Image Enhancement for Underwater Applications

In natural water bodies, images taken under water are mostly characterized by poor visibility conditions. To evaluate the content of such images visually, an image restoration is essential. In this contribution, a concept for the automated restoration of images taken under water is presented. The concept is based on a distance estimation using turbidity indicators, a subsequent color restoration, a model-based deconvolution approach and a concluding image fusion. The restoration is fully based on physical foundations and thus complies with the requirement of an objective enhancement of the image content.