基于深度学习的汽车液晶屏图像识别算法优化  

Optimization of image recognition algorithm for automotive LCD screen based on deep learning

作　　者：陈磊 聂嘉城 王源 马汶锴 宋朋宇 Chen Lei;Nie Jiacheng;Wang Yuan;Ma Wenkai;Song Pengyu(CATARC Component Technology(Tianjin)Co.,Ltd.,Tianjin 300300,China)

机构地区：[1]中汽零部件技术(天津)有限公司,天津300300

出　　处：《汽车知识》2024年第12期116-118,共3页AUTOMOTIVE KNOWLEDGE

摘　　要：为了提升汽车内饰液晶屏数字图像识别的准确性和效率,增加驾驶人员的舒适性,本文提出并优化了基于深度学习的SCT-ResNet50算法,结合自注意力与通道注意力机制,增强对图像特征的提取与识别能力。通过数据预处理与增强技术,有效提升模型在复杂工业环境中的泛化能力。实验结果表明,SCT-ResNet50在高分辨率图像识别中表现出极高的精度和稳定性,推动液晶屏图像识别在汽车内饰智能制造中的广泛应用。To improve the accuracy and efficiency of digital image recognition for automotive interior liquid crystal screens,enhance the comfort of drivers,this paper proposes and optimizes the SCT-ResNet50 algorithm based on deep learning,combining self-attention and channel attention mechanisms to enhance the ability to extract and identify image features.By using data preprocessing and enhancement techniques,the model's generalization ability in complex industrial environments is effectively improved.Experimental results show that the SCT-ResNet50 achieves very high accuracy and stability in high-resolution image recognition,promoting the widespread application of liquid crystal screen image recognition in automotive interior intelligent manufacturing.

关 键 词：深度学习 图像识别 SCT-ResNet50 汽车液晶屏 

分 类 号：U461[机械工程—车辆工程]