A Specular Object Measurement Method Based on Polarization Image and Machine Learning

Abstract

The invention belongs to the technical field of visual measurement, in particular to a method for measuring high-gloss objects based on polarization images and machine learning. The present invention introduces the linear polarizer and the neural network algorithm into the visual measurement system, obtains a plurality of images with different highlight intensity by changing the polarization angle, extracts highlight pixels from them, and uses its V channel value as the training sample of the neural network algorithm. Further suppress the highlight, and finally import the obtained highlight suppression image into the stereo matching algorithm; the invention can perform more accurate visual measurement on the object to be measured with a surface with high reflection characteristics, and improve the visual measurement by suppressing the highlight phenomenon caused by the external light source. The accuracy rate of stereo matching in the process, and finally obtain more accurate measurement results. Experiments prove that the invention effectively suppresses the highlight of the object image, improves the stereo matching accuracy in visual measurement, and obtains more reliable depth information and more accurate measurement results.

Description

technical field [0001] The invention belongs to the technical field of visual measurement, and in particular relates to a method for three-dimensional measurement of an object to be measured with a highly reflective surface based on polarized images and neural network learning. Background technique [0002] Vision measurement technology is widely used in industrial production and daily life. It obtains the three-dimensional information of the object to be measured by processing the images collected by the camera. Therefore, image quality is of great significance to the visual measurement system, and for objects with highly reflective surfaces, the specular reflection of the surface will produce high-light areas of the image, covering the external texture and color characteristics of the measured object, and finally or reduce the image Quality; the commonly used highlight suppression method is mainly based on the two-color reflection model, and the algorithm is improved to pr...

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to overcome the problems that the traditional highlight suppression method easily causes object color information distortion, texture information distortion, and structural information distortion, etc., and provides a method for measuring highlight objects based on polarization images and machine learning, which suppresses the highlight area of ​​the object image, and at the same time Reduce the distortion of its color, texture and structure information

Images