CN-122023296-A - CT image-based jade relic digital segmentation and porosity nondestructive detection method

CN122023296ACN 122023296 ACN122023296 ACN 122023296ACN-122023296-A

Abstract

The invention provides a CT image-based jade cultural relic digital segmentation and porosity nondestructive detection method, and relates to the technical field of cultural relic nondestructive detection and digital display. Aiming at the problems that the surface of a jade cultural relic is often covered with impurities, the internal pore structure is complex and difficult to accurately divide and observe, the invention provides a double-flow processing architecture, which is characterized in that in a calculation channel, accurate thresholds are obtained by utilizing downsampling data in combination with interactive gray profile analysis, an entity envelope mask is constructed by three-dimensional morphological closed operation, the internal porosity is calculated in a constraint mode, in a reconstruction channel, independent three-dimensional models of a jade cultural relic body and a surface attachment are respectively constructed by adopting a double-threshold isosurface extraction technology, a key interaction control mechanism is established, and real-time state switching of integral coexistence and main body display and attachment display is realized.

Inventors

SONG WEITAO
YAN SIYI
LIAO LINGYU
ZHANG HAOTIAN
MA SHINING

Assignees

北京理工大学

Dates

Publication Date: 20260512
Application Date: 20260120

Claims (6)

1. A CT image-based jade cultural relic digital segmentation and porosity nondestructive testing method is characterized by comprising the following steps: step S1, data preprocessing and double-flow architecture construction, namely reading original CT image sequence data of jade cultural relics, converting the original CT image sequence data into a three-dimensional voxel matrix, and then constructing a double-flow channel (1) based on preset high-density artifacts Performing artifact removal processing, performing downsampling processing on the artifact-removed data by adopting a linear interpolation algorithm to generate calculation stream data for real-time interactive analysis and volume calculation; S2, interactive gray feature calibration, namely drawing a detection path on a slice image of a calculation flow in a man-machine interaction mode, generating a gray profile curve in real time, and establishing a four-dimensional threshold value parameter set, namely a body gray interval And pore gray scale interval Quantitatively analyzing the gray level boundary of the jade body, the internal pores and the surface attachments by using the four-dimensional threshold, wherein the gray level interval of the body distinguishes the body from the attachments, the gray level interval of the pores distinguishes the body from the pores, Respectively representing a lower limit value and an upper limit value of the main body gray scale interval; respectively representing a lower limit value and an upper limit value of a pore gray scale interval; S3, performing morphological envelope-based volume quantization, namely extracting a preliminary mask of a jade entity by using the body gray scale interval, performing three-dimensional binary morphological closing operation on the preliminary mask to construct a solid envelope body containing internal pores, and respectively calculating the volume of the jade entity and the volume of the pores by using the envelope body as a space constraint mask through Boolean operation so as to obtain the porosity; and S4, based on rendering stream data, respectively executing independent isosurface extraction and grid optimization processing aiming at the surface attachment and the jade body, constructing two independent three-dimensional geometric objects, and establishing a logic switch in a rendering pipeline to realize independent implicit and explicit control and fusion rendering of the two.
2. The method for digitized segmentation and non-destructive detection of porosity of a jade relic based on CT images as set forth in claim 1, wherein said step S2 comprises: 1) Initializing a visual environment, namely automatically positioning and calculating a central layer slice of stream data by a system, and generating a pseudo color mapping image and a gray histogram of the slice, wherein the histogram counts gray distribution from zero gray to a maximum gray value; 2) Establishing a continuous interaction monitoring cycle, namely enabling the system to enter a timely response mode and continuously monitoring mouse click events of a user; 3) Capturing a group of initial coordinates and termination coordinates clicked by a user on the pseudo-color mapping graph in each iteration of the loop, extracting voxel gray level sequences on the path in real time, refreshing and displaying corresponding gray level profile graphs, and enabling the user to repeatedly detect gray level features of different areas by changing the path positions; 4) The loop termination and parameter establishment, namely, responding to a loop termination instruction of a user, exiting the monitoring loop, and receiving four-dimensional threshold parameters, namely, an ontology gray scale interval, determined by the user based on the repeated detection result And pore gray scale interval 。
3. The method for digitized segmentation and nondestructive detection of porosity of jade relics based on CT images as set forth in claim 1, wherein the step S3 comprises the following steps: 1) Preliminary extraction of entities using main gray scale intervals Performing binarization threshold segmentation on the calculated stream data to obtain a discrete binary image; 2) Removing noise, namely performing connected domain analysis on the binary image, calculating the number of voxels of all connected domains, and only reserving the connected domain with the largest number of voxels as an initial entity mask Removing free noise points; 3) Envelope construction, setting the radius of structural element Masking an initial entity Executing morphological closing operation to fill closed cavity in entity to generate solid envelope mask ; 4) Constraint computing by incorporating computation flow data into the data Binary results in range and solid envelope mask Performing AND logic operation to count total number of body voxels in result Constructing and obtaining an ontology computing mask, and processing computing stream data in a computing system Binary results in range and solid envelope mask Performing AND logic operation to count total number of pore voxels in the result . Calculating volume according to voxel physical distance and according to formula Porosity is obtained.
4. The method for digitized segmentation and non-destructive detection of porosity of a jade relic based on CT images as set forth in claim 1, wherein the step S4 comprises: 1) Additive rendering based on preset high artifact threshold Performing iso-surface extraction on full resolution CT data to be higher than a preset threshold A three-dimensional model of the attachment is constructed in the area of the model, a yellow mark is given, and the model independently characterizes the spatial distribution structure of the attached impurities on the surface of the cultural relic; 2) Jade body rendering, namely performing high-value shielding operation on full-resolution CT data, and enabling gray values to be larger than gray values Is set to zero, preventing high density attachments from being repeatedly calculated. Ontology threshold value based on user acquisition Performing three-dimensional space connectivity analysis on the preliminary grid model, calculating the geometric scale of all independent connected areas, only reserving the connected area with the largest gauge as a jade body three-dimensional model, endowing white marks, and automatically removing other non-connected suspended impurities and noise points; 3) And the multi-material fusion display is carried out by placing the attachment three-dimensional model and the jade body three-dimensional model in the same three-dimensional world coordinate system and carrying out superposition fusion rendering, so that the cultural relic body form and the attachment distribution state are simultaneously presented in a single view, and the digital segmentation and coexistence display of the attachment three-dimensional model and the jade body three-dimensional model are realized.
5. The method for digitized segmentation and non-destructive detection of porosity of a jade relic based on CT images of claim 1, further comprising a joystick camera mode, wherein a rotating response mechanism based on the duration of mouse pressing is established, and when a user presses a left mouse button, the three-dimensional model continuously and automatically rotates according to the vector direction and distance of the mouse away from the click point without repeated dragging.
6. The method for digitally segmenting and nondestructively detecting the jade cultural relics based on the CT images is characterized by further comprising a multi-state key switching mechanism, wherein a keyboard event monitor is arranged, at least three display states are defined, namely a first state is that a jade body model and a surface attachment model are rendered simultaneously, the spatial attachment relation of the jade body model and the surface attachment model is displayed, a second state is that only the jade body model is rendered, the attachment model is hidden, the surface details of the cultural relics after the attachments are removed are displayed, and a third state is that only the surface attachment model is rendered, the jade body model is hidden, and the distribution forms of the attachments are independently displayed.

