Scholarlink

<p>Atmospheric haze often introduces a dominant bluish or yellowish color cast in images, degrading visibility and visual quality. Analysis of hazy image histograms reveals that this color shift is primarily driven by an imbalance in the blue channel, particularly in non-homogeneous haze and sky regions. To address this, we propose a novel Blue Channel Balancing method operating within the CIELAB color space. Our approach calculates the deviation of the B-channel mean from a neutral reference and adaptively corrects it using the normalized luminance (L-channel) as a weighting factor. This effectively mitigates the blue-yellow bias while preserving natural color fidelity. Experimental results on the O-HAZE dataset demonstrate that our method significantly improves image clarity and dehazing performance compared to traditional Dark Channel Prior (DCP) methods, particularly in medium-density haze scenarios. The proposed technique offers a computationally efficient solution for enhancing visibility in outdoor surveillance and autonomous driving 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

Optimising hazy image clarity through blue channel balancing in lab color space

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年2月9日

查看全文
  • Ibrahim Salim Sulaiman Al Farsi,
  • Mohd Shafry Mohd Rahim,
  • Devi Willieam Anggara,
  • Falah Y. H. Ahmed,
  • Asniyani Nur Haidar Abdullah

摘要

Atmospheric haze often introduces a dominant bluish or yellowish color cast in images, degrading visibility and visual quality. Analysis of hazy image histograms reveals that this color shift is primarily driven by an imbalance in the blue channel, particularly in non-homogeneous haze and sky regions. To address this, we propose a novel Blue Channel Balancing method operating within the CIELAB color space. Our approach calculates the deviation of the B-channel mean from a neutral reference and adaptively corrects it using the normalized luminance (L-channel) as a weighting factor. This effectively mitigates the blue-yellow bias while preserving natural color fidelity. Experimental results on the O-HAZE dataset demonstrate that our method significantly improves image clarity and dehazing performance compared to traditional Dark Channel Prior (DCP) methods, particularly in medium-density haze scenarios. The proposed technique offers a computationally efficient solution for enhancing visibility in outdoor surveillance and autonomous driving applications.