Use of Moiré Patterns in Camera Position Estimation
Use of Moiré Patterns in Camera Position Estimation
复制标题
莫尔图案在相机位置估计中的使用
DOI：
--
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
Jun Junghyun
中科院分区：
文献类型：
--
作者：
Samuel Banks;R. Green;Jun Junghyun
先前的工作已经转向使用由双凸透镜形成的莫尔图案来执行短距离的姿态估计。目前的工作已经调查了现有的理论莫尔图案，最显着的傅立叶和第一谐波近似模型。这些模型已经在本文中开发，能够在3D空间中的分离的两个图案生成的莫尔图案的模型。这些模型的准确性已受到现实世界的测试，相机的距离和横向平移估计。从不同的摄像机横向平移的初步测试似乎是准确的近距离测试，约10毫米的精度从160毫米。从不同的摄像机距离的结果显示的承诺，但是，他们并不像预期的精度在0-130毫米的精度从范围100- 2000毫米。
Previous work has gone towards using moire patterns formed with lenticular lenses to perform pose estimation for short ranges. Current work has investigated existing theory of moire patterns, most notably the Fourier and first harmonic approximation models. These models have been developed in this paper to be able to model moire patterns generated from two patterns in 3D space that are separated. The accuracy of these models has been subjected to real-world tests for camera distance and lateral translation estimation. Preliminary tests from varying camera lateral translation appeared to be accurate for the close-range testing, with about 10 mm accuracy from 160 mm. Results from varying camera distance showed promise, however, they weren’t nearly as accurate as anticipated at 0–130 mm accuracy from ranges 100– 2000 mm.