Scholarlink

<p>Currently, most low-light image enhancement algorithms are trained on open-source ideal datasets, which fail to effectively solve problems such as obvious noise and detail degradation observed in images captured in real low light scenes. Therefore, a USB camera is used to capture images and a low light image enhancement algorithm is designed and implemented in this paper. Firstly, the USB camera by using SONY IMX335 as the image sensor is used to capture images at a extremely low illumination of 0.08&#xa0;lx for subsequent image enhancement. Subsequently, the Self-Enhancement Algorithm based on Non-Local Filtering (SENF) is designed, which includes non-local denoising and self-enhancement. The non-local denoising module preserves the image details through a multiscale morphology and global filtering strategy. The self-enhancement module converts images from RGB to HSV and parameter feedback is introduced to effectively avoid image distortion and over exposure. Experimental results show that SENF achieves about 26% higher objective index evaluation scores and better subjective visual effect than other mainstream algorithms. The proposed SENF provides a reference for improving the quality of low light images in practical applications.</p>

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

Low-light image enhancement algorithm for USB camera

  • 期刊论文

  • 发表日期：

    2025年9月11日

查看全文
  • Yi Tang,
  • Xinwei Luo,
  • Sunan Xu,
  • Shengmin Zhou,
  • Juan Kang

摘要

Currently, most low-light image enhancement algorithms are trained on open-source ideal datasets, which fail to effectively solve problems such as obvious noise and detail degradation observed in images captured in real low light scenes. Therefore, a USB camera is used to capture images and a low light image enhancement algorithm is designed and implemented in this paper. Firstly, the USB camera by using SONY IMX335 as the image sensor is used to capture images at a extremely low illumination of 0.08 lx for subsequent image enhancement. Subsequently, the Self-Enhancement Algorithm based on Non-Local Filtering (SENF) is designed, which includes non-local denoising and self-enhancement. The non-local denoising module preserves the image details through a multiscale morphology and global filtering strategy. The self-enhancement module converts images from RGB to HSV and parameter feedback is introduced to effectively avoid image distortion and over exposure. Experimental results show that SENF achieves about 26% higher objective index evaluation scores and better subjective visual effect than other mainstream algorithms. The proposed SENF provides a reference for improving the quality of low light images in practical applications.