[1]唐海涛,吴果林,范广义,等.融合 SIFT 和级联分类器的特种车辆自动检测识别[J].计算机技术与发展,2023,33(09):182-189.[doi:10. 3969 / j. issn. 1673-629X. 2023. 09. 027]
　TANG Hai-tao,WU Guo-lin,FAN Guang-yi,et al.Automatic Detection and Recognition of Special Vehicles Incorporating SIFT and Cascade Classifier[J].,2023,33(09):182-189.[doi:10. 3969 / j. issn. 1673-629X. 2023. 09. 027]

点击复制

融合 SIFT 和级联分类器的特种车辆自动检测识别()

分享到：

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

卷:

33

期数:

2023年09期

页码:

182-189

栏目:

人工智能

出版日期:

2023-09-10

文章信息/Info

Title:

Automatic Detection and Recognition of Special Vehicles Incorporating SIFT and Cascade Classifier

文章编号:

1673-629X(2023)09-0182-07

作者:

唐海涛¹; 2 ; 吴果林¹; 2 ; 范广义³ ; 陈迪三¹

1. 桂林航天工业学院 理学院,广西 桂林 541004;

2. 桂航大数据技术应用研究中心,广西 桂林 541004;

3. 南京高精齿轮集团有限公司,江苏 南京 210000

Author(s):

TANG Hai-tao1; 2 ; WU Guo-lin1; 2 ; FAN Guang-yi3 ; CHEN Di-san1

1. School of Science,Guilin University of Aerospace Technology,Guilin 541004,China;

2. Research Center for Big Data Technology Application in Guat,Guilin 541004,China;

3. Nanjing High-Speed & Accurate Gear Group Co. ,Ltd. ,Nanjing 210000,China

关键词:

尺度不变特征变换; KMeans; RF-RBF 级联分类器; K 折交叉验证; 特种车辆

Keywords:

scale invariant feature transformation; KMeans; RF-RBF cascade classifier; K-fold cross-validation; special vehicle

分类号:

TP391

DOI:

10. 3969 / j. issn. 1673-629X. 2023. 09. 027

摘要:

针对特定场景中特种车辆因多环境影响因素下数据不均衡、检测精度和识别准确率低的问题,提出一种融合尺度不变特征变换( Scale-Invariant Feature Transform,SIFT) 和级联分类器的特种车辆自动检测及识别预测方法。 首先,图像预处理后运用 SIFT 特征提取图像主体区域特征点及特征描述子;其次,结合 SIFT 特征点的密度调整优化

算法实现目标车辆检测;最后,运用 KMeans 聚类算法获得目标检测框中 SIFT 特征描述子的中心聚类点,生成表征目标主体图像的 128 维特征描述子,并最终输入 RF-RBF

( Random Forest-Radial Basis Function) 级联分类器进行学习并识别预测。该文均采用 K 折交叉验证方法保证模型的稳定性和可靠性。实验结果表明,在特定场景下特种车辆目标检测获得了 75.47% 平均交并比,级联分类器在特种车辆识别的综合准确率、精确率、召回率、F1 -Score 值以及 FPS 值分别为 87.35% 、88.17% 、97.27% 、92. 38% 以及 21。 进一步验证融合 SIFT 和级联分类器模型具有较好的自动化检测准确性和识别分类能力。

Abstract:

In order to overcome the unbalanced data and low detection and recognition accuracy of special vehicles in specific scenescaused by a number of environmental influencing factors,a Scale-Invariant Feature Transform ( SIFT) and cascade classifier method forspecial vehicle identification and recognition prediction is proposed. Firstly,the SIFT features are used to extract the feature points andfeature descriptors of the image subject area after image pre-processing. Secondly,the density adjustment optimization algorithm of SIFTfeature points is combined to realize the target vehicle detection. Finally,the KMeans clustering algorithm is used to obtain the centralclustering points of the SIFT feature descriptors in the target detection frame to generate 128 - dimensional feature descriptorscharacterizing the target subject image, and finally input RF - RBF ( Random Forest - Radial Basis Function ) cascade classifier forlearning,identifying and prediction. All K-fold cross-validation methods are used to ensure the stability and reliability of the model. Theexperimental results show that 75. 47% average cross - comparison ratio is obtained for special vehicle target detection in specificscenarios,and the combined accuracy,precision,recall,F1 -Score and FPS values of the cascade classifier in special vehicle recognitionare 87. 35% ,88. 17% ,97. 27% ,92. 38% and 21,respectively. Such model has better automatic detection accuracy and recognition classification ability.

常用功能