计算机科学
人工智能
计算机视觉
图像(数学)
模式识别(心理学)
作者
Jing Guo,Zhen Liu,Zheng Zhang,Zhiqiang Bai,Yaodong Wang,Baoqing Guo
出处
期刊:International Conference on Digital Image Processing
日期:2025-07-22
卷期号:: 58-58
摘要
To address the issues of low accuracy and efficiency associated with traditional manual inspections and existing nondestructive testing methods, this study develops an intelligent rail area identification system. This system encompasses image acquisition, an intelligent identification algorithm, and system software based on machine vision technology. Rail images are captured using image acquisition equipment installed on an intelligent inspection vehicle. The system performs rail surface image recognition utilizing improved Cascade R-CNN and YOLOX algorithms, with the enhanced YOLOX algorithm achieving a mAP value of 0.896 on the test set. The accuracy of the improved algorithm model are validated through comparisons with the original algorithm, highlighting its potential to enhance the quality and efficiency of rail maintenance and its significant application value in ensuring rail traffic safety.
科研通智能强力驱动
Strongly Powered by AbleSci AI