SiamIST: Infrared small target tracking based on an improved SiamRPN

Abstract

The continuous tracking of infrared dim-small targets is significant, due to limited spatial resolution and low thermal features. Tracking algorithms based correlation filter may perform not well referring to infrared information. Therefore, a Deep Learning (DL) model is proposed for the tracking task with public data sets of small targets. To be specific, the Siamese Region Proposal Network (SiamRPN) is improved by the style recalibration module, which can obtain the perception of image styles. Furthermore, the proposed algorithm takes advantage of transfer learning technology referring to labeled target images, obtaining good features. To distinguish the small target from the background edges, the side window filtering is combined with the improved SiamRPN model. The experimental results show the good performance of the proposed small target tracking , namely SiamIST, in public near-infrared videos, compared to several related algorithms. Importantly, the designed algorithm uses the DL model to track small infrared targets for the first time, achieving an overall precision of 78.8%. • A deep learning-based algorithm is proposed for the first time to track a small maneuvering target of the Unmanned Aerial Vehicle (UAV) in NIR videos. • An improved SiamRPN model utilizing SRM-PF is designed to effectively complete small target tracking . • The SWF is adopted to suppress the bad influence of the edge of the image, helping to obtain a good response for small targets. • Transfer learning is employed on a public data set of the IR small target.