Multi-Aperture Synthetic 3D Imaging Technology

Three dimensional technology with high definition and large field of view can reconstruct the real scene,which is widely used in many fields.Increasing the system aperture can effectively expand the field of view,but it will increase the system weight and processing difficulty.To solve this problem,small aperture is used to detect the equivalent large aperture of synthetic imaging.Aiming at the matching accuracy of key technologies of multi aperture synthetic 3D imaging,two parts of multi aperture image mosaic and 3D reconstruction are studied.In the aspect of image mosaic,an image feature point matching algorithm based on ORB and HSV is proposed.The preprocessing of the algorithm combines two methods:bilateral filtering and mean filtering to highlight the edge information and reduce the noise interference.ORB algorithm was used to extract image feature points.K-d Tree algorithm and Hamming distance were used for rough matching of feature points,and HSV information was used to further screen feature point pairs.The experimental results showed that the pre-processing evaluation index value of the proposed algorithm was the best,and the average matching accuracy rate was as high as 98.19%,which was 12.60 percentage points higher than that of unreferenced HSV information screening.Moreover,the final stitching result was of better quality and its NIQE index value was smaller.In the aspect of 3D reconstruction,the dense 3D imaging method based on epipolar matching constraint was studied,and built the experimental platform of binocular parallel optical axis structure system.The algorithm used the Zhang Zhengyou calibration method to calibrate the camera,the region matching method to achieve the feature recombination and matching,and chose the epipolar constraint for feature matching to complete the 3d reconstruction of the target scene.In the experiment,the collected data and images were simulated.Finally,the running results of the algorithm studied in the real scene were reproduced,and the relative error of the morphology accuracy was less than 5%,which had a high degree of reduction and feasibility.Figure 44;Table 8;Reference 50...