【摘要】 为了提高景像匹配导航系统的定位精度,给出了一种基于Markov随机场理论和极大后验概率估计的新型景像匹配算法。考虑到在x方向和y方向位置偏差的实时图与基准图之间的灰度分布关系,利用图像上的灰度分布服从Markvov随机场分布这一特性,建立了景像匹配问题的条件概率分布模型。应用最小二乘法和噪声的先验统计信息估计位置偏差的方差,给出了描述基准图与实时图之间灰度偏差的测量模型,确定了测量的统计特性。基于极大后验概率估计准则计算了位置偏差的估值。由于新算法在计算位置偏差估值时用到了被估量和噪声的统计信息,因此具有很高的精度。将该算法与最小二乘景像匹配算法进行了仿真比较。仿真结果表明,新算法的匹配精度达到了0.1~0.2像素,高于最小二乘匹配算法的匹配精度。更多还原

【Abstract】 In order to improve the positioning precision of the scene matching navigation system,a novel algorithm of scene matching based on Markov random field theory and maximum a posteriori estimate is proposed.On the basis of the gray level relationship between real-time image and referenced image that had x-orientation and y-orientation location disparity,and by modeling the gray distribution of images as Markov random field,the conditional probability distribution model was established for scene matching.The variance of the location disparity was estimated based on the priori statistical information of the noise using the least squares method,the measurement model describing the gray difference between real-time image and referenced image was given,and the statistical property of the measurement was got.The location disparity was estimated based on maximum a posteriori criterion.The novel matching algorithm has high precision by utilizing the statistical information of estimated value and noise to compute the location disparity estimate.It is shown by the simulation results that the matching precision of the novel matching algorithm is 0.1～0.2 pixel which is higher than that of the least square matching algorithm.更多还原