CAJ | 학술논문

为了实现对一组动态特征点的跟踪,基于Gabor变换的特征点提取方法提出了一套跟踪方案.首先采用仿射变换描述点的轨迹模型,引入PDAF算法进行状态估计;然后基于Gabor变换特征向量的匹配算法进行对应点辨识;最终推导出EKF滤波器方程组,对时态信息进行滤波.仿真结果表明,该特征点跟踪算法能够在目标图像辨识率较低的情况下满足目标跟踪的精确性要求.
위료실현대일조동태특정점적근종,기우Gabor변환적특정점제취방법제출료일투근종방안.수선채용방사변환묘술점적궤적모형,인입PDAF산법진행상태고계;연후기우Gabor변환특정향량적필배산법진행대응점변식;최종추도출EKF려파기방정조,대시태신식진행려파.방진결과표명,해특정점근종산법능구재목표도상변식솔교저적정황하만족목표근종적정학성요구.
In order to realize the tracking of a series of dynamic feature points, this paper put forward a tracking scheme based on feature point extraction. First of all, used simple affine transformation to describe the motion of feature points and employed PDAF to obtain the estimation of state vector. Then employed a matching method based on Gabor wavelet transformation to find the corresponding point. Finally cited an EKF to process the temporal information. The simulation of experiment shows that this tracking algorithm can meet the accuracy requirement of object tracking in the situation of low image resolution.