Mesh reconstruction method and apparatus for transparent object, computer device and storage medium

1. A mesh reconstruction method for a transparent object, comprising:

acquiring object images of the transparent object at multiple capture view angles and calibration information corresponding to an image capture device, the image capture device being configured to capture the object images;

generating an initial mesh model corresponding to the transparent object according to the object images acquired at the multiple capture view angles;

determining a light ray refraction loss corresponding to an emergent light ray of the image capture device according to the calibration information, and determining a model loss corresponding to the initial mesh model according to the light ray refraction loss; and

reconstructing the initial mesh model according to the model loss, to obtain a target mesh model corresponding to the transparent object;

wherein the generating the initial mesh model corresponding to the transparent object according to the object images acquired at the multiple capture view angles comprises:

extracting a plurality of contour images corresponding to the transparent object from the object images acquired at the multiple capture view angles;

performing space carving according to the plurality of contour images, to obtain a three-dimensional convex hull corresponding to the transparent object; and

acquiring a target mesh parameter and meshing the three-dimensional convex hull according to the target mesh parameter, to obtain the initial mesh model corresponding to the transparent object;

the mesh reconstruction method further comprising:

acquiring projection contours of the initial mesh model at the multiple capture view angles; and

determining an object contour loss corresponding to the initial mesh model according to a coincidence degree of a projection contour and a corresponding contour image;

wherein the determining the model loss corresponding to the initial mesh model according to the light ray refraction loss comprises:

acquiring a light ray refraction weight corresponding to the light ray refraction loss and a contour weight corresponding to the object contour loss; and

weighting the light ray refraction loss and the object contour loss according to the light ray refraction weight and the contour weight, to obtain the model loss corresponding to the initial mesh model.

2. The method according to claim 1 , wherein the determining the light ray refraction loss corresponding to the emergent light ray of the image capture device according to the calibration information comprises:

determining a first position coordinate corresponding to each of a plurality of emergent light rays of the image capture device according to the calibration information and the object images, there being a corresponding relationship between the first position coordinate and the emergent light ray;

calculating a second position coordinate corresponding to each of the plurality of emergent light rays according to the initial mesh model; and

determining the light ray refraction loss corresponding to the initial mesh model according to a coordinate distance between the first position coordinate and the second position coordinate corresponding to each of the plurality of emergent light rays.

3. The method according to claim 1 , further comprising:

after reconstructing the initial mesh model according to the model loss,

returning to the step of acquiring the target mesh parameter, meshing, according to the acquired target mesh parameter, a mesh model obtained by the reconstructing, and recording a first number of returns; and

stopping returning to the step of acquiring the target mesh parameter when the first number of returns reaches a first threshold.

4. The method according to claim 1 , further comprising:

after reconstructing the initial mesh model according to the model loss,

returning to the step of determining the light ray refraction loss corresponding to the emergent light ray of the image capture device according to the calibration information, and recording a second number of returns; and

stopping returning to the step of determining the light ray refraction loss corresponding to the emergent light ray of the image capture device according to the calibration information when the second number of returns reaches a second threshold.

5. A computer device comprising one or more processors and a memory storing computer-readable instructions, wherein one or more processors, when executing the computer-readable instructions, perform following steps of:

acquiring object images of the transparent object at multiple capture view angles and calibration information corresponding to an image capture device, the image capture device being configured to capture the object images;

generating an initial mesh model corresponding to the transparent object according to the object images acquired at the multiple capture view angles;

determining a light ray refraction loss corresponding to an emergent light ray of the image capture device according to the calibration information, and determining a model loss corresponding to the initial mesh model according to the light ray refraction loss; and

reconstructing the initial mesh model according to the model loss, to obtain a target mesh model corresponding to the transparent object;

wherein the one or more processors, when executing the computer-readable instructions, further perform following steps of:

extracting a plurality of contour images corresponding to the transparent object from the object images acquired at the multiple capture view angles;

performing space carving according to the plurality of contour images, to obtain a three-dimensional convex hull corresponding to the transparent object;

acquiring a target mesh parameter and meshing the three-dimensional convex hull according to the target mesh parameter, to obtain an initial mesh model corresponding to the transparent object;

acquiring projection contours of the initial mesh model at the multiple capture view angles;

determining an object contour loss corresponding to the initial mesh model according to a coincidence degree of a projection contour and a corresponding contour image;

acquiring a light ray refraction weight corresponding to the light ray refraction loss and a contour weight corresponding to the object contour loss; and

weighting the light ray refraction loss and the object contour loss according to the light ray refraction weight and the contour weight, to obtain the model loss corresponding to the initial mesh model.

