Search papers, labs, and topics across Lattice.
This paper introduces a robust image processing framework specifically designed for underwater robots monitoring construction environments, addressing the unique challenges posed by depth-dependent forward scattering and particle-induced degradations like marine snow. By employing a staged processing pipeline that models background and foreground degradations, the authors retrain an existing Joint-ID network using synthetic data derived from real underwater imagery, allowing for an isolated evaluation of dataset realism. Experimental results show significant improvements in visual quality and UIQM scores on real datasets, highlighting the effectiveness of the proposed methods in bridging the synthetic-to-real gap for underwater applications.
Explicitly modeling forward scattering and marine snow effects can drastically enhance the visual quality of underwater imagery, making robotic monitoring more effective.
This paper proposes a robust image processing framework for underwater robot-based construction environment monitoring, targeting complex degradations observed in real marine environments. Unlike conventional approaches that mainly consider absorption and backscattering, real underwater imagery is strongly affected by depth-dependent forward scattering blur and particle-induced degradations such as marine snow. To address this, we introduce a staged processing pipeline that sequentially models background degradation via depth-aware forward scattering and foreground degradation using realistic marine snow patterns extracted from real images. The resulting synthetic data are used to retrain an existing Joint-ID network without modifying its architecture, enabling an isolated evaluation of dataset realism. In addition, a lightweight post-processing scheme is applied to enhance contrast and structural clarity. Experiments on real underwater datasets collected in Korean coastal environments demonstrate consistent improvements in visual quality and UIQM scores. The results indicate that explicitly modeling forward scattering and realistic particle effects effectively reduces the synthetic-to-real gap and improves practical applicability in real-world underwater robotic operations.
