Convolutional neural network-based real-time ROV detection using forward-looking sonar image

Agent system is strategy to enhance the underwater manipulation. The conventional manipulation is generally robot arm-based configuration which has singular points. On the other hand, the agent system is an armless manipulation that the agent vehicle works as the end-effector. If the location of the agent can be measured, the end effector is able to be place to any position. To implement this system, the method of an agent vehicle localization is proposed. The method uses the sonar images of moving agent obtained by forward-looking sonar. To detect the location of the agent in the sonar images, the convolutional neural network is applied. We applied the state-of-art object-detection algorithm to the agent vehicle system. The fast object-detection algorithm based on neural network can fulfil the real-time detection and show the remarkable validity. It means the underwater robot can begin navigation under its feed-back. Through field experiment, we confirm the proposed method can detect and track the agent in the successive sonar images.