Three-dimensional CT image sectioning method and system based on material identification

Abstract

The invention discloses a three-dimensional CT image sectioning method and system based on material identification. The method comprises the following steps: analyzing a to-be-detected luggage according to a material identification algorithm, and generating a sectioning boundary, an article boundary and a sensitivity coefficient; drawing a three-dimensional CT image through an image space volume drawing technology; setting a dangerous goods quantity threshold value, and judging whether the dangerous goods quantity is smaller than the threshold value; and according to a three-point positioning method or a three-dimensional point set convex hull solving algorithm, carrying out common sectioning on the to-be-detected luggage to obtain sectioning data, and taking the sectioning data as a new data source of a ray casting algorithm to carry out redrawing until all articles in the to-be-detected luggage are completely exposed. According to the material identification result, the sectioning boundary is automatically generated, so that the time cost is saved; through a three-point positioning method, a geometric center method and a rapid convex hull method, the sectioning boundary is not limited to a straight line any more, but is a space cutting surface or a three-dimensional convex hull surface based on sensitive points, and usability is improved.

Description

Technical Field

[0001] The present invention relates to a three-dimensional CT image sectioning method based on material identification, and also relates to a corresponding three-dimensional CT image sectioning system, belonging to the technical field of image data processing. Background Art

[0002] In many fields such as public safety and medical diagnosis, three-dimensional CT imaging technology has become a key way to detect the internal structure of objects. This technology uses tomography and three-dimensional reconstruction to produce high-resolution volume data, so that the internal details of the object can be intuitively presented to the user. However, due to the complex characteristics of three-dimensional data, the target object is often easily obscured by other objects. Based on this, in order to expose the area of ​​interest, a sectioning operation is usually used. At present, traditional sectioning methods mainly cover two categories, namely manual sectioning and semi-automatic sectioning. Manual sectioning performs poorly in terms of efficiency due to its reliance on manual operation; semi-automatic sectioning mostly relies on the edge or density difference of the object to generate the sectioning plane by itself, but it is subject to the preset threshold and its adaptability needs to be improved.

[0003] A Chinese invention patent with authorization announcement number CN112598682B discloses a method for sectioning a 3D CT image around any angle. In this technical solution, the user uses the mouse to determine the starting and ending points of the sectioning line. The system then draws a straight line on the screen based on the position of the user's mouse click and release. The system then determines the data to the left of the sectioning line as discarded data, and the data to the right as retained data. Finally, the volume rendering is re-performed based on the discarded data to achieve the sectioning effect. From the above content, it is not difficult to see that if this technical solution wants to fully expose dangerous goods, it can only be achieved through user input, lacks automated auxiliary means, and the sectioning line is limited to a straight line, that is, it is only a regular plane section. Summary of the Invention

[0004] The primary technical problem to be solved by the present invention is to provide a three-dimensional CT image sectioning method based on material identification.

[0005] Another technical problem to be solved by the present invention is to provide a three-dimensional CT image sectioning system based on material identification.

[0006] In order to achieve the above technical objectives, the present invention adopts the following technical solutions:

[0007] According to a first aspect of an embodiment of the present invention, a method for sectioning a three-dimensional CT image based on material identification is provided, comprising the following steps:

[0008] S1: The CT device performs a tomographic scan on the baggage to be inspected, obtains and backs up the tomographic data, and analyzes all items in the baggage to be inspected based on the material recognition algorithm to generate the section boundary, the boundaries of all items in the baggage to be inspected, and the sensitivity coefficient of each item;

[0009] S2: Rendering a 3D CT image using image space volume rendering technology based on the slice data, section boundaries, and object boundaries;

[0010] S3: Set a threshold for the number of dangerous goods, and determine whether the number of dangerous goods is less than the threshold based on the results of the 3D CT image; if the number of dangerous goods is less than the threshold, proceed to step S4; if the number of dangerous goods is greater than or equal to the threshold, proceed to step S5;

[0011] S4: Perform a normal section on the baggage to be inspected according to the three-point positioning method to obtain normal section data, and then proceed to step S6;

[0012] S5: Perform high-level sectioning on the bag to be inspected based on the convex hull solution algorithm of the three-dimensional point set to obtain high-level sectioning data, and then proceed to step S6;

[0013] S6: redrawing the section data as a new data source of the ray casting algorithm, obtaining and backing up the redrawn 3D CT image;

[0014] S7: Repeat steps S3 to S6 until all items in the baggage to be inspected are completely exposed;

[0015] S8: Inspect all exposed items and report the inspection results.

