Scholarlink

While the use of neural radiance fields (NeRFs) in different challenging settings has been explored, only very recently have any contributions focused on using NeRFs in foggy environments. We demonstrate that the traditional NeRF models are able to replicate scenes filled with fog and propose a method to remove the fog when synthesizing novel views. By calculating the global contrast of a scene, we can estimate a density threshold that removes all visible fog when applied. This makes it possible to use NeRF to render clear views of objects of interest in fog-filled environments. Additionally, to benchmark performance on such scenes, we introduce a new dataset that expands some of the original synthetic NeRF scenes by adding fog and natural environments. The code, dataset, and video results can be found on our project page: https://vegardskui.com/fognerf/ .

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

Removing Adverse Volumetric Effects from Trained Neural Radiance Fields

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Andreas Langeland Teigen,
  • Mauhing Yip,
  • Victor P. Hamran,
  • Vegard Skui,
  • Annette Stahl,
  • Rudolf Mester

摘要

While the use of neural radiance fields (NeRFs) in different challenging settings has been explored, only very recently have any contributions focused on using NeRFs in foggy environments. We demonstrate that the traditional NeRF models are able to replicate scenes filled with fog and propose a method to remove the fog when synthesizing novel views. By calculating the global contrast of a scene, we can estimate a density threshold that removes all visible fog when applied. This makes it possible to use NeRF to render clear views of objects of interest in fog-filled environments. Additionally, to benchmark performance on such scenes, we introduce a new dataset that expands some of the original synthetic NeRF scenes by adding fog and natural environments. The code, dataset, and video results can be found on our project page: https://vegardskui.com/fognerf/ .