Urban scene reconstruction method, device, computer equipment and storage medium

Abstract

The present application relates to an urban scene reconstruction method, device, computer equipment and storage medium. The method includes: acquiring a target image frame corresponding to a target city scene, and adjacent image frames of the target image frame; performing object positioning on the target image frame according to the target image frame and the adjacent image frame, Obtain the object point cloud, object image and coordinate transformation matrix corresponding to the target object; determine the global characteristics of the target object and the surface parameters to be selected according to the object point cloud and the object image; The global features determine the surface features to be selected; determine the plane combination matrix of the target object according to the surface features to be selected; rebuild the target city according to the plane combination matrix, the surface parameters to be selected and the coordinate transformation matrix The scene 3D model of the scene. This method can take into account the reconstruction efficiency and accuracy of urban scene reconstruction.

Description

technical field [0001] The present application relates to the technical field of computer vision, in particular to a method, device, computer equipment and storage medium for urban scene reconstruction. Background technique [0002] With the development of computer vision technology, 3D scene reconstruction technology based on 2D image reconstruction of 3D models has developed. 3D scene reconstruction technology is widely used in 3D object recognition, geographic geological survey, automatic driving, urban planning, navigation, and obstacle avoidance. and exploration scenarios. Therefore, how to reconstruct a 3D model based on a 2D image is a problem worthy of attention. [0003] At present, the 3D point cloud is usually reconstructed based on the 2D image, and then complex post-processing is performed on the reconstructed 3D point cloud to reconstruct the 3D model, which has the problem of low reconstruction efficiency. Only 3D models composed of sparse point clouds can b...

AI Technical Summary

Problems solved by technology

[0003] At present, the 3D point cloud is usually reconstructed based on the 2D image, and then complex post-processing is performed on the reconstructed 3D point cloud to reconstruct the 3D model, which has the problem of low reconstruction efficiency. Only 3D models composed of sparse point clouds can be reconstructed, and there is a problem of low reconstruction accuracy

Therefore, the current 3D scene reconstruction method has the problem that the reconstruction accuracy and efficiency cannot be balanced, and it is difficult to apply to the reconstruction of urban scenes

Images