Robot monocular semi-dense map three-dimensional reconstruction method and reconstruction system

The invention discloses a robot monocular semi-dense map three-dimensional reconstruction method and method, and the method comprises the steps: 1, taking an image collected by a robot carrying a monocular camera as a key frame, calculating the depth of the key frame, and forming a semi-dense depth...

Full description

Saved in:
Bibliographic Details
Main Authors MA GUOJUN, DUAN YUNLONG, ZHU JIN, SHEN JIAWEI, MA HONGPENG, LENG JIAJUN
Format Patent
LanguageChinese
English
Published 01.03.2022
Subjects
CALCULATING
COMPUTING
COUNTING
IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
PHYSICS
Online AccessGet full text

Cover

More Information
Summary:The invention discloses a robot monocular semi-dense map three-dimensional reconstruction method and method, and the method comprises the steps: 1, taking an image collected by a robot carrying a monocular camera as a key frame, calculating the depth of the key frame, and forming a semi-dense depth map; combining the key frame with the semi-dense depth map to obtain semi-dense point cloud data; 2, extracting edge line segments of the key frame, wherein each edge line segment is a pixel sequence formed by a plurality of pixels; 3, fitting each edge line segment in the key frame into a first 3D line segment to form a first 3D line segment set LD; and 4, clustering the first 3D line segment set in the same key frame, fitting each clustered 3D line segment cluster into a second 3D line segment again, and representing image details by using the second 3D line segment in the key frame, thereby simplifying semi-dense point cloud data obtained by ORB-SLAM, and realizing three-dimensional reconstruction of a semi-dens
Bibliography:Application Number: CN202111422402