High precision roughness sensor based on annular core optical fiber

Abstract

We present a high precision fiber roughness sensor based on annular core optical fiber (ACF). The ACF has a cone-frustum-shaped tip to enable the annular core to generate a focused ring light. When the incident light passing from the ACF is focused on the rough surface to be measured, the specular light will also be collected by using the ACF, which realizes the integration of the incident optical path and collecting optical path about optical roughness detection. Because the focused ring light has a smaller spot radius than that of the emergent light of a single-mode fiber, the proposed sensor has a higher spatial resolution. Moreover, the ACF with a cone-frustum tip can reduce the diffuse light coupling into the sensor; in this way, we improve the measurement precision. The experimental results show that for the surface roughness of 10 nanoscales, the error between the measured value and the theoretical value is about 5% and the spatial resolution reaches 6 µm. Thus, our sensor can be widely used in high precision nondestructive surface roughness measurements.

A surface roughness grade recognition model for milled workpieces based on deep transfer learning

“…area, non-contact measurement methods have gained increasing attention due to their high flexibility and fast measurement speed without touching the measured object [2][3][4]. Noncontact roughness measurement methods include, but are not limited to, measurement methods based on ultrasonic signals [5,6], methods based on sensors [7,8] and methods based on machine vision. Measurement methods based on machine vision account for a large part of the non-contact measurement methods.…”