Offshore medical ship cone beam CT adaptive image stabilization and rapid diagnosis system

Abstract

The present application relates to the field of ship medical treatment and medical imaging technology, in particular to a sea medical ship cone beam CT adaptive image stabilization and rapid diagnosis system, the system comprises a motion sensing module, a scanning trajectory planning module, a data screening module and an image reconstruction module; the motion sensing module is used for obtaining multi-source data and a motion state vector; the scanning trajectory planning module sets a double-mode scanning mechanism in combination with data, the state vector and historical information, and then plans an adaptive scanning trajectory of the cone beam CT; the scanning trajectory planning module obtains time sequence scanning information output by the cone beam CT based on the scanning trajectory, and the data screening module pairs the time sequence scanning information and the state vector to obtain a target projection sequence; the image reconstruction module performs image reconstruction according to the target projection sequence, and realizes stable imaging and auxiliary diagnosis of the cone beam CT of the medical ship under sea conditions. The present application suppresses motion artifacts through scanning trajectory planning and projection data screening, and clear diagnostic images can be obtained to improve the efficiency of sea first aid.

Description

Technical Field

[0001] This invention relates to the field of marine medical and medical imaging technology, specifically to a cone-beam CT adaptive stabilization and rapid diagnostic system for marine medical ships. Background Technology

[0002] Cone-beam computed tomography (CBCT) primarily uses an X-ray tube to emit a cone-shaped beam of X-rays, which, in conjunction with a flat panel detector, acquires projection data at different angles. It then uses a filtered back-projection algorithm to obtain diagnostic images. CBCT is widely used in medical settings such as oral and maxillofacial imaging, orthopedic diagnosis, interventional surgery guidance, and radiotherapy positioning verification.

[0003] However, existing CBCT systems are primarily designed for terrestrial environments. Their scan trajectory planning, geometric calibration models, and reconstruction algorithms all rely on a stationary scanning platform. Under sea conditions, medical ships experience pitching, rolling, and heave movements due to wind and waves, leading to the following problems in practical applications: 1. Traditional CBCT uses an isocentric circular track for scanning, relying on the relative geometric relationship between the X-ray tube focus and the detector center. However, under sea conditions, the ship's movement causes a shift in the projection angle and detector position, resulting in data that does not meet clinical diagnostic needs; 2. Under sea conditions, the acceleration of ships can reach... First, the motion amplitude far exceeds the error range allowed by the CBCT system, making it impossible to solve the motion artifact problem; second, existing motion compensation technology cannot be adapted to marine application scenarios and cannot achieve stable imaging under sea conditions.

[0004] Based on the diagnostic imaging needs of scenarios such as medical ships, island emergency rescue, and offshore platform operations, it is necessary to design a CBCT adaptive image stabilization and rapid diagnostic system for marine environments to obtain images that meet clinical diagnostic requirements and provide technical support for maritime emergency rescue and telemedicine. Summary of the Invention

[0005] Given the limitations of existing traditional cone-beam CT imaging methods under maritime conditions, and to address the challenges of multi-degree-of-freedom coupled motions (pitch, roll, heave, etc.) on medical ships, this invention provides an adaptive image stabilization and rapid diagnostic system for cone-beam CT on maritime medical ships. The system includes a motion sensing module, a scan trajectory planning module, a data filtering module, and an image reconstruction module. The motion sensing module is designed and installed to collect multi-source data and construct a motion state vector based on this data. The scan trajectory planning module receives the multi-source data and the motion state vector, and, in conjunction with historical diagnostic information, sets a dual-mode scanning mechanism to plan an adaptive scan trajectory for the cone-beam CT. Based on the adaptive scan trajectory, the scan trajectory planning module obtains the temporal scan information output by the cone-beam CT. The data filtering module pairs and filters the temporal scan information and the motion state vector to obtain a target projection sequence. The image reconstruction module reconstructs the image based on the target projection sequence to achieve stable imaging and assisted diagnosis of cone-beam CT on medical ships under maritime conditions.

[0006] The present invention utilizes modules such as motion sensing, adaptive scanning, data filtering and reconstruction to work together to achieve stable imaging and rapid diagnosis of cone-beam CT under sea conditions, providing technical support for maritime emergency rescue and telemedicine.

[0007] Optionally, the design and installation of the motion sensing module, and the acquisition of multi-source data using the motion sensing module, includes: configuring an inertial measurement unit (IMU) and a laser displacement sensor in the motion sensing module; performing coordinate alignment and installation calibration on the IMU and the laser displacement sensor to complete the design and installation of the motion sensing module; acquiring inertial-based ship attitude data using the IMU in the motion sensing module; obtaining laser ranging data through the laser displacement sensor in the motion sensing module; and performing spatiotemporal synchronization alignment processing on the ship attitude data and the laser ranging data in the motion sensing module to output multi-source data. This invention, through the coordinated configuration and spatiotemporal synchronization alignment of the IMU and the laser displacement sensor, achieves real-time perception of the six degrees of freedom motion of the ship, providing a data foundation for subsequent scanning trajectory planning and motion artifact suppression.