[0016] Preferably, the image space volume rendering technology in step S2 is a ray casting algorithm.

[0017] Preferably, the calculation formula of the sensitivity coefficient in step S1 is:

[0018] Sensitivity coefficient = importance parameter / occurrence frequency parameter.

[0019] Preferably, the importance parameters of metals, liquids, flammable, explosive, and radioactive substances are greater than the importance parameters of non-prohibited items; the occurrence frequency parameter is determined according to the occurrence frequency, and the higher the occurrence frequency, the greater the occurrence frequency parameter b.

[0020] Preferably, the three points used in the three-point positioning method in step S4 are the top three points in terms of sensitivity coefficient.

[0021] Preferably, if the sensitivity coefficient ranking fails in step S4, the luggage to be inspected is sectioned using the geometric center method in step S4.

[0022] Preferably, the three-dimensional point set convex hull solution algorithm in step S5 is a fast convex hull algorithm.

[0023] Preferably, based on the fast convex hull algorithm, three points with the largest sensitivity coefficients are first selected, and then points with a similarity difference greater than a preset threshold are searched in the neighborhood of the three points with the largest sensitivity coefficients, and the points with a similarity difference greater than the preset threshold are added to the convex hull solution point set.

[0024] Preferably, after the sectioning process of the 3D CT image is completed, if the original 3D CT image needs to be restored, the ray casting algorithm is used based on the pre-backed up slice data of the test line package or the redrawn 3D CT image to choose to gradually restore or completely restore the 3D CT image.

[0025] According to a second aspect of an embodiment of the present invention, a three-dimensional CT image sectioning system based on material identification is provided, comprising a processor and a memory; wherein the memory is coupled to the processor and is used to store a computer program, and when the computer program is executed by the processor, the processor implements the above method.

[0026] Compared with existing technologies, this method automatically generates cutting boundaries based on material recognition results, eliminating the need for manual line drawing in response to mouse input, saving time and cost. By using three-point positioning, geometric center method, and rapid convex hull method, the cutting boundary is no longer limited to a straight line. Instead, it can be a spatial cutting surface or a three-dimensional convex hull surface based on sensitive points, improving ease of use. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a flow chart of a 3D CT image sectioning method based on material identification in a first embodiment of the present invention;

[0028] Figure 2 is a schematic diagram of a three-dimensional CT image presented after tomography in the first embodiment of the present invention;

[0029] Figure 3 This is a schematic diagram of a method for selecting the top three spatial positions with the highest sensitivity coefficients to determine the cutting plane in the first embodiment of the present invention;

[0030] Figure 4 This is a rendering of the method for selecting the top three spatial positions with the highest sensitivity coefficients to determine the cutting surface in the first embodiment of the present invention;

[0031] Figure 5 A schematic diagram of segmenting slice data using a convex hull cutting plane in the first embodiment of the present invention;

[0032] Figure 6 This is an effect diagram of the first embodiment of the present invention, in which the convex hull cutting surface is used to segment the slice data and perform partial rendering;

[0033] Figure 7 Schematic diagram of a method for generating cutting surfaces parallel to respective coordinate planes using a geometric center as a positioning point in a first embodiment of the present invention;

[0034] Figure 8 This is an effect diagram of a method for generating cutting planes parallel to each coordinate plane using the geometric center as a positioning point in the first embodiment of the present invention;

[0035] Figure 9 FIG. 4 is a structural diagram of a three-dimensional CT image sectioning system based on material identification in a second embodiment of the present invention. DETAILED DESCRIPTION

[0036] The technical content of the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0037] First embodiment

[0038] like Figure 1 As shown, a first embodiment of the present invention provides a 3D CT image sectioning method based on material identification, which includes at least the following steps:

[0039] S1: The CT device performs a tomographic scan on the baggage to be inspected, obtains and backs up the tomographic data, and analyzes all items in the baggage to be inspected based on the material recognition algorithm to generate the section boundary, the boundaries of all items in the baggage to be inspected, and the sensitivity coefficient of each item.

[0040] S2: If Figure 2 As shown, a three-dimensional CT image is rendered using image space volume rendering technology based on the slice data, the section boundary, and the boundary of the object.

[0041] S3: Set a threshold for the number of dangerous goods, and determine whether the number of dangerous goods is less than the threshold based on the results of the three-dimensional CT image; if the number of dangerous goods is less than the threshold, proceed to step S4; if the number of dangerous goods is greater than or equal to the threshold, proceed to step S5.

[0042] S4: As Figure 3 、 Figure 4 As shown, according to the three-point positioning method, the luggage to be inspected is subjected to ordinary sectioning to obtain ordinary sectioning data, and the process proceeds to step S6.

[0043] S5: If Figure 5 、 Figure 6 As shown, according to the three-dimensional point set convex hull solution algorithm, the line hull to be inspected is subjected to high-level sectioning to obtain high-level sectioning data, and the process proceeds to step S6.

[0044] S6: Redraw the section data as a new data source of the ray casting algorithm to obtain and back up the redrawn three-dimensional CT image.

[0045] S7: Repeat steps S3 to S6 until all items in the baggage to be inspected are completely exposed.

[0046] S8: Inspect all exposed items and report the inspection results.

[0047] In one embodiment of the present invention, the image space volume rendering technique in step S2 is a ray casting algorithm.

[0048] In one embodiment of the present invention, the calculation formula of the sensitivity coefficient in step S1 is:

[0049] Sensitivity coefficient = importance parameter / frequency parameter

[0050] Among them, the importance parameter is set according to the national security inspection standards, that is, the importance parameters of metals, liquids, flammable, explosive, and radioactive materials are greater than the importance parameters of non-prohibited items; the occurrence frequency parameter is determined according to the occurrence frequency, and the higher the occurrence frequency, the greater the occurrence frequency parameter.

[0051] The three points used in the three-point positioning method in step S4 are the top three points in terms of sensitivity coefficient.

[0052] Preferably, Figure 7 、 Figure 8 As shown, if the sensitivity coefficient sorting fails, the geometric center method is used to section the baggage to be inspected in step S4.

[0053] In one embodiment of the present invention, the convex hull solution algorithm for the three-dimensional point set in step S5 is a fast convex hull algorithm.

[0054] Preferably, after the 3D CT image sectioning process is complete, if the original 3D CT image needs to be restored, the ray casting algorithm within image-space volume rendering technology is used based on pre-backed slice data of the packet to be tested or a redrawn 3D CT image to gradually or completely restore the 3D CT image. This allows security personnel to compare the images before and after the sectioning process, further reducing missed inspections.

[0055] The following is a detailed description of the above steps:

[0056] First, in step S1, the CT device performs a tomographic scan on the baggage to be inspected. This scan involves emitting X-rays into the baggage and collecting attenuated data from multiple angles after the rays have penetrated the baggage, thereby generating tomographic data. This tomographic data contains detailed structural information about the items within the baggage. This data is then analyzed using a material recognition algorithm, which accurately distinguishes the various items within the baggage based on the differences in X-ray absorption characteristics of different materials. Once the analysis is complete, a sensitivity coefficient ranking is generated for all items within the baggage to be inspected. This step aims to accurately obtain detailed information about the items within the baggage. The sensitivity coefficient ranking can help subsequent steps quickly locate potentially hazardous areas, providing a key reference for subsequent inspections and improving inspection efficiency and accuracy.

[0057] The present invention uses a material recognition algorithm to enable the system to automatically generate a cutting boundary according to the result of material recognition, without the need to respond to mouse input and manually draw lines, thus saving time and cost.

[0058] Next, in step S2, a raycasting algorithm converts the 2D slice data acquired by the tomography scan into a 3D image with a three-dimensional feel. This process requires real-time refreshes at 30, 60, or even higher frequencies per second to respond to user screen operations such as rotation, dragging, and zooming. During each refresh, the system processes the volume data through coordinate transformation, raycasting, and other processing steps, ultimately mapping it to screen pixels to generate a dynamic 3D image. Compared to traditional 2D slice data, 3D imaging technology can intuitively visualize the shape, position, and spatial relationships of items within a bag, providing security personnel with more comprehensive visualization information, significantly improving the accuracy of contraband identification and inspection efficiency.

