Infrared target detection in marine images with heavy waves via local patch similarity

Abstract

Infrared detection is a crucial method for marine searching and securing. The accuracy of detecting the targets is a decisive and challenging problem, especially in severe environments. To detect the target in heavy wave infrared images, we analyze the image features in this situation. Based on the difference between targets and waves in the similarity of neighboring patches, a local patch-based similarity descriptor is designed to suppress the wave. Based on the significance of targets in the horizontal direction, an oriented filter is proposed to enhance the targets. According to experiments, the proposed method solves the difficulty in detecting the targets that emerged in the heavy wave, and it makes a balance between accuracy and efficiency. It meets the real-time requirement in the application and has higher detection accuracy than several state-of-the-art methods on several heavy waves infrared datasets.