CN-121616762-B - Three-dimensional scanning method, scanning device, computing device, and nonvolatile storage medium

Abstract

The application discloses a three-dimensional scanning method, a scanning device, a computing device and a nonvolatile storage medium. The method comprises the steps of obtaining a plurality of grids generated by scanning an object in multiple rounds by scanning equipment, aligning the grids to the same coordinate system, splicing the aligned grids to obtain spliced grids, determining the visibility of each patch in the spliced grids corresponding to texture images obtained in the multiple rounds of scanning, and rendering the patches by using at least one visible texture image. The application solves the technical problem that the background and the scanned object are easy to be misplaced when the multi-working-procedure files generated by multiple rounds of scanning are fused in the dynamic environment in the related technology.

Inventors

• LUO YANGJUN
• Jiang tengfei

Assignees

• 先临三维科技股份有限公司

Dates

Publication Date

20260512

Application Date

20260130

Claims (12)

1. 1. A three-dimensional scanning method, comprising: acquiring a plurality of grids generated by scanning equipment for carrying out multi-round scanning on an object, and aligning the grids to the same coordinate system, wherein the multi-round scanning is completed under at least two space environments, and the at least two space environments are at least partially different; The multiple grids after being aligned are spliced to obtain a spliced grid, and the visibility of each panel in the spliced grid corresponding to the texture image acquired in the multi-round scanning is determined, wherein the method comprises the steps of, for a first panel of a current grid in the spliced grid, responding to a second panel capable of searching at least one other grid around the first panel, determining that the first panel is visible to a first texture image corresponding to the first panel and a second texture image corresponding to the second panel, or responding to a second panel incapable of searching at least one other grid around the first panel, determining that the first panel is visible only to the first texture image corresponding to the first panel, and The patch is rendered using the at least one texture image that is visible.
2. 2. The method of claim 1, further comprising: aligning the multiple grids under the same coordinate system according to the relative pose transformation relation among the multiple grids, and The texture images are aligned to the same coordinate system.
3. 3. The method of claim 2, wherein the same coordinate system is a coordinate system of a grid generated by any one of the multiple rounds of scanning.
4. 4. The method of claim 1, wherein each round of scanning comprises a plurality of scanning processes, a point cloud image and a texture image of the object are acquired in each scanning process using the scanning device, and the grid is generated from the plurality of point cloud images acquired in each round of scanning.
5. 5. The method of claim 1, wherein the multiple rounds of scanning comprise a first round of scanning in a first spatial environment and a second round of scanning in a second spatial environment, the plurality of grids comprising a first grid generated by the first round of scanning and a second grid generated by the second round of scanning, wherein the first spatial environment is at least partially different from the second spatial environment.
6. 6. The method of claim 1, wherein determining the visibility of each patch in the stitched grid corresponding to a texture image acquired in the multiple passes comprises: For a first panel of a current grid in the spliced grid, determining a position relationship between the first panel and a second panel of at least one other grid; And determining the visibility of the first panel to the second texture image corresponding to the second panel according to the position relation.
7. 7. The method of claim 6, wherein determining the visibility of the first panel to the second texture image corresponding to the second panel in accordance with the positional relationship comprises: determining a first panel center and a first normal vector of the first panel and a second panel center and a second normal vector of the second panel; determining a center distance between the first panel center and the second panel center, and determining a normal cosine value between the first normal vector and the second normal vector; determining that the first panel is invisible to the second texture image corresponding to the second panel in response to the center distance being greater than a preset distance threshold or the normal cosine value being less than a preset cosine value threshold; And determining the visibility of the first panel to the second texture image corresponding to the second panel according to the projection relationship between the center of the second panel and the first panel in response to the center distance not being greater than the preset distance threshold and the normal cosine value not being less than the preset cosine value threshold.
8. 8. The method of claim 7, wherein determining the visibility of the first panel to the second texture image corresponding to the second panel as a function of the projected relationship between the second panel center and the first panel, comprises: Projecting the center of the second panel to a target plane along the direction of the second normal vector to obtain a projection intersection point, wherein the target plane is a plane taking the center of the first panel as the center and taking the direction of the first normal vector as the normal direction; Determining that the first panel is visible to the second texture image corresponding to the second panel in response to the projection intersection being located inside the first panel, and And determining that the first panel is not visible to the second texture image corresponding to the second panel in response to the projection intersection point not being located inside the first panel.
9. 9. The method of claim 1, wherein rendering the patch using the at least one texture image that is visible comprises: Determining one or more patches corresponding to the visible texture image; And performing projection rendering on the one or more patches according to relevant parameters of the visible texture image, wherein the relevant parameters comprise at least one of pose parameters of the texture image and texture camera internal parameters.
10. 10. A scanning device comprising a memory and a processor for executing a program stored in the memory, wherein the program is run to perform the three-dimensional scanning method of any one of claims 1 to 9.
11. 11. A computing device comprising a memory and a processor for executing a program stored in the memory, wherein the program is run to perform the three-dimensional scanning method of any one of claims 1 to 9.
12. 12. A non-volatile storage medium comprising a stored computer program, wherein the device in which the non-volatile storage medium resides performs the three-dimensional scanning method of any one of claims 1 to 9 by running the computer program.

