Scholarlink

Haze caused by atmospheric particles or uneven lighting severely degrades the image quality impacting applications such as object detection. Image dehazing plays a significantly important role in tasks related to computer vision, mitigating degradation of visibility caused by haze. This paper presents an advanced approach for dehazing an image aimed at improving visual quality and restoring clarity in both outdoor and indoor environments. Building on Dark Channel prior (DCP) technique, our approach integrates weighted Atmospheric Light estimation, Transmission map estimation and its refinement through guided filter to preserve image details while minimizing artifacts. For outdoor images, a contrast improvement method called Contrast-Limited Adaptive Histogram Equalization (CLAHE) along with some blending techniques are applied as post processing to enhance contrast and ensure natural appearance. For Indoor images, Gamma Correction with contrast stretching are applied as post processing steps to bring out finer details and improve illumination. Experiments are carried out using SOTS and ITS subsets of the standard RESIDE dataset. The framework’s performance is quantitatively evaluated using Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index Measure (SSIM). Its dynamic adaptability to environmental conditions makes it suitable for diverse applications, including surveillance, autonomous systems and digital photography.

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

Dynamic Image Dehazing: A Framework for Improved Visibility in Varying Environments

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • K. Kishore Raju,
  • D. Karthikeya,
  • K. Raj Kumar,
  • B. Kranthi Swaroop,
  • J. Surya Deepthi Narasimha

摘要

Haze caused by atmospheric particles or uneven lighting severely degrades the image quality impacting applications such as object detection. Image dehazing plays a significantly important role in tasks related to computer vision, mitigating degradation of visibility caused by haze. This paper presents an advanced approach for dehazing an image aimed at improving visual quality and restoring clarity in both outdoor and indoor environments. Building on Dark Channel prior (DCP) technique, our approach integrates weighted Atmospheric Light estimation, Transmission map estimation and its refinement through guided filter to preserve image details while minimizing artifacts. For outdoor images, a contrast improvement method called Contrast-Limited Adaptive Histogram Equalization (CLAHE) along with some blending techniques are applied as post processing to enhance contrast and ensure natural appearance. For Indoor images, Gamma Correction with contrast stretching are applied as post processing steps to bring out finer details and improve illumination. Experiments are carried out using SOTS and ITS subsets of the standard RESIDE dataset. The framework’s performance is quantitatively evaluated using Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index Measure (SSIM). Its dynamic adaptability to environmental conditions makes it suitable for diverse applications, including surveillance, autonomous systems and digital photography.