A multi-camera reconstruction method based on dlp surface structured light

Abstract

The invention provides a multi-camera reconstruction method based on DLP surface structured light, including: building a four-eye structured light detection system, calibrating the camera, generating Gray code and phase shift code pattern at the PC end, controlling the projection of the DLP projector, and triggering the camera with a level signal Shoot the modulated image of the coding pattern; correct the captured image and perform denoising processing; decode and unwrap the denoised Gray code pattern and phase shift coding pattern respectively to calculate the truncated phase, and combine the results of the two for absolute Calculate the phase value; perform stereo matching to obtain the point cloud information of the object to be measured; filter the point clouds obtained by the two sets of binocular structured light systems, and perform preliminary splicing of the two sets of point clouds through rotation and translation transformation; The point cloud is fused and processed, and the surface is reconstructed. The present invention completes the high-precision encoding process by combining the Gray code with the phase-shift code, improves the accuracy of encoding and decoding and the accuracy of the final reconstruction, and meets the requirements of detection in the industry.

Description

technical field [0001] The present invention relates to the technical field of three-dimensional reconstruction technology, in particular to a multi-camera reconstruction method based on DLP surface structured light. Background technique [0002] Three-dimensional reconstruction technology can be divided into two types: contact and non-contact. The method of contact three-dimensional reconstruction mainly obtains the point cloud data of the contour through the direct contact between the sensor and the target. The detection accuracy of this method is high, but the overall efficiency is low, and at the same time, certain damage may be caused to the object to be measured during the contact process. The non-contact 3D reconstruction technology mainly uses light, sound, electromagnetic waves, etc. to contact the target to obtain 3D information of the object, or perform 3D reconstruction through the principle of multi-view geometry, and use multiple cameras to shoot simultaneously...

AI Technical Summary

Problems solved by technology

The photometric stereo method uses a camera and multiple light sources with the same light intensity, changes the direction of the light source, and takes images of the object under different illumination conditions at the same time, and performs three-dimensional reconstruction according to the surface normal information. Poor anti-interference

The time-of-flight method calculates the depth distance on the surface of the measured object by recording the beam propagation time. This method can avoid the influence of obstructions, but the overall accuracy is low

Laser triangulation is to use the camera to collect the line laser of the surface deformation of the object, and then obtain the three-dimensional information of the object according to the principle of triangulation, but because the projection pattern is a straight line, the measurement efficiency is low

The structured light projection method is a 3D reconstruction method that is widely used at present. It has the characteristics of high precision and fast measurement speed. It consists of a structured light projector, but the field of view measured by this method is limited. If you want to get the complete point cloud information, you need to take multiple shots, and some highly reflective objects are likely to cause the lack of point clouds. Therefore, this project plans to use a four-eye structure. Optical technology, through the projection of gray code and phase shift coding pattern to realize the precise coding of each pixel, and combined with point cloud splicing and fusion technology for 3D reconstruction

Images