Scholarlink

To address the numerous challenges associated with the current vehicle defrosting performance verification process, which relies heavily on traditional manual methods, this paper proposes an intelligent verification method based on machine vision. The proposed method leverages a hybrid neural network architecture, utilizing the Swin-Transformer (Liu, Z. et al. “Swin Transformer: Hierarchical Vision Transformer using Shifted Windows”. IEEE/CVF International Conference on Computer Vision (ICCV), 202) to fully retain and utilize the self-attention mechanism of the Transformer (Vaswani, Ashish, Shazeer, Noam, Parmar, Niki, Uszkoreit, Jakob, Jones, Llion, Gomez, Aidan N, Kaiser, Łukasz, and Polosukhin, Illia. “Attention is All You Need”. Advances in Neural Information Processing Systems, 2017: 5998–6008.; Fu, Jun; Liu, Jing; Tian, Haijie; Li, Yong; Bao, Yongjun; Fang, Zhiwei; Lu, Hanqing. “Dual Attention Network for Scene Segmentation”. IEEE Conference on Computer Vision and Pattern Recognition, 2019, Pages 3146–315; Tang, Jinhui, and Lu, Hanqing. “Adaptive Context Network for Scene Parsing”. IEEE/CVF International Conference on Computer Vision, 2019: 6748–6757.;) while incorporating the local perception capabilities of Convolutional Neural Networks (CNNs). This combination enables the automatic extraction and intelligent analysis of various frost layer characteristics at different time points. The method supports comprehensive data recording and the intelligent analysis of key metrics, such as defrosting rates, throughout the verification process, effectively eliminating the influence of human subjectivity and other factors on the verification results. This approach significantly improves the data quality, operational efficiency, and consistency of results throughout the verification lifecycle. This research not only demonstrates the broad application prospects of intelligent technology in automotive quality control but also offers the industry an innovative and reliable method for vehicle defrosting performance verification.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Design and Research of an Intelligent Verification Method for Vehicle Defrosting Performance Based on Machine Vision

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Zhang Yunbo,
  • Zhao Haitao,
  • Si Wenjuan,
  • Lu Jianlin,
  • Liu Xiujun,
  • Li Po

摘要

To address the numerous challenges associated with the current vehicle defrosting performance verification process, which relies heavily on traditional manual methods, this paper proposes an intelligent verification method based on machine vision. The proposed method leverages a hybrid neural network architecture, utilizing the Swin-Transformer (Liu, Z. et al. “Swin Transformer: Hierarchical Vision Transformer using Shifted Windows”. IEEE/CVF International Conference on Computer Vision (ICCV), 202) to fully retain and utilize the self-attention mechanism of the Transformer (Vaswani, Ashish, Shazeer, Noam, Parmar, Niki, Uszkoreit, Jakob, Jones, Llion, Gomez, Aidan N, Kaiser, Łukasz, and Polosukhin, Illia. “Attention is All You Need”. Advances in Neural Information Processing Systems, 2017: 5998–6008.; Fu, Jun; Liu, Jing; Tian, Haijie; Li, Yong; Bao, Yongjun; Fang, Zhiwei; Lu, Hanqing. “Dual Attention Network for Scene Segmentation”. IEEE Conference on Computer Vision and Pattern Recognition, 2019, Pages 3146–315; Tang, Jinhui, and Lu, Hanqing. “Adaptive Context Network for Scene Parsing”. IEEE/CVF International Conference on Computer Vision, 2019: 6748–6757.;) while incorporating the local perception capabilities of Convolutional Neural Networks (CNNs). This combination enables the automatic extraction and intelligent analysis of various frost layer characteristics at different time points. The method supports comprehensive data recording and the intelligent analysis of key metrics, such as defrosting rates, throughout the verification process, effectively eliminating the influence of human subjectivity and other factors on the verification results. This approach significantly improves the data quality, operational efficiency, and consistency of results throughout the verification lifecycle. This research not only demonstrates the broad application prospects of intelligent technology in automotive quality control but also offers the industry an innovative and reliable method for vehicle defrosting performance verification.