A Small Target Detection Method for Sea Surface Based on Guided Filtering and Local Mean Gray Difference

The traditional small target detection algorithm often results in a high false alarm rate on the sea surface background. To address this issue, a small target detection method based on guided filtering and local average gray level difference was proposed in this paper for the sea surface. Firstly, the method enhanced the details of the small targets by employing guided filtering to suppress the background clutter and noise in the sea surface image. Subsequently, the local average gray level difference of each point in the image was calculated to further distinguish the targets from other interference points. Finally, the threshold segmentation method was utilized to obtain the actual small targets on the sea surface. After conducting experiments on various sea surface scenes, the LSCRG, BSF, and ROC curve were computed for the proposed method and five other algorithms. Comparative analysis with BS, Top-hat, TDLMS, Max-median, and LCM demonstrates the superiority of the proposed method for infrared small target detection on the sea surface.

Threat Assessment Method Considering Target Instantaneous and Historical States

The sea surface escort formation faces various threats in reality. For example, suicide boats may carry explosives or other dangerous items, aiming to cause maximum damage by colliding or detonating escort targets. Since suicide boats have a certain degree of concealment, it is necessary to establish a threat assessment algorithm to timely identify and respond to such fast and concealed threats. This paper establishes a threat assessment model that considers the instantaneous and historical states of the target. The instantaneous state of the target takes into account six evaluation indicators, including target category, target distance, target heading, target speed, collision risk, and ship automatic identification system(AIS) recognition status;in terms of historical state information mining, a target typical intention recognition method based on graph neural network is proposed to achieve end-to-end target typical intention recognition. Furthermore, this paper introduces a multi-attribute decision analysis method to weight the evaluation indicators, improves the relative closeness calculation method between different evaluation schemes and positive and negative ideal schemes, and determines the target threat ranking based on relative closeness. Based on Unity3D, a set of unmanned boat confrontation simulation system is designed and developed, and typical intention recognition data sets and threat assessment scenario simulation data are generated through real-life confrontation. Comparative analysis shows that the threat assessment model in this paper can accurately and timely detect raid target threats and give scientific and reasonable target threat ranking results.