[0008] Optionally, the step of constructing a motion state vector based on the multi-source data includes: the motion sensing module setting inertial component fusion weight coefficients and laser component fusion weight coefficients based on the ship attitude data and laser ranging data in the multi-source data; the motion sensing module performing frequency domain complementary weighting and fusion processing on the multi-source data based on the inertial component fusion weight coefficients and the laser component fusion weight coefficients to construct a standardized six-degree-of-freedom motion state vector.

[0009] This invention performs frequency domain complementary weighting and fusion processing on multi-source data to construct a high-precision standardized six-degree-of-freedom motion state vector, which helps to monitor complex ship motion.

[0010] Optionally, the scanning trajectory planning module receives the multi-source data and the motion state vector, and sets up a dual-mode scanning mechanism in conjunction with historical diagnostic information, including: the scanning trajectory planning module receives historical diagnostic information; the scanning trajectory planning module classifies sea state levels; the scanning trajectory planning module analyzes the amplitude and rate of change of the translation component based on the multi-source data and the motion state vector; the scanning trajectory planning module sets up a dual-mode scanning mechanism based on the sea state level, the historical diagnostic information, the amplitude of the translation component, and the rate of change; the dual-mode scanning mechanism includes a motion compensation scanning mechanism and a non-isocentric large FOV scanning mechanism.

[0011] The motion compensation scanning mechanism and the non-isocentric large FOV scanning mechanism of this invention realize adaptive dynamic planning of scanning trajectory under sea conditions, overcoming the problems of geometric relationship destruction and imaging range limitation in traditional cone-beam CT scanning, and helping to improve the imaging quality and diagnostic efficiency of maritime emergency rescue.

[0012] Optionally, the scanning trajectory planning module sets up a dual-mode scanning mechanism based on the sea state level, the historical diagnostic information, the translation component amplitude, and the rate of change, including: when the sea state level is below level three, the dual-mode scanning mechanism adopts a motion compensation scanning mechanism; a spatial coordinate update model is established in the motion compensation scanning mechanism; real-time spatial coordinate information of the X-ray tube focal point and the detector center is obtained using the spatial coordinate update model; a relative position constant constraint function is constructed in the motion compensation scanning mechanism based on the real-time spatial coordinate information; the scanning trajectory planning module constrains the distance between the X-ray tube and the detector through the relative position constant constraint function to maintain the constant projection geometry.

[0013] This invention establishes a spatial coordinate update model under low sea state conditions through a motion compensation scanning mechanism, which can maintain the constancy of the projection geometry under ship motion disturbance.

[0014] Optionally, the scanning trajectory planning module sets up a dual-mode scanning mechanism based on the sea state level, the historical diagnostic information, the translation component amplitude, and the rate of change, including: when the sea state level is greater than or equal to level three, the dual-mode scanning mechanism adopts a non-isocentric large FOV scanning mechanism; a non-isocentric geometric constraint function is set in the non-isocentric large FOV scanning mechanism; an effective imaging volume expansion model is constructed in the motion compensation scanning mechanism by combining the similar triangle imaging principle and the non-isocentric geometric constraint function; the scanning trajectory planning module uses the effective imaging volume expansion model to analyze the effective imaging range of the target area within the detector to ensure imaging integrity under sea state conditions.

[0015] This invention dynamically analyzes the effective imaging range of the target area within the detector using a non-isocentric large FOV scanning mechanism under high sea state conditions, ensuring the integrity and feasibility of imaging even under sea conditions.

[0016] Optionally, the step of planning the adaptive scanning trajectory of cone-beam CT based on the dual-mode scanning mechanism includes: setting joint judgment conditions for the dual-mode scanning mechanism in the scanning trajectory planning module based on the dual-mode scanning mechanism operation mechanism; the scanning trajectory planning module completing the adaptive switching between the motion compensation scanning mechanism and the non-isocentric large FOV scanning mechanism according to the joint judgment conditions, and planning the adaptive scanning trajectory of cone-beam CT.

[0017] This invention sets joint judgment conditions for a dual-mode scanning mechanism to achieve adaptive switching between low sea state motion compensation mode and high sea state non-isocentric large FOV mode, thereby improving imaging stability in sea state scenarios.

[0018] Optionally, the scanning trajectory planning module obtains the temporal scanning information output by cone-beam CT based on the adaptive scanning trajectory, including: the scanning trajectory planning module adjusting the rotating gantry turntable device of the cone-beam CT based on the adaptive scanning trajectory; and the scanning trajectory planning module acquiring the temporal scanning information output by the cone-beam CT by controlling the X-ray tube and detector in the rotating gantry turntable device.

[0019] This invention adjusts the rotating gantry turntable of cone-beam CT to make the exposure nodes and acquisition timing match the real-time motion state of the ship, reducing invalid projections and providing data for subsequent data screening and artifact-free reconstruction.

