⁠Marine Ecosystem Monitoring Based on Remote Sensing Using Underwater Image Analysis for Biodiversity Conservation Model

The growing focus on underwater observation and use of marine resources in recent years has led to an active research area in underwater image processing and analysis. Because of its importance in marine engineering as well as aquatic robotics, underwater picture enhancement serves as a preprocessing step to increase performance of high-level vision tasks like underwater object detection as well as recognition. While a number of studies demonstrate that underwater image enhancement methods (synthetic underwater environment image) increase the detectors’ F-measure, no research has examined the connection between these two jobs. This research proposed novel technique in remote sensing–based underwater image analysis in ecosystem monitoring in biodiversity conservation model using machine learning techniques. Here, the underwater-based remote sensing image based on marine ecosystem has been collected and analysed for noise removal with normalisation. The processed image feature extracted and classified using Gaussian spatial gradient neural network (GSGNN) and multilayer pyramid Bayes network (MPBN). The output images show coral surface texture. The experimental analysis has been carried out for various underwater dataset in terms of kappa coefficient, classification accuracy, random precision, F-measure, and normalised square error. Proposed technique kappa coefficient is 94%, classification accuracy is 98%, random precision is 95%, F-measure is 93%, and Normalised square error is 63%.