Target Classification of Marine Debris using Deep Learning

Abstract

Marine Debris is human-created waste dumped into the sea or ocean which make the life of ocean species into danger and make environment polluted. Removal of marine debris from ocean is necessary to remove pollution and to make species life safe. As species lives are affected from marine debris. For detection of marine debris inside the ocean, a robust and automatic system is necessary that detect unnecessary litter of plastic and other garbage at real-time which can be further removed from the ocean. In this study, we have proposed to use deep learning based architectures for the detection and classification of marine debris. We have used Histogram Equalization technique combined with Median Filter(HE-MF) to enhance the contrast of images which would help deep learning based algorithms to easily detect the debris and improvement in performance. Results assessment of our proposed technique achieved recall of (96%) and Mean Overlap bounding boxes of (3.78%). Visual and qualitative assessment of proposed methodology shows the effectiveness of proposed technique