School of Astronautics, Northwestern Polytechnical University, Xi'an 710072, China.
Shanghai Aerospace Control Technology Institute, Shanghai 201109, China.
Sensors (Basel). 2019 Mar 14;19(6):1289. doi: 10.3390/s19061289.
In this paper, we focus on developing an algorithm for infrared-imaging guidance that enables the aircraft to be reliably tracked in the event of interference. The key challenge is to track the aircraft with occlusion caused by decoys and drastic appearance changes resulting from a diversity of attacking angles. To address this challenge, an aircraft-tracking algorithm was proposed, which provides robustness in tracking the aircraft against the decoys. We reveal the inherent structure and infrared signature of the aircraft, which are used as discriminative features to track the aircraft. The anti-interference method was developed based on simulated images but validate the effectiveness on both real infrared image sequences without decoys and simulated infrared imagery. For frequent occlusion caused by the decoys, the mechanism of occlusion detection is exploited according to the variation of the model distance in tracking process. To have a comprehensive evaluation of tracking performance, infrared-image sequences with different attack angles were simulated, and experiments on benchmark trackers were performed to quantitatively evaluate tracking performance. The experiment results demonstrate that our aircraft-tracking method performs favorably against state-of-the-art trackers.
本文专注于开发一种红外成像制导算法，使飞机在受到干扰的情况下能够可靠地被跟踪。关键的挑战是在诱饵造成的遮挡和由于多种攻击角度导致的外观剧烈变化的情况下跟踪飞机。为了解决这个挑战，我们提出了一种飞机跟踪算法，它为跟踪飞机提供了对诱饵的鲁棒性。我们揭示了飞机的固有结构和红外特征，这些特征被用作跟踪飞机的判别特征。该抗干扰方法是基于模拟图像开发的，但在没有诱饵的真实红外图像序列和模拟红外图像上验证了其有效性。对于诱饵频繁造成的遮挡，我们利用跟踪过程中模型距离的变化来检测遮挡的机制。为了全面评估跟踪性能，我们模拟了不同攻击角度的红外图像序列，并在基准跟踪器上进行了实验，以定量评估跟踪性能。实验结果表明，我们的飞机跟踪方法优于最先进的跟踪器。