[0020] Optionally, the data filtering module performs pairing and filtering of the temporal scan information and the motion state vector to obtain the target projection sequence, including: the data filtering module determining the acquisition time of each frame of projection from the temporal scan information; the data filtering module performing data pairing based on the acquisition time and the motion state vector to obtain the motion state vector corresponding to the acquisition time of each frame of projection; combining the motion state vector corresponding to the acquisition time of each frame of projection to obtain a paired data stream of projection and motion; obtaining a reference projection sequence based on the redundancy angle and the paired data stream of projection and motion; introducing a real-time motion perturbation weight factor and improving the SSIM algorithm formula with reference to the reference projection sequence and the paired data stream to obtain an improved SSIM algorithm formula; extracting the image to be compared from the temporal scan information and determining the reference image from the reference projection sequence; analyzing the structural similarity between the image to be compared and the reference image based on the reference projection sequence and the improved SSIM algorithm formula; and the data filtering module filtering the images to be compared in the temporal scan information based on the structural similarity to obtain the target projection sequence.

[0021] This invention introduces a real-time motion perturbation weighting factor to improve the traditional SSIM structural similarity algorithm, reducing the probability of motion artifacts under sea conditions and further improving the clarity and reliability of diagnostic images.

[0022] Optionally, the image reconstruction module performs image reconstruction based on the target projection sequence to achieve stable imaging and auxiliary diagnosis of cone-beam CT on a medical ship under sea conditions, including: the image reconstruction module receiving the target projection sequence, the multi-source data, and the motion state vector; the image reconstruction module reconstructing the CT image based on the target projection sequence, the multi-source data, and the motion state vector to obtain the reconstructed CT image; and achieving stable imaging and auxiliary diagnosis of cone-beam CT on a medical ship under sea conditions based on the reconstructed CT image.

[0023] This invention reconstructs and optimizes based on target projection sequences, multi-source motion data, and motion state vectors, achieving geometric offset compensation caused by ship motion and providing a reliable diagnostic image processing technology for maritime emergency rescue. Attached Figure Description

[0024] Figure 1 This is a flowchart of the adaptive image stabilization and rapid diagnosis system for cone-beam CT on a marine medical ship according to the present invention; Figure 2 This is a structural diagram of the cone-beam CT adaptive image stabilization and rapid diagnostic system for marine medical vessels according to the present invention. Detailed Implementation

[0025] Specific embodiments of the present invention will now be described in detail. It should be noted that the embodiments described herein are for illustrative purposes only and are not intended to limit the invention. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the invention. However, it will be apparent to those skilled in the art that these specific details are not necessary to practice the invention. In other instances, well-known circuits, software, or methods have not been specifically described to avoid obscuring the invention.

[0026] Throughout this specification, references to "an embodiment," "an embodiment," "an example," or "an example" mean that a particular feature, structure, or characteristic described in connection with that embodiment or example is included in at least one embodiment of the invention. Therefore, the phrases "in an embodiment," "in an embodiment," "an example," or "an example" appearing in various places throughout the specification do not necessarily refer to the same embodiment or example. Furthermore, specific features, structures, or characteristics can be combined in one or more embodiments or examples in any suitable combination and / or sub-combination. Moreover, those skilled in the art will understand that the illustrations provided herein are for illustrative purposes and are not necessarily drawn to scale.

[0027] To address the technical shortcomings of existing traditional cone-beam CT imaging methods in marine environments, and to overcome the multi-degree-of-freedom coupled motions (pitch, roll, heave, etc.) experienced by medical ships due to wind and waves, this invention aims to obtain medical images that meet clinical diagnostic standards, providing technology for offshore emergency medical services and remote medical systems. For details on the adaptive image stabilization and rapid diagnostic system for cone-beam CT on maritime medical ships provided by this invention, please refer to [link to related documentation]. Figure 1 It includes the following steps: This embodiment incorporates a motion sensing module, a scan trajectory planning module, a data filtering module, and an image reconstruction module in the cone-beam CT adaptive stabilization and rapid diagnostic system for marine medical ships.

[0028] S1. Design and install the motion sensing module, collect multi-source data using the motion sensing module, and construct motion state vectors based on the multi-source data. The specific implementation details are as follows: First, the motion sensing module was designed and installed.

[0029] In this embodiment, in order to ensure high-precision imaging of the shipborne CBCT system under complex sea conditions, the design and installation of the motion sensing module must be carried out in strict accordance with the technical specifications. The motion sensing module in this embodiment adopts a dual-sensor architecture that integrates an inertial measurement unit (IMU) and a laser displacement sensor, and provides a data foundation for subsequent motion compensation and trajectory planning through multi-source data complementarity.

[0030] The first step is to configure an inertial measurement unit and a laser displacement sensor in the motion sensing module.

[0031] The hardware in the motion sensing module must meet the adaptability requirements of the shipboard environment. All components in the module must undergo wide-temperature range operation testing before assembly (the wide-temperature range in this example is...). to Salt spray corrosion protection testing and functional stability testing under continuous random vibration environment are conducted to effectively avoid performance drift or failure under long-term offshore operation conditions.

