Vehicle detection Using Convolutional Neural Networks in Sonar Images
DOI:
https://doi.org/10.17762/msea.v70i2.1577Abstract
An agent system is used to improve underwater operations. Single-point robotic arm-based positioning is often used in conventional operations. In contrast to this armless operation, agent systems use agent vehicles as end effectors. The end effector can be placed anywhere as long as the agent can be located. As a means of realizing this system, technology is provided to identify the location of the agent's vehicle. This method uses look-ahead sonar imagery of moving agents. CNN is used to find out agent within sonar image. An Agent system had the advanced algorithms for object detection. Neural- network object detection methods will meet real-time detection requirements and will give exceptional effectiveness. That is the underwater robot can perform the task by its own. In real-world this approach is effective in locating agents in next sonar photos and track the progress.