Advancing Plastic Pollution Detection in Underwater Environments Using CNN Architectures

Most of the plastic debris from land that ends up in the ocean originates from human waste, making ocean pollution one of the most serious environmental problems. The creatures, the local economy, and the equilibrium of the marine ecosystem are all at risk from these contaminants. Aquatic life and humans will undoubtedly be impacted by this. Various advanced models and approaches are used to detect and measure plastic pollution in the water. This research focuses on a number of well-known methods, such as VGG16, MobileNetV2, DenseNet, and a bespoke convolutional neural network (CNN) architecture. Using these cutting-edge models, we want to improve the precision and effectiveness of plastic garbage detection in underwater environments. MobileNetV2, which demonstrated its better performance with 80% accuracy and computational efficiency in this setting, produced the most promising results among the tested models.