滑动置信度约束的红外弱小目标跟踪算法研究

为了提高红外视频弱小目标的跟踪精度，提出了滑动置信度约束的弱小目标跟踪方法。在快速自适应中值滤波的红外图像背景抑制技术的基础上，设计了正交变换和置信域约束的轨迹预测，利用加权参数增强目标函数的收敛性能，提高下一位置初的预测准确度；通过轨迹相邻点的位置差计算搜索窗口的大小，搜索与之相匹配的特征点进行关联处理，完成对初预测点的筛选；以滑动轨迹置信度检验为准则判决轨迹的真实性，并进行目标轨迹更新以实现对弱小目标的准确跟踪。通过红外弱小目标视频对所提算法进行了实验验证，结果表明,该算法对红外弱小目标的跟踪轨迹误差有较小的均方偏差与均方差，在噪声消除和对图像整体信息保护方面都具有良好的性能。

Research on Infrared Dim-small Target Tracking Algorithm with Template Sliding Confidence Constraint

The infrared dim-small target has a small portion of image pixel and low SNR,which makes it difficult to detect and track the target especially in noise and clutter.A dim-small target tracking method with fixed template sliding confidence constraint,on the basis of the fast adaptive median filter to suppress the infrared background,is presented.A prediction of trajectory with orthogonal transformation and confidence region is proposed.Weighted parameters are used to enhance the convergence of target function and the prediction accuracy of the next position.The size of sliding search window is calculated from the position difference of the adjacent points on trajectory to search the matched feature point with the predicted position in the next frame and make further parallel processing.The trajectory sliding confidence is used to verify the authenticity of trajectory and update the target trajectory for accurate tracking.The proposed algorithm is tested with infrared dim-small target video.The results show that the proposed method shows better performance in target tracking with lower mean square error deviation and mean square error.Furthermore,the good performance of noise elimination and image information protection also verifies the effectiveness of the algorithm.