Abstract

The detection of distant small ship targets in the marine environment is a critical and challenging issue that urgently needs to be addressed in the realization of accurate marine information control in the complex environment. It is of great significance for monitoring Marine environment and safeguarding maritime sovereignty In the process of acquiring target information on ships at sea, the images captured typically contain information of moving targets within dynamic scenes. Traditional, singular methods are inadequate for obtaining complete information on these moving targets. Based on this, the article proposes an integrated method combining motion target detection algorithms, edge detection operators, and deep learning-based target detection algorithms. This method constructs an improved motion target detection algorithm to achieve comprehensive information acquisition and detection of the position, size, and type of moving ship targets in complex marine environments. Experimental simulation has validated the network performance and practical value. The network has been deployed on an Nvidia Jetson TX2 development board for real-world testing, confirming its performance in detecting dynamic ship targets in actual marine environments, and providing a viable technical approach and theoretical support for enhancing the refined target selection capability.