RGBT双模态加权相关滤波跟踪算法

RGBT Dual-modal Tracking with Weighted Discriminative Correlation Filters

与基于可见光的单模态跟踪算法相比,基于可见光/热红外(RGB-Thermal,RGBT)的双模态跟踪算法对光照变化具有更强的鲁棒性。但在实际应用场景中,双模态跟踪算法仍然会受到局部遮挡目标形变的影响。为了解决以上问题,本文提出了一种可见光/热红外RGBT双模态加权相关滤波跟踪算法。该算法首先利用可见光图像和热红外图像联合求解权重图,然后利用权重图引导相关滤波器求解过程,最后根据权重图推断前景目标是否被遮挡。该算法在公开数据集RGBT234上的结果表明,本文提出的RGBT双模态加权相关滤波跟踪算法能够有效处理目标局部遮挡和目标畸变等情况,实现复杂场景下鲁棒持续的目标跟踪。

Compared with single-modal RGB trackers,dual-modal RGBT(RGB-Thermal)trackers are more robust to illumination variation.In real scenarios,however,dual-modal trackers are severely influenced by partial occlusion and shape deformation.To tackle above problems,in this paper,we propose a weighted DCF(Discriminative Correlation Filter)based RGBT tracker.This tracker derives a weight map from a RGBT image pair and guides correlation filter training with this map.The occlusion state of the foreground target is inferred from the weight map.Experimental results on the public RGBT234 dataset demonstrate that our tracker is able to cope well with partial occlusion and shape deformation and achieves robust and persistent tracking in complex scenarios.