Passive polarization three-dimensional reconstruction method based on CNN correction

Abstract

The invention relates to a passive polarization three-dimensional reconstruction method based on CNN correction. The method comprises the following steps: acquiring a plurality of polarization imagesof a to-be-reconstructed object at different polarization angles; obtaining the first surface normal gradient field of the to-be-reconstructed object according to the polarization image; acquiring thesecond surface normal gradient field of the to-be-reconstructed object by using a convolutional neural network; correcting the first surface normal gradient field by using the second surface normal gradient field to obtain the third surface normal gradient field of the to-be-reconstructed object; and realizing three-dimensional reconstruction of the to-be-reconstructed object according to the third surface normal gradient field. According to the three-dimensional reconstruction method, the problem of three-dimensional reconstruction result distortion caused by binaryzation of the normal azimuth angle is solved, three-dimensional reconstruction can be realized under natural light, the application scene is wide, and the cost is relatively low.

Description

technical field [0001] The invention belongs to the technical field of computational imaging, and in particular relates to a passive polarization three-dimensional reconstruction method based on CNN correction. Background technique [0002] 3D reconstruction technology can provide depth information that cannot be obtained from 2D images, and is an important way to obtain data in today's information age. Through 3D reconstruction technology, we can obtain complete 3D information including structure, texture, scale and other objects. With the wide application of 3D reconstruction technology in the fields of face anti-spoofing, face recognition, facial expression animation and mobile face payment, especially in the field of mobile face payment, people's detailed information about the reconstruction results of 3D reconstruction technology The requirements for richness, accuracy and high precision are getting higher and higher. [0003] At present, for traditional 3D imaging me...

AI Technical Summary

Problems solved by technology

[0003] At present, for traditional 3D imaging methods such as Time of Flight (TOF), binocular shadow shape recovery (Shape from Shading, SFS) and laser scanning, etc., due to the complexity of imaging equipment, poor information interpretation and imaging conditions Harsh and other factors greatly limit its development

Therefore, the technicians proposed a method of monocular polarization 3D reconstruction. The method of monocular polarization 3D reconstruction can only deal with objects with relatively simple and single target shapes. When facing a human face, the light intensity information of the human face is very complicated. And there is the problem that the normal azimuth angle is uncertain, and it is not possible to simply use the gradient field of light intensity to correct the gradient field obtained by polarization

In addition, this technology requires that the incident light source be uniformly irradiated and the direction of the irradiated light needs to be obtained. Therefore, the integrating sphere needs to be used as an active light source to irradiate the target to reconstruct the result, which makes the method costly to implement, the application range is small, and the reconstruction is difficult. Higher, lower reconstruction accuracy

Images