【摘要】 提出一种改进Harris特征点的机器人精确定位方法,通过改进特征点提取、匹配、跟踪策略,为运动估计提供更加可靠的输入,提高运动估计结果的准确性。具体实现策略是在图像高斯金字塔中改进Harris特征点提取策略,提高其对旋转和尺度变化的适应性,实现特征点的均匀分布且数量可控。然后利用惯导先验信息预测特征点跟踪的搜索区域,提高特征点跟踪的准确性和稳定性,提高搜索效率。最后采用四元数和最小二乘方法估计车体姿态信息。实验结果表明:算法可较好地实现惯导信息与视觉信息的融合,不仅改善了传统视觉定位算法中特征点特性影响定位精度的问题,且几乎不会受到惯导漂移引起定位精度下降的问题,可实现机器人精确定位。更多还原

【Abstract】 A precise self-localization algorithm based on improved Harris feature point is proposed.It provides more reliable input for motion estimation through improving the strategies of feature detecting,matching and tracking,which improves the accuracy of motion estimation.The improved strategy of extracting Harris corners in Gaussian scale spaces is presented.The method not only makes the features be rotation and scale invariant,but also controls the quantity of the uniform distributed features.Using the IMU prior data to estimate the areas of the tracked features,which ensures the accuracy of feature-tracking and reduces the computing time.Quaternion and least square method are used to estimate the attitude information of the vehicle.Results of the experiments show that the algorithm fuses IMU data and visual information excellently,makes feature matching and tracking stably,avoids the drift problem of IMU data and achieves a higher localization precision.更多还原