Description

Three-dimensional scanning method, scanning device, computing device, and nonvolatile storage medium Technical Field The present application relates to the field of three-dimensional imaging technologies, and in particular, to a three-dimensional scanning method, a scanning device, a computing device, and a nonvolatile storage medium. Background In the related art, multiple rounds of three-dimensional scanning are required to obtain multi-frame point clouds and textures for three-dimensional reconstruction of the same object, and a direct fusion strategy is generally adopted for texture and point cloud fusion, namely, a texture image obtained by each scanning is fused onto a constructed final three-dimensional grid model according to corresponding point cloud data. This process relies on a comparison of the point cloud frame depth to the mesh rendering depth to determine whether to select textures, in order to avoid regions of too great a depth difference being fused, thereby reducing misalignment. However, this solution presents a significant limitation in the case of multiple rounds of scanning in a dynamic environment, and is prone to problems of misalignment of the background and the scanned object. Disclosure of Invention The embodiment of the application provides a three-dimensional scanning method, scanning equipment, computing equipment and a nonvolatile storage medium, which at least solve the technical problem that a background and a scanning object are easy to be misplaced when multi-procedure files generated by multiple rounds of scanning are fused in a dynamic environment in the related technology. According to one aspect of the embodiment of the application, a three-dimensional scanning method is provided, which comprises the steps of obtaining a plurality of grids generated by scanning equipment for carrying out multi-round scanning on an object, aligning the grids to the same coordinate system, splicing the aligned grids to obtain a spliced grid, determining the visibility of each patch in the spliced grid corresponding to texture images obtained in the multi-round scanning, and rendering the patches by using at least one visible texture image. Optionally, the method further comprises aligning the grids under the same coordinate system according to the relative pose transformation relation among the grids, and aligning the texture image under the same coordinate system. Optionally, the same coordinate system is the coordinate system of the grid generated by any one of the multiple scans. Optionally, each round of scanning includes a plurality of scanning processes, a point cloud image and a texture image of the object are acquired in each scanning process using the scanning device, and a grid is generated according to the plurality of point cloud images acquired in each round of scanning. Optionally, the multiple passes are performed in at least two spatial environments, at least partially different from each other. Optionally, the multiple rounds of scanning include a first round of scanning performed in a first spatial environment and a second round of scanning performed in a second spatial environment, the multiple grids including a first grid generated by the first round of scanning and a second grid generated by the second round of scanning, wherein the first spatial environment is at least partially different from the second spatial environment. Optionally, determining the visibility of each panel in the spliced grid corresponding to the texture image acquired in the multi-pass scanning comprises, for a first panel of a current grid in the spliced grid, determining that the first panel is visible to a first texture image corresponding to the first panel and a second texture image corresponding to the second panel in response to a second panel surrounding the first panel capable of searching for at least one other grid, and for a first panel of the current grid in the spliced grid, determining that the first panel is visible to only the first texture image corresponding to the first panel in response to a second panel surrounding the first panel incapable of searching for at least one other grid. Optionally, determining the visibility of each panel in the stitching grid corresponding to the texture image acquired in the multi-pass scanning includes determining, for a first panel of a current grid in the stitching grid, a positional relationship between the first panel and a second panel of at least one other grid, and determining, according to the positional relationship, the visibility of the first panel to a second texture image corresponding to the second panel. Optionally, determining the visibility of the first panel to the second texture image corresponding to the second panel according to the position relation comprises determining a first panel center and a first normal vector of the first panel and a second panel center and a second normal vector of the second panel, determinin