[0032] To effectively acquire attitude and acceleration data, the inertial measurement unit (IMU) selected a nine-axis MEMS inertial sensor that meets the shipboard vibration resistance requirements, with a zero-bias stability of less than [value missing]. This ensures reliable rotational and translational acceleration data even under continuous ship vibration. The laser displacement sensor uses a high-sampling-rate triangular reflector or Time-of-Flight (ToF) ranging sensor with a sampling rate no lower than [missing information]. The resolution needs to be better than This allows for the real-time measurement of the relative distance between the detector and the object being inspected.

[0033] The second step involves coordinate alignment and installation calibration of the inertial measurement unit and the laser displacement sensor to complete the design and installation of the motion sensing module.

[0034] This embodiment establishes a unified coordinate mapping relationship to further achieve effective motion compensation. The inertial measurement unit (IMU) should be physically installed near the center of gravity of the CBCT frame to minimize measurement errors caused by the lever arm effect. Simultaneously, during installation, the three axes of the IMU must be strictly aligned with the ship's six-degree-of-freedom coordinate system.

[0035] The definition of rotational degrees of freedom in the coordinate system of the above coordinate mapping relationship is as follows: The axis is roll. The axis corresponds to the pitch. The axis corresponds to yaw; the translational degrees of freedom correspond to... Surge Sway The heave mechanism, which maps the six degrees of freedom coordinate system to the above-mentioned coordinate system, applies to the entire processing flow of this invention, achieving coordinate alignment and providing reference coordinates for subsequent motion compensation and trajectory planning.

[0036] Module installation and calibration are crucial steps in eliminating systematic errors. In this embodiment, the installation angle deviation of the inertial measurement unit (IMU) needs to be controlled within 0.6 degrees. Simultaneously, the IMU must be statically calibrated before installation, followed by temperature compensation coefficient calibration under varying temperature conditions. This helps to eliminate the impact of drastic temperature changes in the marine environment on measurement accuracy.

[0037] The laser displacement sensor needs to be securely mounted on a fixed bracket on the side wall of the gantry, ensuring that the laser beam is absolutely aligned with the detector panel. This facilitates the real-time acquisition of relative distance data. In this embodiment, to prevent additional errors introduced by thermal expansion and contraction or vibration coupling, the material and mechanical structure of the fixed bracket must be completely consistent with the CBCT gantry. After the sensor is installed, a zero-point calibration procedure is performed, that is, the reference distance is measured under both no-load and full-load conditions. The two sets of reference values ​​are used as the reference for normalization in subsequent distance weighting correction. Through the above standardized design and installation process, a reliable motion sensing module can be obtained.

[0038] After completing the design and installation of the motion sensing module, the following implementation steps are followed to collect and process multi-source data in a spatiotemporal manner, providing data support for subsequent motion compensation.

[0039] The first step is to use the inertial measurement unit (IMU) in the motion sensing module to collect inertial-based hull attitude data.

[0040] Inertial measurement unit (IMU) data acquisition is fundamental to obtaining dynamic information about the ship's hull. In this example, the minimum data output frequency of the IMU must first be set to... This helps to continuously collect raw motion data of the ship's three-axis angular velocity and three-axis acceleration. At the same time, it makes full use of the digital motion processor (DMP) built into the inertial measurement unit, which can realize sensor bias compensation and correction of operating temperature drift, further ensuring the accuracy of the raw data.

[0041] After obtaining the above data, attitude adjustment is performed based on the ship's six-degree-of-freedom coordinate system mapping rules. According to the above implementation details, the specific mapping relationship is as follows: , , , , , ;in The axis corresponds to the roll (Roll) axis. The axis corresponds to the pitch. The axis corresponds to yaw. The axis corresponds to Surge, The axis corresponds to the sway. The axis corresponds to the heave (Heave) axis. The module also needs to simultaneously output the calculated data of the ship's three-axis rotation angles and three-axis translational accelerations that match the above dimensions.

[0042] The second step involves acquiring laser ranging data using the laser displacement sensor in the motion sensing module.

[0043] The laser displacement sensor in this embodiment can provide high-frequency relative distance correction information. This sensor is fixed to... Operating at its designated frequency, the system can acquire the raw relative distance between the probe and the measured ship structure in real time. In this embodiment, considering the high-frequency electromagnetic interference and ranging jitter noise present in the shipboard environment, a 5-point window sliding mean filtering algorithm is used at the module's data acquisition end to ensure both the stability and feasibility of the output data. This algorithm extracts the arithmetic mean of five consecutive sampling points, effectively smoothing and filtering out high-frequency noise interference, ultimately outputting stable and reliable relative distance data, thus obtaining laser ranging data.

[0044] The third step, the motion sensing module, performs spatiotemporal synchronization and alignment processing on the ship's attitude data and laser ranging data to output multi-source data.

