Target body surface reconstruction method in three-dimensional image

Abstract

The invention discloses a target body surface reconstruction method in a three-dimensional image. The method includes the steps of firstly, obtaining outer shell data of a target body through a dot density analytical method and an improved grid division method, namely, a depth-first traversal method; secondly, conducting surface reconstruction of the target body through an energy function method based on image segmentation by taking the outer shell data, obtained in the first step, of the target body as point cloud in the computer graphics. According to the target body surface reconstruction method in the three-dimensional image, firstly, the obtained volume data, namely, the earthquake sampling point data, of the target body are processed through the improved image segmentation technology so that the volume data can be applied to the point cloud surface reconstruction method in the computer graphics, then, surface reconstruction is conducted on the target body through the improved energy function method based on image segmentation, the difficult problem of conducting surface reconstruction on the three-dimensional target body which takes sampling points as the volume data is solved, and the target body surface reconstruction method is suitable for surface reconstruction of the target body which is high in noise content or flat in appearance, inconsistent in sampling in all directions or undersampled.

Description

technical field [0001] The invention belongs to the technical field of three-dimensional images, and in particular relates to a surface reconstruction method of a target body in a three-dimensional image of seismic data. Background technique [0002] The surface reconstruction of the target body in the 3D image is widely used in scientific research and production fields such as geophysics, geological exploration, medical and health care, and remote sensing and telemetry. For example, in geological interpretation work, the identification, tracking and modeling of subsurface objects and stratigraphic structures are very important; for example, in medical and health work, 3D modeling and surface reconstruction of internal tissues and organs of the human body are very important. Construction is also a very important process. [0003] For the surface reconstruction of the 3D target body whose sampling points are the surface points of the target body, the point cloud surface reco...

AI Technical Summary

Problems solved by technology

However, since the 3D image data is volume data, the surface reconstruction method based on the point cloud cannot be directly used for the surface reconstruction of the target body in the 3D image.

[0004] For the surface reconstruction of the 3D target body whose sampling points are volume data, the current main solutions can be divided into two categories. One is based on the image segmentation technology in image processing, which is used to segment the 3D image and separate the target body; One is to use volume rendering, volume perspective, point and other scalar forms to directly visualize three-dimensional volume data, such as the commercial software petrel, which expresses the target volume in the form of volume perspective or point rendering, but these methods are not very good Reconstruct the surface of the target body, especially the seismic target body in the geological field. This is because the surface reconstruction of the 3D image target body is very difficult, mainly manifested in:

[0005] (1) The 3D image data is the volume sampling point in the scene where the target body is located, and the original data reconstructed based on the point cloud surface is the sampling point on the surface of the target body, which cannot directly use the point cloud surface reconstruction method in computer graphics

[0006] (2) Especially for seismic data in the geological field, the shape of the target body in the seismic image is irregular and unpredictable, and it is impossible to make any geometric constraints on the contour shape of the target body in advance

[0007] (3) The target body in the seismic image has the characteristics of ambiguous boundaries and low signal-to-noise ratio of seismic data, which makes the surface reconstruction work more difficult

Images