Urban scene grid model optimization method with additional multi-line-of-sight feature constraints

Abstract

The invention provides an urban scene grid model optimization method with additional multi-line-of-sight feature constraints. The method comprises the following steps: carrying out straight line segment extraction on each frame of multi-view image in a multi-view image sequence to obtain a plurality of two-dimensional straight line segments, calculating a binary descriptor, and matching the two-dimensional straight line segments between two continuous frames of multi-view images by utilizing the binary descriptor to obtain a plurality of groups of homonymous two-dimensional straight line segments; selecting the longest two-dimensional straight line segment in each group of homonymous two-dimensional straight line segments to construct a longest two-dimensional straight line segment set, back-projecting all pixels of all line segments in the set to the three-dimensional grid model through camera parameters, and all vertex sets of the triangular patch where all back projection points arelocated being multiple edge vertex sets of the three-dimensional grid model; and after an edge vertex set is obtained, constructing an energy function about the vertex of the three-dimensional grid model, and gradually optimizing the vertex position of the three-dimensional grid model through a gradient descent method. The invention has the maximum linear information amount, and the precision andthe aesthetic degree of the edge of the model can be recovered to the maximum extent.

Description

technical field [0001] The invention belongs to the field of remote sensing surveying and mapping, and in particular relates to an urban scene grid model optimization method with additional multi-line-of-sight feature constraints. Background technique [0002] 3D reconstruction based on multi-view stereo reconstruction technology is one of the important ways to obtain the triangular mesh model of the object. Its main process includes dense matching to generate point cloud, point cloud network to generate model, and model texture mapping. However, due to the limitation of the dense matching algorithm, the final generated point cloud will inevitably contain a certain number of noise points, especially in the edge area of ​​the reconstructed object. This kind of noise is difficult to remove by means of point cloud filtering, etc., resulting in false ups and downs in the reconstructed 3D urban real scene model, especially distortion or excessive smoothness in the edge area of ​​...

AI Technical Summary

Problems solved by technology

To achieve the above functions, the technical problem to be solved by the present invention: the edge of the three-dimensional grid model is easily affected by noise, and the phenomenon of excessively smooth distortion occurs. This is due to the limitation of the dense matching algorithm, and the optimization method based on the consistency of image gray can't solve the problem

Images