[0045] To achieve effective fusion of ship attitude data and laser ranging data, this embodiment performs spatiotemporal synchronization alignment on the two types of heterogeneous data. During time synchronization, the two types of data can be synchronized via hardware trigger signals, ensuring that the timestamp deviation of the devices is controlled within a specified range. Within this range. Simultaneously, the module uses a unified shipborne local clock or GPS satellite clock as the time reference, and timestamps all collected data based on this to further eliminate clock drift between multiple devices; during the timing alignment process, in order to accurately match the timing requirements of the laser sensor, it is necessary to... The ship's attitude data undergoes upsampling processing. In this embodiment, a zero-order hold algorithm is used to increase the frequency of the ship's attitude data to [a higher level]. After upsampling and timestamp alignment, the inertial motion data and laser ranging data can be synchronized in time and matched in spatial dimension, thus completing the integrated spatiotemporal alignment of multi-source data and finally obtaining the multi-source data of this embodiment.

[0046] Finally, a standardized six-degree-of-freedom motion state vector is constructed based on multi-source data.

[0047] After completing the acquisition of multi-source data and spatiotemporal synchronization alignment, this embodiment further constructs a standardized six-degree-of-freedom motion state vector through a frequency domain complementary weighted fusion strategy. Based on this, the low-frequency dynamic response characteristics of the inertial measurement unit (IMU) and the high-frequency absolute position measurement characteristics of the laser displacement sensor can be combined to further eliminate the drift error caused by long-term integration of the inertial measurement unit (IMU) and at the same time make up for the signal jitter defect of the laser sensor under severe vibration.

[0048] The first step of the motion perception module sets the inertial component fusion weight coefficient and the laser component fusion weight coefficient based on the ship attitude data and laser ranging data from the multi-source data.

[0049] This embodiment employs a frequency-domain complementary weighted fusion strategy. First, the fusion weighting coefficient matrix for the inertial and laser components is defined. Inertial data is set to dominate the system's low-frequency stability, while laser data is responsible for correcting high-frequency disturbances. In this embodiment, the fusion weighting coefficients for the inertial components are defined. The fusion weighting coefficient for the laser components is 0.7. The value is 0.3, and the weighting coefficient satisfies... With the normalization constraint, the above weight allocation can ensure that the motion state vector maintains the overall attitude trend stability while having the ability to respond quickly and correct for high-frequency micro-amplitude vibrations.

[0050] The second step, the motion sensing module, performs frequency domain complementary weighting and fusion processing on the above multi-source data based on the inertial component fusion weighting coefficient and the laser component fusion weighting coefficient, which can then construct a standardized six-degree-of-freedom motion state vector.

[0051] Next, the embodiment performs frequency domain complementary weighting and fusion processing on the spatiotemporally aligned multi-source data based on the aforementioned weighting coefficients, wherein the real-time six-degree-of-freedom motion state vector of the ship's hull is set as follows: And the vector equation expression for the motion state vector is as follows: , in, express The standardized six-degree-of-freedom motion state vector output at each moment. This is represented as the original motion vector obtained by the IMU after bias compensation, temperature correction, and attitude calculation. This represents error compensation and vector superposition operations based on frequency domain characteristics. The fusion weighting coefficient represents the inertial component. This represents the fusion weighting coefficient of the laser components. This represents the high-frequency displacement correction vector after laser filtering, noise reduction, and differential correction.

[0052] Based on the vector equation expression, it can be seen that the above motion state vector is the weighted fusion result of the inertial solution component and the laser correction component; the original motion vector mainly includes the rotation angle and the translation integral displacement.

[0053] After fusion computation, to make the six-degree-of-freedom real-time motion state vector The timing requirements of the matching system are mainly comprised of six physical quantities, and their mathematical expressions are as follows: , in, These are three-axis rotational components, corresponding to the ship's roll, pitch, and yaw angles, respectively, with the unit set to milliradians (mrad). These components are mainly derived from the calculations of the inertial measurement unit (IMU), and the zero drift situation is effectively suppressed by using a complementary filtering algorithm.

[0054] at the same time, These are three-axis translational components, corresponding to the hull's surge, sway, and heave displacements, respectively, with the unit set in millimeters (mm). These components mainly utilize laser ranging data to perform high-frequency real-time correction on the inertial measurement unit's quadratic integration results, ensuring the absolute accuracy of displacement measurement and dynamic response characteristics. The motion state vector can be constructed through the above steps.

[0055] S2. The scan trajectory planning module receives the above multi-source data and motion state vectors, and sets up a dual-mode scanning mechanism in conjunction with historical diagnostic information. Based on the dual-mode scanning mechanism, it plans the adaptive scan trajectory of cone-beam CT. The specific implementation steps and contents are as follows: First, the scanning trajectory planning module can receive multi-source data and motion state vectors in real time, and combine them with historical diagnostic information to set up a dual-mode scanning mechanism, which can ensure that the scanning trajectory planning module can adjust its strategy according to the actual sea conditions and the state of the object being inspected.

[0056] The first step is for the trajectory planning module to receive historical diagnostic information.

