Underwater heritage image enhancement and color restoration integrating partitioned background light estimation and deep fusion

Authors:
Open Access
Publisher website
Google Scholar
Abstract
The accurate recording of high-quality underwater cultural heritage images is crucial for archeological research. However, underwater images frequently suffer from color distortion and reduced clarity, which compromises image quality. Existing underwater image enhancement methods often lead to either over-enhancement or under-enhancement, thereby obscuring artifact details and hindering archeological research. This study proposes a method for enhancing quality and restoring color in underwater cultural heritage images, based on an underwater physical imaging model. First, a different background light estimation algorithm based on brightness segmentation (DBE-BS) is developed to enable adaptive multi-region background light estimation, thereby mitigating the impact of uneven lighting on the image. Next, the depth-saturation fusion transmission (DSFT) map estimation algorithm integrates depth information with the inverse saturation map, improving transmission map accuracy. Finally, the Depth-integrated Color Compensation Model (DICC) is introduced to optimize color correction using image depth data, enhancing the image’s visual quality.
Funding Information
  • Postgraduate Education and Teaching Quality Improvement Project of Beijing University of Civil Engineering and Architecture (J2023002)
  • National Natural Science Foundation of China (61902016, 61902016)
  • Graduate Student Innovation Projects of Beijing University of Civil Engineering and Architecture (PG2024121)
  • Education and Teaching Research Project of China Construction Education Association (2023007)
  • Beijing Education Science 14th Five Year Plan Project (CDDB24252)

This publication has 28 references indexed in Scilit: