A Complex Convolution Kernel-Based Optical Displacement Sensor
A Complex Convolution Kernel-Based Optical Displacement Sensor
复制标题
DOI：
10.1109/jsen.2020.2986240
复制
发表时间：
2020-09
影响因子：
4.3
通讯作者：
Matthew Southwick;Zhu Mao;C. Niezrecki
中科院分区：
文献类型：
--
作者：
Matthew Southwick;Zhu Mao;C. Niezrecki
近十年来，各种光学位移测量技术得到了发展、商业化和广泛应用。本文提出了一种新型光学位移传感器，该传感器在现有技术的基础上，具有更高的亚像素分辨率、更高的精度和更高的计算效率。本文开发的基于复卷积核的光学传感器（KBOS）利用复值核函数的相位幅度表示来实现用于二维运动测量的优越光学位移传感器。传感器的性能和特性是根据其计算标记运动的能力来量化的。该指标的评估将解释传感器如何能够实现优于当前技术状态的性能。然后通过一系列操作条件下的实验验证该设计，以证明该方法优于其他现有位移测量技术。最后，将传感器的性能与文献中发现的其他先进光学位移传感器进行定性比较。该传感器显示具有亚像素分辨率之间的1/100和1/2500的像素取决于空间采样密度。对传感器的动态范围进行了研究，发现它只受相机采样速度的限制。
During the last decade, a variety of optical displacement measurement techniques have been developed, commercialized and widely used. This paper presents a new type of optical displacement sensor that improves on the prior art with higher sub-pixel resolution, better accuracy, and better computation efficiency. The Complex Convolution Kernel-Based Optical Sensor (KBOS) developed in this paper leverages the phase-magnitude representation of a complex valued kernel function to achieve a superior optical displacement sensor for 2D motion measurements. The sensor’s performance and properties are quantified in terms of its ability to compute the marker motions. Evaluation of this metric will explain how the sensor is able to achieve superior performance over the state of the art. The design is then validated with a set of experiments over a range of operating conditions to demonstrate the effectiveness of the approach over other existing displacement measurement techniques. Finally, the sensor’s capability is compared qualitatively against other state of the art optical displacement sensors found in the literature. The sensor is shown to possess a subpixel resolution of between 1/100 and 1/2500 of a pixel depending upon spatial sampling density. The sensors dynamic range was investigated and shown to be only limited by the camera sampling speeds.