[0057] In this embodiment, the scanning trajectory planning module can automatically retrieve the historical diagnostic records of the subject and extract three types of key prior information to provide an initial benchmark for the scanning mechanism, avoiding the need for the algorithm to be calculated from scratch. In a specific embodiment, firstly, the spatial location and size range parameters of the target anatomical region are extracted to determine the required scanning field of view, which is then used as the initial parameters for the geometric design under the non-isocentric large FOV scanning mechanism. Secondly, the motion artifact distribution characteristics recorded in previous scans are extracted to initialize the weight parameters of the motion compensation algorithm, enabling the system's scanning trajectory planning module to predict motion compensation for a specific subject during the first frame acquisition, effectively avoiding large-area artifacts in the first frame. Finally, the basic parameters of the subject are obtained, allowing the initial distance between the detector and the X-ray tube to be set, effectively improving the rationality of the initial settings.

[0058] The second step, the scanning trajectory planning module, classifies the sea state levels. In this embodiment, the scanning trajectory planning module combines external environmental data and ship attitude data collected by the motion sensing module to perform real-time comparative evaluation of the current sea state level. Specifically, sea state level three is used as the core judgment threshold. When the sea state level is below level three, it is determined to be a stable condition with small ship sway, and the system can maintain the motion compensation scanning mechanism to maintain the stability of the projection geometry through a relative position constant constraint function to obtain the highest quality image. When the sea state level is greater than or equal to level three, it is determined to be a disturbance condition with large ship sway. The system triggers a non-isocentric large FOV scanning mechanism, which expands the effective imaging volume through a non-isocentric geometric constraint function to ensure the integrity of the imaging under motion interference. The above-mentioned three-level sea state classification is the key trigger condition for the adaptive switching of the dual-mode scanning mechanism, enabling the system to adapt to ship sway in moderate sea states in advance and avoid scanning failures due to excessive motion amplitude.

[0059] The third step, the trajectory planning module, analyzes the magnitude and rate of change of the translation component based on multi-source data and motion state vectors.

[0060] The scanning trajectory planning module is based on The high-frequency real-time receiving motion sensing module pushes standardized six-degree-of-freedom motion state vectors. The above vector is formed by weighted fusion of the inertial component and the laser correction component, and its expression is: And specifically includes the following six-dimensional physical quantities. .

[0061] To ensure the reliability of the input data, the module receiver needs to perform validity checks on each frame of data. When a packet loss rate exceeding a certain threshold is detected, the receiver will immediately initiate a validation process. When an error occurs, the frame will be automatically discarded and the system will wait for the next frame of valid data, which can prevent the error information from being passed downstream.

[0062] Meanwhile, the magnitude and rate of change of the translation component are obtained based on multi-source data and motion state vectors.

[0063] The scanning trajectory planning module in this embodiment is mainly based on motion state vectors. The three-axis translation components The amplitude and rate of change of the translation component are extracted and calculated in real time. The specific processing flow is as follows: First, the magnitude of the translation component is calculated, and the translation vector at the current moment is extracted. The spatial Euclidean norm is calculated to analyze the magnitude of the translational component of the ship's absolute displacement at the current moment.

[0064] Next, the rate of change of the translation component is calculated. The module uses a sliding time window mechanism to extract two adjacent sampling periods, with a time step of [value missing]. The translation vector within the time step is calculated, and the difference vector between the two is calculated. The magnitude of the difference vector is divided by the time step to obtain the instantaneous translation rate at the current moment, which is the translation speed amplitude in this embodiment.

[0065] The module performs smoothing filtering on the amplitude and rate of change of the above translation components, eliminates instantaneous extreme values ​​caused by high-frequency noise from the sensor, and updates the processed feature parameters to the decision buffer of the trajectory planning module in real time, providing a reference for the subsequent quantitative analysis of the dynamic switching dual-mode scanning mechanism.

[0066] The dual-mode scanning mechanism in this embodiment mainly includes a motion compensation scanning mechanism and a non-isocentric large FOV scanning mechanism. When the system assesses that the amplitude and rate of change of the translational component exceed the conventional threshold, or the sea state reaches level three or above, the motion compensation scanning mechanism is activated or strengthened to cope with hull vibration. At the same time, combined with the anatomical region and body shape parameters in the historical diagnostic file, the non-isocentric large FOV scanning mechanism is activated in a coordinated manner. By dynamically adjusting the geometric distance between the detector and the X-ray tube, complete coverage of the target area is ensured under complex motion conditions. However, the effective working boundary of this dual-mode scanning mechanism is set at sea state three, where the wave height reaches 2.5-4.0 meters. When the sea state is greater than or equal to level three, because the amplitude of ship motion exceeds the system safety and imaging physical limits, the dual-mode scanning mechanism will stop being applied and the scanning operation will be terminated to avoid invalid exposure or equipment damage.

[0067] In one embodiment, the motion compensation scanning mechanism in the dual-mode scanning mechanism operates as follows: When the sea state level is below level three, the dual-mode scanning mechanism adopts a motion compensation scanning mechanism.