[0059] The ray casting algorithm is one of the most commonly used methods for displaying CT data. Starting from a screen pixel, it emits a virtual ray along the observation direction. As the ray propagates, data is sampled at each point it passes through. The sampled information includes information such as the density and absorption coefficient of the material. After sampling, the information from all sampled points is comprehensively considered to accurately calculate their contribution to the screen pixel. This calculation takes into account many factors, such as the distance between the sampled points and the screen pixel and the sampled points' properties. Through calculation, the contributions of these sampled points are integrated to ultimately synthesize the color of the screen pixel. The ray casting algorithm shares many similar principles with other similar imaging methods. Their general approach is to use two-dimensional pixels to represent three-dimensional information. In this process, occlusion between sampled points is an inevitable problem. Because data acquisition and processing occur in three-dimensional space, sampled points are positioned in a forward-backward relationship. Sampled points in the rear may be obscured by those in the front. This occlusion can interfere with the final image. For example, in security inspection CT images, when observing an item inside a baggage to be inspected, the item in front may block the item behind, making it difficult for security personnel to accurately obtain the complete shape and information of the item when interpreting the image.

[0060] Therefore, in order to improve the accuracy of interpretation, it is crucial to find effective interference removal methods. By adopting appropriate algorithms and technical means, the occlusion relationship between sampling points can be accurately analyzed and processed. For example, using a depth sorting algorithm, the sampling points are sorted according to the distance from the observation point, and sampling points with a short distance are prioritized to reduce the impact of occlusion. Alternatively, an adaptive sampling strategy can be adopted to increase the density of sampling points in areas prone to occlusion, obtain richer information, and thus synthesize images more accurately. These effective interference removal methods can significantly improve the quality of images presented by the ray casting algorithm, allowing security personnel, doctors, scientific researchers and other users to understand the internal structure of objects more clearly and accurately when observing CT data, providing a more reliable basis for security inspections, medical diagnosis, industrial testing and other fields.

[0061] Next, in step S3, a threshold value for the number of dangerous goods is set, and based on the results of the three-dimensional imaging, it is determined whether the number of dangerous goods is less than the threshold value. In actual security inspection scenarios, a reasonable threshold value for the number of dangerous goods is set according to different security inspection standards and requirements. Security personnel identify the dangerous goods in the luggage based on the items in the three-dimensional imaging and count their number. If the number of dangerous goods is less than the threshold value, ordinary sectioning is performed; if the number of dangerous goods is greater than or equal to the threshold value, advanced sectioning is performed. The purpose of this step is to preliminarily judge the complexity of the security inspection based on the number of dangerous goods, so as to adopt different sectioning strategies. For cases with a small number of dangerous goods, a relatively simple ordinary sectioning method can be used; for cases with a large number of dangerous goods, a more advanced sectioning method is required to ensure a comprehensive and accurate inspection of the items in the luggage.

[0062] If the ordinary sectioning method is selected (i.e., step S4), the baggage to be inspected is subjected to ordinary sectioning according to the three-point positioning method to obtain ordinary sectioning data. The three-point positioning method is a method of determining the sectioning plane by determining the positions of three points. Three appropriate points on the baggage are selected to determine a plane, and then sectioning is performed along this plane. Through ordinary sectioning, the internal structure data of the baggage on a specific plane can be obtained, and the distribution of items on the sectioning plane can be displayed. These ordinary sectioning data provide the basis for subsequent redrawing and further analysis, helping security personnel to observe the items in the baggage from different angles and to more carefully inspect the details and potential dangers of the items.

[0063] Among them, the three-point positioning method is a key method used to determine the position of the cutting plane when cutting through the baggage to be inspected, and plays an important role in security inspections. When selecting the three points on the baggage, certain principles must be followed. These three points should be representative and unique, and should not be too concentrated in a single local area. Generally, points are selected on different sides and edges of the baggage so that the plane determined in this way can better reflect the overall structural characteristics of the baggage interior. On a rectangular baggage, one point can be selected at the top left corner, another point at the middle of the bottom right corner of the front, and the third point at the middle of the side near the bottom. This selection method allows the cutting plane to cover as many items as possible at different locations within the baggage, avoiding missing important information. In the present invention, the three points are selected based on the sensitivity coefficient. This selection is because the sensitivity coefficient represents the degree of suspicion (or danger) of the item being tested. The larger the coefficient, the more suspicious it is. Therefore, by sorting to determine the three most suspicious points, and then cutting, it can more accurately and efficiently perform security inspections on the baggage.

