Phase principal value extraction method based on full convolutional neural network

Abstract

A phase principal value extraction method based on a full convolutional neural network comprises the following steps: 1) projecting a fringe pattern which is distributed according to sine and is required by pre-coding and a pre-coded fringe pattern on a computer to a to-be-detected object, and collecting the fringe pattern of the to-be-detected object by using an industrial camera; 2) constructinga full convolutional neural network model, setting a training parameter and a loss function, inputting the picture obtained in the step 1) into the neural network, and operating the neural network toobtain a required phase principal value; and (3) unwrapping the phase main value obtained in the step (2) by using a method based on quality map guidance to obtain an accurate phase value. The invention provides the phase main value extraction method based on the full convolutional neural network, which has the advantages that the image acquisition quantity is small, a training data set and a training process are not needed, and the precision is relatively high.

Description

technical field [0001] The invention relates to an image processing method, in particular to a phase principal value extraction method based on a fully convolutional neural network. Background technique [0002] With the rapid development of information technology and the diversification of social needs, the spatial three-dimensional measurement of object contours has been widely used in many fields such as industrial automatic inspection, product quality control, reverse design, biomedicine, virtual reality, cultural relics reproduction, anthropometry, etc. application. In particular, three-dimensional measurement using optical methods has been greatly developed in the past ten years because of its non-contact measurement, high measurement accuracy, large amount of data acquisition, and easy realization of optical, mechanical, and electrical integration under the computer. [0003] Phase measuring profilometry (PMP) is a three-dimensional measurement method that combines s...

AI Technical Summary

Problems solved by technology

[0004] In order to overcome the shortcomings of the existing image phase principal value extraction method, which require too many pictures, long shooting time, and low precision in actual measurement, in order to be able to extract the precise phase principal value in a single frame fringe image, based on In this way of thinking, the present invention proposes a phase principal value extraction method based on a fully convolutional neural network with higher precision

Images