[0068] The motion compensation scanning mechanism in this embodiment is mainly used in environments with sea state levels below level three. Under such low sea state conditions, the translational motion component of the ship is within the effective compensation range of the mechanical actuator. The core of the above mechanism lies in driving the X-ray tube and the detector to perform synchronous reverse motion, further offsetting the influence of ship displacement, thereby maintaining a constant relative position of the projected illumination line with respect to the object under test, and ensuring the stability of the imaging geometry. The specific implementation details are as follows: The first step is to establish a spatial coordinate update model in the motion compensation scanning mechanism.

[0069] This embodiment first constructs a spatial coordinate update model to analyze the spatial position information of the X-ray tube focus and the detector center in real time. The model uses the coordinates of the motion compensation trigger time (or the time of locking the previous frame of valid data) as the initial reference. Combined with the real-time translational velocity of the subject, the cumulative displacement deviation changes over time is calculated. After obtaining the cumulative displacement deviation through integration, the system substitutes it into the model formula and subtracts the deviation value from the initial reference coordinates to obtain the real-time spatial coordinates of the X-ray tube focus and the detector center at different times.

[0070] The spatial coordinate update model in this embodiment satisfies the following relationship: , in, and They represent Real-time spatial coordinates of the X-ray tube focal point and the detector center. , These represent the initial reference coordinates at the motion compensation trigger time (or the time when the valid data of the previous frame is locked), respectively. Indicates based on the real-time translation speed of the subject The cumulative displacement deviation is obtained by integration.

[0071] And the cumulative displacement deviation satisfies the following relationship: , in, Indicates based on the real-time translation speed of the subject In time interval The cumulative displacement deviation obtained by internal integration. Indicates that the subject is in The real-time translation speed at any given moment.

[0072] The second step involves updating the model using the aforementioned spatial coordinates. The module processes the collected motion data in real time to obtain the real-time spatial coordinate information of the X-ray tube focus and the detector center at the current moment, providing data support for subsequent constraint control.

[0073] The third step involves constructing a relative position constant constraint function within the motion compensation scanning mechanism based on real-time spatial coordinate information.

[0074] The embodiment constructs a relative position constant constraint function based on real-time acquired spatial coordinate information, and satisfies the following relationship: , in, and They represent Real-time spatial coordinates of the X-ray tube focal point and the detector center. , These represent the initial reference coordinates at the motion compensation trigger time (or the time when the previous frame's valid data was locked).

[0075] The relative position constant constraint function can forcibly lock the spatial distance between the X-ray tube and the detector, keeping it constant. By maintaining a constant distance, the stability of the scanning magnification can be guaranteed, and the projected illumination line can be made to be relatively stationary in the coordinate system of the object under test, thereby effectively eliminating motion blur caused by the swaying of the ship.

[0076] The fourth step, the scanning trajectory planning module, constrains the distance between the X-ray tube and the detector through a relative position constant constraint function to maintain the constant projection geometry. This means that the motion trajectory of the X-ray tube and the detector is adjusted in real time to further limit the change in the distance between them, ensuring that the projection geometry remains highly constant throughout the scanning process and effectively achieving high-precision motion compensation scanning.

[0077] In one embodiment, the non-isocentric large FOV scanning mechanism in the dual-mode scanning mechanism is as follows: In one embodiment, the non-isocentric large FOV scanning mechanism in the dual-mode scanning mechanism is specifically configured as follows: when the system assesses the current sea state level as greater than or equal to level three, or detects that the real-time amplitude and rate of change of the ship's translational component continuously exceed the effective tolerance range of the motion compensation mode, the system will trigger the non-isocentric large FOV scanning mechanism. This embodiment's mechanism can abandon absolute isocentric locking of the relative position of the projection line, instead dynamically adjusting geometric optical parameters, such as source-image distance or offset, to obtain a larger motion tolerance space, thereby ensuring complete coverage of the target anatomical region even under conditions of significant swaying. The specific implementation steps are as follows: First, set the non-isocentric geometric constraint function in the non-isocentric large FOV scanning mechanism.

[0078] First, establish non-isocentric geometric constraint functions as boundary conditions for scanning trajectory planning, satisfying the following relationship: , in, This indicates the distance from the detector to the center of rotation of the object being inspected. This indicates the distance from the focal point of the X-ray tube to the object being examined. This represents the minimum total source-detector distance required to cover the target area.

[0079] This design allows the detector to be as close as possible to the object being examined to expand the field of view, while also meeting the system's minimum imaging distance requirements.

[0080] In geometric constraints This allows the detector to be placed as close as possible to the object being inspected, thereby expanding the field of view; geometric constraints in It can meet the minimum imaging distance requirement of the system.

[0081] The function described above includes two constraints: one is the distance from the detector to the center of rotation of the object being inspected. It needs to be much smaller than the distance from the X-ray tube focal spot to the subject. (Right now Secondly, the sum of the two must satisfy the minimum total source-detector distance required to cover the target area. (Right now ), based on this execution The constraints force the detector to be as close as possible to the object being inspected, thereby maximizing the field of view in physical space; at the same time, through The constraints ensure that the system always meets the minimum imaging distance requirement and guarantees the feasibility of the imaging optical path.

