Phase information extraction method based on convolutional neural network, storage medium and equipment

Abstract

The invention discloses a phase information extraction method based on a convolutional neural network, a storage medium and equipment. The method comprises the steps: constructing a Hypercolums convolutional neural network model, carrying out the analysis and prediction of an interference fringe image, and obtaining the phase data corresponding to the interference fringe image; four different mathematical functions of a sine / cosine data set, adopting a quadric surface data set, a waveform data set and a free-form surface data set respectively for generating phase data in a sample set, and thenobtaining an interference image corresponding to the phase data through an interference fringe image light intensity distribution formula; equally dividing the generated samples to jointly form N groups of data in the training set and M groups of data in the verification set; then, based on all the generated sample data, training a Hypercolums convolutional neural network model; and adopting a polynomial three-dimensional surface fitting method to eliminate local errors of the initial prediction result of the Hypercolums convolutional neural network, and realizing the phase extraction resultoptimization. The method is high in processing speed and high in phase extraction precision, and a single-frame interferogram phase extraction function can be realized.

Description

technical field [0001] The invention belongs to the technical field of interference fringe image processing in the field of three-dimensional contour precision measurement, and in particular relates to a phase information extraction method, storage medium and equipment based on a Hypercolumns convolutional neural network. Background technique [0002] Optical three-dimensional profilometry technology is widely used in industrial manufacturing, reverse engineering, aerospace, medical diagnosis and other fields. It has the characteristics of non-contact, high precision and high resolution, and is recognized as one of the most promising profilometry methods. There are many methods to realize optical profilometry, including time method, structured light method, projection method, interferometry and so on. In the interferometric measurement process, the test light is reflected by the workpiece and superimposed and interfered with the reference light to form an interference fringe...

AI Technical Summary

Problems solved by technology

However, the application of this method has certain limitations, and the phase information cannot be accurately extracted for images with closed fringes

Images