US 12,444,066 B2
Model generation method and apparatus based on multi-view panoramic image
Dan Chen, Shenzhen (CN); Yuyao Zhang, Shenzhen (CN); and Zhigang Tan, Shenzhen (CN)
Assigned to KANDAO TECHNOLOGY CO., LTD., Shenzhen (CN)
Appl. No. 18/041,413
Filed by KANDAO TECHNOLOGY CO., LTD., Shenzhen (CN)
PCT Filed Apr. 19, 2021, PCT No. PCT/CN2021/088002
§ 371(c)(1), (2) Date Feb. 13, 2023,
PCT Pub. No. WO2021/244161, PCT Pub. Date Dec. 9, 2021.
Claims priority of application No. 202010501846.9 (CN), filed on Jun. 4, 2020.
Prior Publication US 2023/0237683 A1, Jul. 27, 2023
Int. Cl. G06T 7/55 (2017.01); G06T 7/80 (2017.01)
CPC G06T 7/55 (2017.01) [G06T 7/80 (2017.01); G06T 2207/20084 (2013.01); G06T 2207/20221 (2013.01)] 15 Claims
OG exemplary drawing
 
1. A model generation method based on a multi-view panoramic image, comprising:
obtaining a reference image and a plurality of corresponding source images, wherein the source images and the reference image have a coinciding viewing angle;
obtaining source camera parameters of the source images and a reference camera parameter of the reference image, and calculating an image rectification rotation matrix of the source images and the reference image based on the source camera parameters and the reference camera parameter;
extracting a reference image feature of the reference image and source image features of the source images, and calculating cost volumes of the reference image and the source images based on the reference image feature and the source image features;
performing coordinate system transformation on the cost volumes by using the image rectification rotation matrix to obtain rectified cost volumes of the source images and the reference image;
performing a fusion operation on the rectified cost volumes of the plurality of source images corresponding to the reference image to obtain a final cost volume;
calculating a phase difference distribution estimation of the reference image under a set resolution based on the final cost volume, and calculating an estimated phase difference under the set resolution;
fusing an estimated phase difference of the reference image under each layer of resolution to obtain a final phase difference of the reference image; and
generating a depth map of the reference image based on the final phase difference of the reference image, and constructing a corresponding stereo vision model according to the depth map.