During his three years at Fraunhofer IGD, Martin Radolko focused in his dissertation, “Change Detection in Combination with Spatial Models and its Effectiveness on Underwater Scenarios,” on the automated detection of fish in underwater footage. In his defense on January 21, 2019, he explained his approach to developing reliable and precise object recognition despite the specific challenges posed by underwater imagery. Based on self-recorded videos, he developed a tracking method that identifies and classifies fish according to various features.

A next step would be growth analysis as part of automated size measurement. Possible future fields of application include aquaculture, where the growth rate and health status of fish could be monitored without subjecting them to stressful manual examinations. His work thus provides an important foundation for the goal of offering users in the field of underwater imaging software that delivers monitoring-relevant data at a glance, without the need for time-consuming and costly manual inspections.