[0064] Furthermore, if the sensitivity coefficient ranking fails, meaning the sensitivity coefficients are equal (mean), the geometric center method can be used instead of the three-point positioning method. The geometric center method was chosen as an alternative because it is more efficient and less costly than other algorithms. In the security inspection field, inspection time (or speed) is a crucial assessment criterion for security inspectors. Typically, a baggage inspection must be completed within 1 to 3 seconds. Therefore, the speed of the algorithm becomes a key factor in its selection.

[0065] If the advanced sectioning method is selected (i.e., step S5), the baggage to be inspected is subjected to advanced sectioning according to the three-dimensional point set convex hull solving algorithm to obtain advanced sectioning data. The three-dimensional point set convex hull solving algorithm will calculate a minimum convex polyhedron containing all points, namely the convex hull, based on the three-dimensional point set information of the items in the baggage. Then, sectioning is performed along certain faces or specific directions of the convex hull. This sectioning method can more comprehensively and specifically display the complex structure and mutually obstructed items in the baggage. Compared with ordinary sectioning data, advanced sectioning data can provide more detailed and comprehensive item information, and is especially suitable for baggage inspections with a large number of dangerous goods and complex item structures.

[0066] Among them, the common algorithms for solving the convex hull of three-dimensional point sets include the incremental method, the divide-and-conquer method, the gift wrapping method, and the fast convex hull algorithm. Due to the time requirements in the security inspection process mentioned above, the present invention selects the fast convex hull algorithm among the convex hull solving algorithms for three-dimensional point sets. Its average time complexity is good, and the convex hull calculation can usually be completed in near-linear time. It has good randomness and adaptability, and has good performance for point sets of various distributions, which is very consistent with the needs of the security inspection field. In the present invention, based on the fast convex hull algorithm, with the principle of minimizing the accelerated calculation of convex hull points, first select the three points with the largest sensitivity coefficients, and then search for points with a similarity difference greater than a preset threshold in the neighborhood of the three points with the largest sensitivity coefficients, and add the points with a similarity difference greater than the preset threshold in the neighborhood to the convex hull solution point set.

[0067] It should be noted that the above neighborhood search method can be iterated once or multiple times according to actual scenario requirements, and the present invention is not limited to this.

[0068] The present invention uses a three-point positioning method, a geometric center method and a fast convex hull method so that the cutting boundary is no longer limited to a straight line, but a spatial cutting surface or a three-dimensional convex hull surface based on sensitive points, thereby improving ease of use.

[0069] Next, in step S6, after the sectioning is complete, the redrawn sectioning algorithm uses the sectioning data to reconstruct the image using a ray casting algorithm. This step presents the sectioned data as a 3D image, allowing security personnel to more intuitively observe the 3D form of the object on the sectioning plane. Compared to simple 2D sectioning data, the redrawn 3D image provides richer information, making it easier for security personnel to identify the object's shape, size, and relationship to surrounding objects, helping to more accurately determine whether the object is dangerous.

[0070] In the technical solution of this invention, the redrawing sectioning algorithm is a key component in achieving efficient cross-sectional viewing. The presentation of 3D CT images relies on a ray casting algorithm, which requires continuous refresh operations when displaying images on the screen to meet the user's real-time interactive needs (such as rotation and zooming). The redrawing sectioning algorithm cleverly utilizes the imaging calculation mechanism during this refresh process.

[0071] During conventional 3D CT image display, the coordinates of volume data points undergo a complex series of transformations to be correctly displayed on the screen. These transformations include model view transformation, projection transformation, and viewport transformation. Before the sectioning operation, the volume data points undergo these transformations according to the normal process to form the original 3D image. While the sectioning operation alters the entire data processing process, the redrawing sectioning algorithm is optimized based on the existing coordinate transformation mechanism.

[0072] The reason for utilizing refresh-time imaging calculations is that refreshes are a continuous and frequent operation in 3D imaging systems. Each refresh requires the recalculation and display of volume data points, which provides a natural opportunity for the sectioning algorithm to perform computations. Creating a separate computational process for sectioning would not only increase system complexity but also waste computing resources and reduce efficiency. Utilizing the existing refresh mechanism maximizes the reuse of existing computing resources and processes, achieving more efficient sectioning calculations.