Description

CT image-based jade relic digital segmentation and porosity nondestructive detection method Technical Field The invention belongs to the technical field of digital protection and nondestructive detection of cultural relics, and particularly relates to a CT image-based method for digital segmentation and nondestructive detection of a jade cultural relic. Background The jade cultural relics are used as important carriers of China cultural, and the internal texture, the structural texture and the degree of being absorbed (reflected as the change of porosity) are important bases for researching the manufacturing process, the source flow of the production area and the history age. The Computer Tomography (CT) technology is used as a high-precision nondestructive testing means, can acquire three-dimensional density distribution information inside the cultural relics in a non-contact mode, and becomes a standard tool for digital construction and research of the cultural relics. However, in practical applications, the surface of the unearthed jade relic is often covered with hard-bonded soil, rust or other high-density secondary attachments, which often appear as structural irregularities in the CT image. In the prior art, a single global threshold is generally adopted for segmentation, so that surface attachments are difficult to thoroughly remove while the details of a jade body are reserved, the surface roughness of a three-dimensional reconstruction model is caused, impurities remain, and the digital display effect of cultural relics is seriously affected. Furthermore, quantitative calculation of porosity is a difficulty in jade research. Because of the complex morphology of the jade relics, the traditional volume calculation method is difficult to automatically distinguish 'closed pores in the relics' from 'background air outside the relics'. If the precise morphological treatment is not performed, the outside air is often misjudged as pores, or the internal communication pores are not counted, so that the deviation of the calculation result of the porosity is larger. Meanwhile, the existing visualization software lacks an independent control mechanism for cultural relic components, researchers cannot switch and view the pure body after removing attachments and the spatial distribution of the attachments in the same three-dimensional scene in real time, and the erosion relation of the attachments to the cultural relics is difficult to visually evaluate. Disclosure of Invention In view of the above, the invention aims to provide a CT image-based jade cultural relic digital segmentation and porosity nondestructive detection method, which aims to solve the problems of uncleanness of a three-dimensional reconstruction model, non-visual jade model structure, inaccurate calculation of internal porosity and single visual interaction means caused by the interference of surface attachments of the cultural relics in the prior art. A CT image-based jade relic digital segmentation and porosity nondestructive detection method comprises the following steps: step S1, data preprocessing and double-flow architecture construction, namely reading original CT image sequence data of jade cultural relics, converting the original CT image sequence data into a three-dimensional voxel matrix, and then constructing a double-flow channel (1) based on preset high-density artifacts Performing artifact removal processing, performing downsampling processing on the artifact-removed data by adopting a linear interpolation algorithm to generate calculation stream data for real-time interactive analysis and volume calculation; S2, interactive gray feature calibration, namely drawing a detection path on a slice image of a calculation flow in a man-machine interaction mode, generating a gray profile curve in real time, and establishing a four-dimensional threshold value parameter set, namely a body gray interval And pore gray scale intervalQuantitatively analyzing the gray level boundary of the jade body, the internal pores and the surface attachments by using the four-dimensional threshold, wherein the gray level interval of the body distinguishes the body from the attachments, the gray level interval of the pores distinguishes the body from the pores,Respectively representing a lower limit value and an upper limit value of the main body gray scale interval; respectively representing a lower limit value and an upper limit value of a pore gray scale interval; S3, performing morphological envelope-based volume quantization, namely extracting a preliminary mask of a jade entity by using the body gray scale interval, performing three-dimensional binary morphological closing operation on the preliminary mask to construct a solid envelope body containing internal pores, and respectively calculating the volume of the jade entity and the volume of the pores by using the envelope body as a space constraint mask through Boolean operation so as to obtain t