Research On The Key Problems Of 3D Surface Reduction In Surface Topography Measurement By Optical Fiber Sensor

At present,the 3D visualization measurement and inspection of the surface morphology has not been well solved yet.This is mainly reflected in the fact such as the low efficiency or accuracy of measurement and the complexity of surface texture modeling and processing.Due to the advent of low-loss optical fibers and the development of related sensors,optical fiber sensing technology has been gradually applied to surface measurement.Among the fiber optical sensor,the reflective intensity modulation type(RIM-FOS)is simple and easy to design.It is also the first optical fiber sensing technology.Therefore,it has important theoretical significance and application value to research how to apply RIM-FOS in 3D surface topography measurement.On the basis of elaborating the background,purpose,research value and current research direction of the research,this paper discusses the measurement theory from two aspects,parameter measurement and topography measurement.Therefore,a measurement model and a reduction scheme for three-dimensional surface reduction with RIM-FOS are proposed,and its feasibility is verified.Due to the summary and analysis of research on parameter measurement,we present some key problems in the 3D surface reduction,such as the interference problem,the deceptive problem and the lateral resolution problem.To solve these problems,this paper adopts neural network algorithm to weaken the influence of light source fluctuation on the measurement,and improves neural network by a new improved bat algorithm;this paper study to introduce recurrent neural network(RNN)into the field of measurement to solve the problem of probe deception,and improves the system's ability to integrate the overlapping information brought by continuous sampling;And the multichannel quadrant optical fiber probe is proposed to further enrich the available surface information and improve the transverse resolution of the probe.According to the determined scheme,we carry out the overall design of the measurement system.Calibrating and estimating the error of every components of the measurement system,then considering its effect on the accuracy of the measurement.Finally,the testing software and data processing program of the measurement system is compiled,which verifies the basic functions of the measurement system,and realizes the displacement calibration and neural network compensation of the tested workpiece.And 3D morphology of the objects are reduced by using the calibration parameters and the scanned surface data.The experimental results show that the optimized neural network algorithm can effectively weaken the influence of light source fluctuation,and the repeatability of the sensor has increased by 50%.The surface reduction scheme can effectively identify and reduce the three-dimensional surface with the step and the ring textures,and the measurement error of the step height identified in the 3D map is reduced by 20 um.