基于深度学习的汽车轮胎受损检测方法研究  

Research on Automobiles Tire Detection Method Based on Deep Learning

作　　者：赫忠乐 刘国民 温秀兰[1] 焦良葆 唐颖 李子康 HE Zhongle;LIU Guomin;WEN Xiulan;JIAO Liangbao;TANG Ying;LI Zikang(School of Automation,Nanjing Institute of Technology,Nanjing 211167,China;Province Research Center of Intellisense Technology and System,Nanjing 211167,China)

机构地区：[1]南京工程学院自动化学院,江苏南京211167 [2]江苏省智能感知技术与装备工程研究中心,江苏南京211167

出　　处：《南京工程学院学报（自然科学版）》2024年第2期14-21,共8页Journal of Nanjing Institute of Technology(Natural Science Edition)

基　　金：江苏省智能感知技术与装备工程研究中心开放基金(ITS202103);江苏省产学研合作项目(BY2022076);江苏省大学生实践创新训练项目(202311276083Y)。

摘　　要：针对汽车轮胎受损尺度小、特征与背景相似导致检测困难等问题,提出一种基于改进YOLOX(YOLOX-O)的深度学习轮胎受损检测方法.以YOLOX模型为框架,对原YOLOX网络进行下采样层剪枝,使最小特征图尺度增加以实现对小尺度受损目标检测;为解决目标特征和背景相似导致受损难以检测的问题,在主干网络CSP结构中添加最大池化层和1×1卷积层,以保留最突出的目标特征;为平衡受损样本不均衡问题,将原网络中用于计算损失函数的BCE-loss函数替换为Focal-loss.利用Aidlux平台对模型进行压缩和优化,并将其部署到手机上实现移动端目标检测.试验结果表明:本文方法能够检测尺度小、特征和背景相似的受损轮胎,检测平均准确率均值达到90.8%,而且能够在手机端实现快速检测,适于推广应用.Aiming at the problems of small scale damaged tires and the difficulty in detection due to the similarity between features and background,a deep learning tire damage detection method based on improved YOLOX(YOLOX-O)is proposed.Using the YOLOX model as the framework,downsampling layer pruning is performed on the original YOLOX network to increase the minimum feature map scale for detecting small-scale damaged targets.To solve the difficult damage caused by similar features and backgrounds,a max pooling layer and 1×1 convolutional layer are added to the CSP structure of the backbone network to preserve the most prominent target features.To balance the problem of imbalanced target samples,the BCE-loss function used to calculate the loss function in the original network is replaced with Focal loss.The model is compressed and optimized using the Aidlux platform and deployed to mobile devices to achieve target detection on the mobile terminal.Experimental results show that the proposed method can not only detect tires of small damage scales and similar features and backgrounds with a detection average accuracy of up to 90.8%,but also achieve fast detection on mobile devices.It is suitable for promotion and application.

关 键 词：轮胎受损检测 改进YOLOX Focal loss 模型压缩优化 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]