基于多域学习的红外空中目标跟踪算法

Infrared Air Target Tracking Algorithm Based on Multi-Domain Learning

自深度学习技术被提出以来,迅速风靡各个学术领域,极大地推动了图像处理技术的发展。红外目标跟踪技术是红外导引领域的一项关键技术,但目前深度学习技术在图像处理中的应用主要集中在可见光领域,在红外领域鲜有应用。同时,由于红外场景的复杂性,红外空中目标跟踪的效果遭遇瓶颈。该文基于多域学习训练思想,设计开发了一种应用于红外领域的目标跟踪卷积神经网络,利用VOT2016红外数据集训练后,在仿真红外空中目标序列上达到了优秀的跟踪速度和跟踪精度,并具备一定的抗干扰能力。

Since the introduction of deep learning technology,it has quickly become popular in various academic fields and has greatly promoted the development of image processing technology.Infrared target tracking technology is a key technology in the field of infrared guidance,but the current application of deep learning technology in image processing is mainly concentrated in the field of visible light,with few applications in the field of infrared.At the same time,due to the complexity of infrared scenes,the effect of infrared air target tracking encounters a bottleneck.Based on the idea of multi-domian learning,this paper designs and develops a target tracking convolutional neural network applied in the infrared field.After training with the VOT2016 infrared datasets,it achieves excellent tracking speed and tracking accuracy on the simulated infrared air target sequence.And have anti-interference ability.