Color Constancy in Aquatic Environment: Application to Live Fish Recognition

摘要

We present in this paper some aspects of color restoration of underwater images and an application to live fish segmentation and recognition. The underwater images are characterized by a strong and non-uniform color cast due to the aquatic environment and to artificial (and/or natural) lighting. The application considered in this paper is about an information system dedicated to aquariums called Aqu@theque. The proposed automatic color correction techniques are based on the perceptual ACE model, and PHM color constancy method. The restored images give better results when displayed or processed (fish segmentation and recognition). The presented preliminary results are satisfying and promising.