Scholarlink

Pedestrian causalities and harms happen in most pedestrian crashes in the clouded places. However, researchers have contributed lot to maximize accuracy to develop pedestrian detection system. Still, the system is not accurate. The general issue in pedestrian detection is the minimize accuracy in night-time, but system might work well in the different lighting conditions. This research introduces the assessment of performance of pedestrian detection in dynamic lightning conditions, and included to apply multi-spectral image and deep learning to increase the performance. The system is compared on the basis of RGB and multispectral images for the evaluation of the pedestrian recognition. Additionally, optimized design for the deep learning is processed to get maximize recognition rate and minimize processing time for the given detection system. The achieved results indicate that multi-spectral image is very useful in pedestrian detection in dynamic lightning conditions.

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

Multispectral Images-Based Enhanced Pedestrian Detection System Using Deep Learning

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Mayur Rahul,
  • Himanshu Shukla,
  • Amit Virmani,
  • Akhilesh Singh,
  • Prashant Srivastava,
  • Deepak Kumar Mishra,
  • Vikash Yadav

摘要

Pedestrian causalities and harms happen in most pedestrian crashes in the clouded places. However, researchers have contributed lot to maximize accuracy to develop pedestrian detection system. Still, the system is not accurate. The general issue in pedestrian detection is the minimize accuracy in night-time, but system might work well in the different lighting conditions. This research introduces the assessment of performance of pedestrian detection in dynamic lightning conditions, and included to apply multi-spectral image and deep learning to increase the performance. The system is compared on the basis of RGB and multispectral images for the evaluation of the pedestrian recognition. Additionally, optimized design for the deep learning is processed to get maximize recognition rate and minimize processing time for the given detection system. The achieved results indicate that multi-spectral image is very useful in pedestrian detection in dynamic lightning conditions.