[7]高翔,朱婷婷,刘洋. 多摄像头系统的目标检测与跟踪方法研究[J].计算机技术与发展,2015,25(07):221.
　GAO Xiang,ZHU Ting-ting,LIU Yang. Research of Target Detection and Tracking Method for Multi-camera System[J].,2015,25(12):221.
[8]章文洁[][],黄旻[],张桂峰[]. 滤光片多光谱成像中运动目标场景误配准修正[J].计算机技术与发展,2016,26(01):18.
　ZHANG Wen-jie[][],HUANG Min[],ZHANG Gui-feng[]. Misregistration Correction for Moving Object Scene in Filter-type Multispectral Imaging[J].,2016,26(12):18.
[9]施泽浩,赵启军.基于全卷积网络的目标检测算法[J].计算机技术与发展,2018,28(05):55.[doi:10.3969/j.issn.1673－629X.2018.05.013]
　SHI Ze-hao,ZHAO Qi-jun.Object Detection Algorithm Based on Fully Convolutional Neural Network[J].,2018,28(12):55.[doi:10.3969/j.issn.1673－629X.2018.05.013]
[10]张夏清,茅耀斌. 一种改进的ViBe背景提取算法[J].计算机技术与发展,2016,26(07):36.
　ZHANG Xia-qing,MAO Yao-bin. An Improved ViBe Background Generation Method[J].,2016,26(12):36.
[11]韩 彤,曹铁勇,郑云飞,等.迷彩伪装目标检测的视觉特征偏好研究[J].计算机技术与发展,2023,33(12):193.[doi:10. 3969 / j. issn. 1673-629X. 2023. 12. 027]
　HAN Tong,CAO Tie-yong,ZHENG Yun-fei,et al.Research on Visual Feature Bias of Camouflaged Object Detection[J].,2023,33(12):193.[doi:10. 3969 / j. issn. 1673-629X. 2023. 12. 027]

更新日期/Last Update: 2022-12-10