Self-tuning of a sunlight-deflickering filter for moving scenes underwater

This work deals with the design of a real-time approach for automatic parameter tuning in a deflickering process of videos of a moving landscape scene underwater perturbed with sunlight waves. The tuner performs a continuous parameter optimization inside a basic filter, which employes feedback in order to improve the performance. Also it is able to adapt the optimum to eventual changing of the scenario conditions. Moreover, issues like the parameter sensitivity and loop stability are addressed equally. The results have been illustrated by means of a worked-out case-study employing real film material of a scenario underwater. The approach was motivated by the SLAM applications with monocamera vision-based vehicles underwater and simple hardware implementations in FPGA structures.