基于多数据融合的多路视频图像目标跟踪方法  

Multi-channel Video Image Target Tracking Method Based on Multi-data Fusion

作　　者：范博 FAN Bo(Pinduoduo,Shanghai 200051)

机构地区：[1]拼多多,上海200051

出　　处：《长江信息通信》2025年第1期19-21,共3页Changjiang Information & Communications

摘　　要：现有的跟踪方法实际位置与目标位置存在误差,为此研究基于多模态大模型的多路视频图像目标跟踪方法。通过光学影像与SAR影像结合的方式构建多模态数据大模型,通过数据融合方式提取多路视频图像多种特征,并采用通道叠加方法将不同模态特征合并融合,实现优势互补。利用布莱卡滤波方法预测目标的可能位置,然后以该预测位置为中心,对周围区域内的像素点进行整合。根据导向概率来构建一个导向图,统计各视频块中的目标点与总样本点数的比率。通过与预设的阈值比较,动态调整跟踪窗口的大小,完成目标跟踪。实验结果表明,不论是单目标还是多目标的情况,通过扫描发现目标位置与实际区域达到最佳匹配,记录到最终位置的像素坐标为(30,550),与预期效果相符合。达到了良好的跟踪效果,实现了对多路视频图像中目标的精确跟踪。Due to the error between the actual position and the target position of existing tracking methods,a multi-channel video image target tracking method based on multi-modal large model is studied.A large model of multi-modal data is built by combining optical image and SAR image,and multiple features of multi-channel video images are extracted by data fusion,and different modal features are merged and fused by channel superposition method to achieve complementary advantages.Breca filter method is used to predict the possible position of the target,and then the predicted position is used as the center to integrate the pixels in the surrounding area.According to the guiding probability,a guiding chart is constructed to calculate the ratio of target points in each video block to the total sample points.By comparing with the preset threshold,dynamically adjusting the size of the tracking window completes the target tracking.The experimental results show that no matter in the case of single target or multi-target,the optimal match between the target position and the actual area is found through scanning,and the pixel coordinate of the recorded final position is(30,550),whichis consistent with the expected effect.Good tracking effect is achieved,and the target tracking in multi-channel video image is realized accurately.

关 键 词：多模态特征融合 视频 图像 目标跟踪 多模态 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]