Image Detection of Seaweed, Seagrass, and Coral in Coastal and Underwater Marine Ecosystems

Marine ecosystem monitoring is crucial for understanding and preserving the health of coastal and underwater environments. This research presents an innovative approach to automatically detect and classify key components of marine ecosystems, namely seaweed, seagrass, and coral, from the images captured in coastal and underwater environments. We leverage the You Only Look Once (YOLO) object detection framework for its high-accuracy object recognition capabilities. To enhance the detection accuracy and improve the quality of input images, a combination of Generative Adversarial Networks (GANs) and the Contrast Limited Adaptive Histogram Equalization (CLAHE) algorithms is used during training. GANs generate synthetic images, expanding the diversity of training data. The CLAHE algorithm enhances the contrast and detail in both real and synthetic images, making the object detection process more robust. The integration of YOLO with GAN and CLAHE algorithms helps address challenges encountered in underwater imaging conditions, such as low visibility and changing lighting. Compared to the original YOLO-V7, it demonstrates an increase of 8% to 10% in detecting seaweed, seagrass, and coral using the new approach. This new approach showcases the potential for advancing the monitoring of coastal and underwater environments.