Online Platform of Convolutional Neural Network for Ship Detection of Optical Remote Sensing Satellite Imagery Data

Monitoring illicit fishing in Indonesia’s extensive oceans requires a significant amount of time and work. Remote sensing technology is an ideal solution for monitoring, controlling, and surveying (MCS) systems. The problem of pictures with varying locations is of particular importance to the ship identification process utilizing a remote sensing system due to the ship’s rectangular and narrow dimensions and the high viewing angle from space. This circumstance changes if the firing is done close to items on the Earth’s surface. Given the difficulty of capturing data from a remote location, the convolutional neural network (CNN) approach for processing remote sensing picture data is still being refined, with a focus on recognizing ships. The most recent version of the CNN algorithm demonstrates its capacity to see vessels by accounting for the orientation of objects at sea level with near-perfect accuracy. The issue is that the system must be able to recognize ship movements and kinds in detail as recorded by remote sensing satellite sensors. An online platform for automated ship identification based on satellite imaging data is required to safeguard Indonesia’s maritime sovereignty. Image data were obtained from arbitrary data and captured The image processing step then includes reading and dividing the dataset, building a CNN, and then training and evaluating the CNN model on online platform. The accuracy values for detecting ship and non-ship objects using the CNN model are 99.78% and 99.24%, respectively.