[0073] In this way, the present invention integrates the slice calculation into conventional imaging calculations, laying a solid foundation for achieving rapid cross-sectional visualization. This design not only simplifies the algorithm flow but also improves the overall performance of the system, enabling rapid response to user slice operations even when processing large-scale volume data.

[0074] At the next refresh after a slice, the redrawing slice algorithm begins to demonstrate its efficient computational capabilities. At this point, the volume data points undergo a conventional forward coordinate transformation. Notably, the coordinate transformations for each data point are independent of each other, with no dependencies. This feature makes it possible to exploit the parallel computing capabilities of graphics hardware.

[0075] Graphics hardware (such as GPUs) has powerful parallel processing capabilities. They have a large number of processing cores and can handle multiple data tasks simultaneously. In the redrawing sectioning algorithm, since the coordinate transformation of each data point is independent, these coordinate transformation tasks can be assigned to different processing cores of the GPU for simultaneous calculation.

[0076] Compared to traditional CPU computing methods, GPUs offer significant advantages in parallel computing. While CPUs typically focus on complex logic control and sequential processing, GPUs excel at handling large-scale parallel data computations. When performing cross-sectional calculations, CPUs require sequential coordinate transformations and related calculations for each data point, which can be time-consuming when processing large amounts of data. GPUs, on the other hand, can perform calculations on all data points simultaneously, significantly reducing computation time.

[0077] By leveraging the parallel capabilities of graphics hardware, the redrawing sectioning algorithm can achieve simultaneous calculation of all data points. This parallel computing method completes the sectioning calculation with extremely low time overhead, greatly improving the efficiency of the sectioning calculation. For example, when processing a three-dimensional CT image containing millions of individual data points, the traditional sequential computing method may take tens of milliseconds or even longer to complete the sectioning calculation. However, by utilizing the redrawing sectioning algorithm provided by the present invention combined with the parallel computing of graphics hardware, it only takes a few milliseconds to complete, greatly improving the user experience and the real-time performance of the system. This enables operators to see the image effect after sectioning in a short period of time, which is of great significance for quickly locating and analyzing areas of interest. Whether it is for quickly detecting dangerous goods in the security inspection field or assisting doctors in diagnosing diseases in the medical field, it can significantly improve work efficiency and accuracy.

[0078] Next, in step S7, the 3D image is repeatedly sliced ​​and redrawn by continuously changing the position of the cutting plane, gradually revealing the contents of the bag at different locations. Each slice and redraw allows security personnel to see more previously unseen parts of the bag. As this process continues, all items in the bag are eventually fully revealed. This series of operations ensures a comprehensive and detailed inspection of all items in the bag, preventing any potentially dangerous items from being missed and ensuring the integrity and accuracy of the security inspection.

[0079] Next, in step S8, after the previous cutting operation has completely exposed all items in the bag, security personnel conduct a detailed inspection of these items. Based on relevant security standards and regulations, they determine whether the items are prohibited items or pose a safety hazard. Upon completion, the inspection results are collated and reported to relevant personnel. This step is the final step in the entire security inspection process, ensuring the integrity of the inspection and providing an accurate basis for subsequent decision-making and handling.

[0080] Second embodiment

[0081] Based on the above method, the second embodiment of the present invention provides a three-dimensional CT image sectioning system based on material recognition. Figure 9 As shown, the system includes one or more processors and a memory. The memory is used to store one or more programs; the programs are executed by the processor to implement the method in the above embodiment.

[0082] The processor is used to control the overall operation of the system to complete all or part of the steps of the above method. The processor can be a central processing unit (CPU), a graphics processing unit (GPU), a field programmable gate array (FPGA), an application-specific integrated circuit (ASIC), a digital signal processing (DSP) chip, etc. The memory is used to store various types of data to support operations in the system. These data may include, for example, instructions for any application or method operating on the system, as well as application-related data. The memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, etc.

[0083] In an exemplary embodiment, the system can be implemented in a variety of ways to perform the above-mentioned method and achieve the corresponding technical effects. Specifically, it can be constructed based on a computer chip or physical form, or it can be implemented with the help of products with specific functions. Common embodiments include CT equipment, security inspection equipment, non-destructive testing equipment, and computers. In the computer field, there are many types of specific devices, such as personal computers, laptop computers, in-vehicle human-computer interaction devices, cellular phones, camera phones, smartphones, personal digital assistants, media players, navigation devices, email devices, game consoles, tablet computers, wearable devices, and systems composed of any combination of these devices.

