基于改进粒子滤波的视觉目标跟踪算法研究

Research on Visual Target Tracking Algorithm Based on Improved Particle Filter

单纯依靠一种颜色特征往往使得目标跟踪效果不理想,为了在复杂背景及光照突变等情况下实现对目标稳定而准确的跟踪,提出一种基于颜色直方图(CHF)和彩色二进制局部不变特征(CBLIF)相结合的粒子滤波跟踪方法,在CBLIF目标模板和当前搜索区域之间进行特征匹配,当特征点数量大于前几帧数目平均值时,说明跟踪算法有效,此时用跟踪结果的颜色目标模板和CBLIF模板对颜色目标模板和CBLIF模板进行更新;否则,不更新颜色目标模板。实验结果表明,该算法具有较高精度和适应性。

The effect of target tracking algorithm is poorer based on one color features,in order to complete the advantages of fast tracking stability and accuracy at a complex back round and light mutation.A new method is proposed in this paper to handlew with these situations,in which the target model is built by(color histogram feature,CHF)color binary local invariant features(CBLIF)feature,and the particle filter is integrated to achieve target tracking.Whether the colour target model is to be updated or not depends on the number of matching feature points between the tracking result in current frame and the CBLIF model.If the number of matching feature points is greater than the average,tracking results effectively,the color target model will be updated through the numbers of matching points between the current search area and the CBLIF target model,otherwise,do not update.Experiments show that the proposed method can improve the tracking precision especially and stradily compared with traditional methods.