Visual system based on photometric stereo vision and 3D reconstruction method thereof

Abstract

The invention discloses a visual system based on photometric stereo vision. The system includes a spot generator which can project dense spots, at least three LEDs of which each forms a point light source, two cameras which constitute a pair of stereo cameras, and a control circuit which is in communication connection with the spot generator, the at least three LEDs and the two cameras in order to make the spot generator project dense spots, make the at least three LEDs turned on in turn and only one LED turned on at the same time, make the two cameras triggered simultaneously to collect an image of the surface of an object under the irradiation of the spot generator and make one of the two cameras triggered to collect an image of the surface of the object under the irradiation of the LEDs. The invention further discloses a corresponding 3D reconstruction method. The visual system of the invention is small, and has a fast measurement mode and an accurate measurement mode.

Description

technical field [0001] The invention relates to the field of optical three-dimensional measurement, in particular to a vision system based on photometric stereo vision and a three-dimensional reconstruction method thereof. Background technique [0002] 3D reconstruction technology has a wide range of applications in 3D modeling, reverse engineering, 3D printing, anthropometry, human-computer interaction, film and animation production, etc. Systems based on sinusoidal fringes, which have high measurement accuracy, are expensive, making them difficult to use in consumer applications. In recent years, cheap 3D imaging sensors such as Kinect have greatly reduced the cost of 3D data acquisition, but these cheap sensors have limited precision and cannot directly obtain high-precision 3D reconstruction. At present, in order to improve the accuracy of 3D reconstruction, one type of method is to perform weighted average of 3D point clouds observed at multiple points, and the other t...

AI Technical Summary

Problems solved by technology

In addition, the photometric stereo vision system based on point light sources has a more compact structure than the stereo vision system based on parallel light, but in order to ensure the accuracy of normal direction estimation, a large distance needs to be maintained between each point light source, which makes Stereo vision system is bulky

Some stereo vision methods assume that the target surface has a uniform reflectivity, which works well for the reconstruction of targets with uniform reflectivity, but is not suitable for targets with multiple reflectivities

Images