6. The computer device according to claim 5 , wherein the one or more processors, when executing the computer-readable instructions, further perform following steps of

determining a first position coordinate corresponding to each of a plurality of emergent light rays of the image capture device according to the calibration information and the object images, there being a corresponding relationship between the first position coordinate and an emergent light ray;

calculating a second position coordinate corresponding to each of the plurality of emergent light rays according to the initial mesh model; and

determining the light ray refraction loss corresponding to the initial mesh model according to a coordinate distance between the first position coordinate and the second position coordinate corresponding to each of the plurality of emergent light rays.

7. The computer device according to claim 5 , wherein the one or more processors, when executing the computer-readable instructions, further perform following steps of:

returning to the step of acquiring the target mesh parameter, meshing, according to the acquired target mesh parameter, a mesh model obtained by the reconstructing, and recording a first number of returns; and

stopping returning to the step of acquiring the target mesh parameter when the first number of returns reaches a first threshold.

8. One or more non-transitory computer-readable storage media storing computer-readable instructions, wherein one or more processors, when executing the computer-readable instructions, perform following steps of:

acquiring object images of the transparent object at multiple capture view angles and calibration information corresponding to an image capture device, the image capture device being configured to capture the object images;

generating an initial mesh model corresponding to the transparent object according to the object images acquired at the multiple capture view angles;

determining a light ray refraction loss corresponding to an emergent light ray of the image capture device according to the calibration information, and determining a model loss corresponding to the initial mesh model according to the light ray refraction loss; and

reconstructing the initial mesh model according to the model loss, to obtain a target mesh model corresponding to the transparent object;

wherein the one or more processors, when executing the computer-readable instructions, further perform following steps of:

extracting a plurality of contour images corresponding to the transparent object from the object images acquired at the multiple capture view angles;

performing space carving according to the plurality of contour images, to obtain a three-dimensional convex hull corresponding to the transparent object;

acquiring a target mesh parameter and meshing the three-dimensional convex hull according to the target mesh parameter, to obtain the initial mesh model corresponding to the transparent object;

acquiring projection contours of the initial mesh model at the multiple capture view angles;

determining an object contour loss corresponding to the initial mesh model according to a coincidence degree of a projection contour and a corresponding contour image;

acquiring a light ray refraction weight corresponding to the light ray refraction loss and a contour weight corresponding to the object contour loss; and

weighting the light ray refraction loss and the object contour loss according to the light ray refraction weight and the contour weight, to obtain the model loss corresponding to the initial mesh model.

9. The storage medium according to claim 8 , wherein the one or more processors, when executing the computer-readable instructions, perform following steps of:

determining a first position coordinate corresponding to each of a plurality of emergent light rays of the image capture device according to the calibration information and the object images, there being a corresponding relationship between the first position coordinate and an emergent light ray;

calculating a second position coordinate corresponding to each of the plurality of emergent light rays according to the initial mesh model; and

determining a light ray refraction loss corresponding to the initial mesh model according to a coordinate distance between the first position coordinate and the second position coordinate corresponding to each of the plurality of emergent light rays.

10. The storage medium according to claim 8 , wherein the one or more processors, when executing the computer-readable instructions, perform following steps of:

returning to the step of acquiring the target mesh parameter, meshing, according to the acquired target mesh parameter, a mesh model obtained by the reconstructing, and recording a first number of returns; and

stopping returning to the step of acquiring the target mesh parameter when the first number of returns reaches a first threshold.