Automatic underwaTer Image REstoration System

Currently, underwater image restoration (color, denoising) mainly focuses on single cameras and is not connected with automatic identification. Algorithms that use multiple cameras for underwater reconstruction are rare, and the issue of automatic identification remains separate from image restoration.

To improve image quality, ANERIS will develop computational photography methods for two sources: i) single underwater images captured by individuals using smartphones, and ii) images taken by multiple low-cost cameras connected together, utilizing information from various viewpoints. In both cases, the methods will include denoising, visibility enhancement, and color correction.

Multiple single-image reconstruction methods were already developed, with an example of the algorithm shown on Fig.1. A new and improved single-image algorithm is currently under development (Fig.2). Video examples can be found here. Read more about the technology in the following article.

Figure 1. Example of our single image algorithm, already developed.  

Figure 2. Example of our single image algorithm in development.

Target Stakeholders

Academia: Marine Research Observatories

Contact / Technology lead:

Tali Treibitz (

Partners involved