Method and system for geometric correction based on image tracking

1 . A system for geometric correction based on image tracking, comprising:

at least one storage device including a set of instructions; and

at least one processor in communication with the at least one storage device, wherein when executing the set of instructions, the at least one processor is directed to perform operations including:

obtaining a plurality of projection images of a geometric model, each of the plurality of projection images including a projection marker corresponding to an original marker in the geometric model, wherein one of the plurality of projection images corresponding to one of the plurality of imaging angles;

obtaining an initial correspondence between the original markers in the geometric model and the projection markers in one of the plurality of projection images;

determining a target correspondence between the original markers in the geometric model and the projection markers in the tracked projection image by tracking, based on the initial correspondence and a correlation of the plurality of projection images, the projection markers in the plurality of projection images; and

performing, based on the target correspondence, a geometric correction.

2 . The system of claim 1 , wherein obtaining the plurality of projection images of the geometric model includes:

obtaining a imaging parameter of the geometric model; and

generating the plurality of projection images of the geometric model, based on the imaging parameter.

3 . The system of claim 2 , further comprising:

determining, based on the imaging parameter, an angle difference of at least one group of adjacent scans; and

determining, based on the angle difference, at least one of the plurality of imaging angles.

4 . The system of claim 1 , wherein obtaining an initial correspondence between the original markers in the geometric model and the projection markers in one of the plurality of projection images includes:

determining the projection image satisfying a first preset condition as a reference projection image; and

obtaining the initial correspondence between the original markers in the geometric model and the projection markers in the reference projection image.

5 . The system of claim 4 , further comprising:

in response to determining that none of the plurality of projection images satisfy the first preset condition, extracting an similar projection image satisfying a second preset condition;

determining, based on the similar projection image, an optimized angle; and

generating an optimized projection image by scanning, based on the optimized angle, the geometric model; and

determining the optimized projection image satisfying the first preset condition as the reference projection image.

6 . The system of claim 5 , wherein the optimized angle is determined based on a deviation degree, the deviation degree being a deviation value between a position of the projection marker in the one of the plurality of projection images and a preset position corresponding to the first preset condition.

7 . The system of claim 4 , wherein obtaining the initial correspondence between the original markers in the geometric model and the projection markers in the one of the plurality of projection images includes:

selecting, from the plurality of projection images by an automatic detection algorithm, the reference projection image that satisfies the preset feature; and

automatically recognizing the initial correspondence between the projection markers in the reference projection image and the original markers.

8 . The system of claim 1 , wherein determining a target correspondence between the original markers in the geometric model and the projection markers in the tracked projection image by tracking, based on the initial correspondence and a correlation of the plurality of projection images, the projection markers in the plurality of projection images includes:

determining, based on a first projection image of the plurality of projection images, a first projection marker;

determining a first correspondence between the original marker in the geometric model and the first projection marker in the first projection image;

determining, based on a second projection image of the plurality of projection images, a second projection marker; and

determining, based on a position relationship between the second projection marker and the first projection marker, a mapping relationship between the second projection marker and the first projection marker; and

determining the target correspondence between the original markers in the geometric model and the projection markers in the tracked projection image, based on the first correspondence and the mapping relationship between the second projection marker and the first projection marker.

9 . The system of claim 8 , wherein the second projection image is generated in the adjacent scans of the first projection image.

10 . The system of claim 8 , wherein the position relationship includes a distance between coordinates of the second projection marker and coordinates of the first projection marker.

11 . The system of claim 10 , wherein determining, based on the position relationship between the second projection marker and the first projection marker, the mapping relation between the second projection marker and the first projection marker includes:

in response to determining that an absolute value of a coordinate difference between a second projection marker in the second projection image and the first projection marker is the smallest, determining the second projection marker and the first projection marker as the same projection corresponding to the original marker.

12 . The system of claim 1 , wherein determining a target correspondence between the original markers in the geometric model and the projection markers in the tracked projection image by tracking, based on the initial correspondence and a correlation of the plurality of projection images, the projection markers in the plurality of projection images includes:

tracking the projection markers in the plurality of projection images using a recognition model, the recognition model being a machine learning model.

13 . The system of claim 12 , wherein:

an input of the recognition model includes the projection images generated in adjacent scans; and

an output of the recognition model includes mapping relationships of the projection markers in the projection images generated in the adjacent scans.

14 . A method for geometric correction based on image tracking, method for geometric correction based on image tracking, implemented on a device including one or more processing devices and one or more storage devices, the method comprising:

obtaining a plurality of projection images of a geometric model, each of the plurality of projection images including a projection marker corresponding to an original marker in the geometric model, wherein the plurality of projection images corresponding to a plurality of imaging angles;

obtaining an initial correspondence between the original markers in the geometric model and the projection markers in one of the plurality of projection images;

determining a target correspondence between the original markers in the geometric model and the projection markers in the tracked projection image by tracking, based on the initial correspondence and a correlation of the plurality of projection images, the projection markers in the plurality of projection images; and

performing, based on the target correspondence, a geometric correction.

15 . The method of claim 14 , wherein obtaining the plurality of projection images of the geometric model includes:

obtaining a imaging parameter of the geometric model; and

generating the plurality of projection images of the geometric model by performing, based on the imaging parameter.

16 . The method of claim 15 , further comprising:

determining, based on the imaging parameter, an angle difference of at least one group of adjacent scans, and

determining, based on the angle difference, at least one of the plurality of imaging angles.

17 . The method of claim 14 , wherein obtaining an initial correspondence between the original markers in the geometric model and the projection markers in one of the plurality of projection images includes:

determining the projection image satisfying a first preset condition as a reference projection image; and

obtaining the initial correspondence between the original markers in the geometric model and the projection markers in the reference projection image.

18 . The method of claim 17 , wherein the at least one processor is directed to perform the operations including:

in response to determining that none of the plurality of projection images satisfy the first preset condition, extracting an similar projection image satisfying a second preset condition;

determining, based on the similar projection image, an optimized angle

generating an optimized projection image by scanning, based on the optimized angle, the geometric model; and

determining the optimized projection image satisfying the first preset condition as the reference projection image.

19 . The method of claim 14 , wherein determining the optimized projection image satisfying the first preset condition as the reference projection image by tracking, based on determining the optimized projection image satisfying the first preset condition as the reference projection image includes:

determining, based on a first projection image of the plurality of projection images, a first projection marker;

determining a first correspondence between the original marker in the geometric model and the first projection marker in the first projection image;

determining, based on a second projection image of the plurality of projection images, a second projection marker; and

determining, based on a position relationship between the second projection marker and the first projection marker, a mapping relationship between the second projection marker and the first projection marker; and

determining the target correspondence between the original markers in the geometric model and the projection markers in the tracked projection image, based on the first correspondence and the mapping relationship between the second projection marker and the first projection marker.

20 . A non-transitory computer readable medium, comprising at least one set of instructions for geometric correction based on image tracking, wherein when executed by one or more processors of a computing device, the at least one set of instructions causes the computing device to perform a method, the method comprising:

obtaining a plurality of projection images of a geometric model, each of the plurality of projection images including a projection marker corresponding to an original marker in the geometric model, wherein the plurality of projection images corresponding to a plurality of imaging angles;

obtaining an initial correspondence between the original markers in the geometric model and the projection markers in one of the plurality of projection images;

determining a target correspondence between the original markers in the geometric model and the projection markers in the tracked projection image by tracking, based on a correlation of the plurality of projection images, the projection markers in the plurality of projection images; and

performing, based on the target correspondence, a geometric correction.