基于微透镜阵列空间角度编码的高精度视觉位姿测量方法

Microlens Array-Based Spatial Angle Encoding for High-Precision Visual Pose Measurement

为提升现有视觉标记点法的角度测量精度与鲁棒性,提出了一种融合微透镜阵列与随机微图形阵列编码的高精度视觉位姿测量方案。首先,在标定阶段中,利用微透镜阵列下动态变化的随机微图形阵列成像模式对角度空间进行密集的高精度编码,通过汉明码构建高精度空间角度坐标基准。然后,在测量阶段中,基于观测到的显示图像构建观测角度编码,并联合最邻近角度基准加权计算出测量位置的精确角度信息,最终实现超高精度的三轴角度测量与位姿估计。同时,进行了大量的真实实验以验证所提方法的有效性。实验结果表明,所提方法可实现高于0.08°的角度测量精度,与先前工作相比,在保证位移测量精度的情况下将角度测量平均误差减小了97%。此外,所提算法的鲁棒性保证了自然光照条件下的高精度位姿测量。最后,通过充分的定量比较全面地分析了阈值分割与角度编码参数等因素对测量精度的影响。

To improve the precision and robustness of angle measurement by existing visual marker methods,a highprecision visual pose measurement approach that combines microlens arrays and random micrographics arrays for encoding is proposed.First,in the calibration stage,dense and high-precision encoding is carried out by dynamic imaging modes of random micrographics arrays under microlens arrays,and a high-precision coordinate datum of spatial angles is constructed by the Hamming code.Then,in the measurement stage,observation angle encoding is completed on the basis of observed display images,the weighted fusion of which with the nearest angle datum is applied to obtain the precise angles of measuring positions.In this way,the triaxial angle measurement and pose estimation with high precision can be achieved.At the same time,extensive experiments are performed to verify the validity of the proposed method.The results reveal that the angle measurement precision is higher than 0.08°,and compared with the results of previous work,the average angle measurement error is reduced by 97%without sacrificing the displacement measurement precision.In addition,the robustness of the proposed algorithm ensures high-precision pose measurement under natural illumination.Finally,by sufficient quantitative comparison,the effects of factors on angle measurement precision,such as threshold segmentation and angle encoding parameters,are comprehensively analyzed.