[0084] In another exemplary embodiment, the present invention further provides a computer-readable storage medium including program instructions, which, when executed by a processor, implement the steps of the method described in any of the above embodiments. For example, the computer-readable storage medium may be the aforementioned memory including the program instructions, which may be executed by a processor to perform the above method and achieve the same technical effects as the above method.

[0085] It should be noted that the above embodiments are merely examples, and the technical solutions of the various embodiments may be combined and are all within the scope of protection of the present invention.

[0086] The above describes in detail the material identification-based 3D CT image sectioning method and system provided by the present invention. For those skilled in the art, any obvious modification without departing from the essence of the present invention would constitute an infringement of the present invention's patent rights and would result in corresponding legal liability.

Claims

1. A 3D CT image sectioning method based on material identification, characterized in that The following steps are involved: S1: The CT device performs a tomographic scan on the baggage to be inspected, obtains and backs up the tomographic data, and analyzes all items in the baggage to be inspected based on the material recognition algorithm to generate the section boundary, the boundaries of all items in the baggage to be inspected, and the sensitivity coefficient of each item; S2: Rendering a 3D CT image using image space volume rendering technology based on the slice data, section boundaries, and object boundaries; S3: Set a threshold for the number of dangerous goods, and determine whether the number of dangerous goods is less than the threshold based on the results of the 3D CT image; if the number of dangerous goods is less than the threshold, proceed to step S4; if the number of dangerous goods is greater than or equal to the threshold, proceed to step S5; S4: Perform a normal section on the baggage to be inspected according to the three-point positioning method to obtain normal section data, and then proceed to step S6; S5: Perform high-level sectioning on the bag to be inspected based on the convex hull solution algorithm of the three-dimensional point set to obtain high-level sectioning data, and then proceed to step S6; S6: redrawing the section data as a new data source of the ray casting algorithm, obtaining and backing up the redrawn 3D CT image; S7: Repeat steps S3 to S6 until all items in the baggage to be inspected are completely exposed; S8: Inspect all exposed items and report the inspection results.

2. The three-dimensional CT image sectioning method according to claim 1, wherein The image space volume rendering technology in step S2 is a ray casting algorithm.

3. The three-dimensional CT image sectioning method according to claim 1, wherein The calculation formula of the sensitivity coefficient in step S1 is: Sensitivity coefficient = importance parameter / occurrence frequency parameter.

4. The three-dimensional CT image sectioning method according to claim 3, wherein The importance parameters of metals, liquids, flammable, explosive, and radioactive materials are greater than those of non-prohibited items; The frequency parameter is determined according to the frequency of occurrence. The higher the frequency of occurrence, the larger the frequency parameter.

5. The three-dimensional CT image sectioning method according to claim 1, wherein The three points used in the three-point positioning method in step S4 are the top three points in terms of sensitivity coefficient.

6. The three-dimensional CT image sectioning method according to claim 1, wherein If the sensitivity coefficient ranking fails in step S4, the luggage to be inspected is sectioned using the geometric center method in step S4.

7. The three-dimensional CT image sectioning method according to claim 1, wherein The convex hull solution algorithm for the three-dimensional point set in step S5 is a fast convex hull algorithm.

8. The three-dimensional CT image sectioning method according to claim 7, wherein Based on the fast convex hull algorithm, three points with the largest sensitivity coefficients are first selected, and then points with a similarity difference greater than a preset threshold are searched in the neighborhood of the three points with the largest sensitivity coefficients, and the points with a similarity difference greater than the preset threshold are added to the convex hull solution point set.

9. The three-dimensional CT image sectioning method according to claim 1, wherein After the 3D CT image sectioning process is completed, if the original 3D CT image needs to be restored, the ray casting algorithm is used based on the pre-backed up slice data of the test line package or the redrawn 3D CT image to select the method of gradually restoring or completely restoring the 3D CT image.

10. A 3D CT image sectioning system based on material identification, characterized in that The method comprises a processor and a memory; wherein the memory is coupled to the processor and is used to store a computer program, and when the computer program is executed by the processor, the processor implements the method according to any one of claims 1 to 9.

Images