[0082] The second step combines the principle of similar triangle imaging with the non-isocentric geometric constraint function to construct an effective imaging volume expansion model in the motion compensation scanning mechanism.

[0083] By combining the imaging principle of similar triangles with the aforementioned non-isocentric geometric constraint function, an effective imaging volume expansion model is constructed, which can quantify and optimize the field of view size. The model formula is defined as follows: , in, Indicates the effective imaging coverage volume. Indicates the instantaneous scan layer thickness along the ray direction. Indicates the effective area of ​​the detector. This indicates the distance from the detector to the center of rotation of the object being inspected. This indicates the distance from the focal point of the X-ray tube to the object being examined.

[0084] Based on the model, it can be seen that when sea conditions are severe, the control logic reduces [the impact of these factors]. and increase , making the ratio Significantly reduced in detector area Under constant conditions, the above-mentioned geometric parameter adjustments can significantly expand the field of view. The above measures further ensure that even when significant ship rocking causes a severe shift in the X-ray center, the target dissection area still falls within the effective imaging range of the detector, thus guaranteeing the integrity and continuity of the scanning data.

[0085] The third step, the scanning trajectory planning module, uses an effective imaging volume expansion model to analyze the effective imaging range of the target area within the detector, ensuring imaging integrity under sea conditions.

[0086] In this embodiment, the scanning trajectory planning module utilizes an effective imaging volume expansion model to perform real-time analysis and dynamic planning of the effective imaging range of the target area within the detector. In one embodiment, firstly, the scanning trajectory planning module receives motion data from the ship's attitude sensor in real time, and combines it with the effective imaging volume expansion model to calculate the optimal solution of geometric parameters under the current sea state severity. The model uses quantization ratios... By observing the changes, the effective imaging volume that the detector can cover under the current geometric configuration can be calculated. .

[0087] Finally, based on the above analysis, the scanning trajectory planning module can maximize the imaging range in harsh environments with sea state level greater than or equal to level three. The mechanism of this embodiment effectively avoids the truncation of anatomical structures or loss of projection data caused by large-scale hull movements, thereby ensuring the integrity of imaging data.

[0088] To further clarify the dual-mode scanning mechanism of the scanning trajectory planning module in this embodiment, please refer to Table 1 for its specific triggering conditions and execution strategies.

[0089] Table 1. Information on the dual-mode scanning mechanism in the scanning trajectory planning module.

[0090] Then, the scan trajectory planning module plans the adaptive scan trajectory of cone-beam CT based on the dual-mode scanning mechanism.

[0091] The first step is to set the joint decision conditions for the dual-mode scanning mechanism in the scan trajectory planning module. These joint decision conditions do not rely on a single sea state signal, but rather comprehensively consider environmental parameters and system execution capabilities. The specific decision logic is as follows: The scanning trajectory planning module first reads the sea state monitoring data and uses sea state level 3 as the core judgment threshold. Combined with the triggering conditions of the motion compensation scanning mechanism, when the sea state level is lower than level 3, the system further evaluates the real-time amplitude and rate of change of the ship's translation component. Only when both of the above dynamic indicators are within the effective compensation bandwidth of the mechanical actuator is the motion compensation condition determined to be met.

[0092] The triggering conditions for the non-equicenter large FOV scanning mechanism are as follows: when any of the following conditions are met, the module determines that the mechanical compensation has failed and triggers the non-equicenter large FOV scanning mechanism: first, the sea state level reaches or exceeds level three; second, although the sea state level is lower than level three, the rate of change of the ship's translational component continues to exceed the dynamic tolerance range of the motion compensation mode.

[0093] The second step, the scan trajectory planning module, completes the adaptive switching between motion compensation scanning mechanism and non-isocentric large FOV scanning mechanism based on the joint judgment conditions, and plans the adaptive scan trajectory of cone-beam CT.

[0094] Trajectory planning under motion compensation scanning mechanism: When the joint judgment conditions meet the motion compensation requirements, a scanning trajectory with relative position locking needs to be planned. In the above trajectory, the X-ray tube focal spot and the detector center are controlled to move synchronously in opposite directions, and the cumulative displacement deviation is calculated and offset. ,make sure The trajectory planning method, which always equals the initial baseline difference, ensures that the projected radiation is absolutely stationary relative to the subject, which is beneficial for obtaining clear images with high spatial resolution during cone-beam CT scanning.

[0095] Trajectory planning under a non-equicenter large FOV scanning mechanism: When the joint decision condition triggers a tolerance exceedance, the module needs to generate a scan trajectory with optimized geometric parameters. In this trajectory, the module does not perform synchronous reverse locking, but instead relies on the non-isocentric geometric constraint function (…). and A motion path is planned that brings the detector extremely close to the object being examined while simultaneously increasing the distance between the detector and the X-ray tube. This trajectory maximizes the effective imaging volume. Even if the center of the cone beam shakes violently with the hull, the target dissection area can still fall within the detector's wide field of view.

[0096] Through the above steps, the scanning trajectory planning module realizes the intelligent and adaptive nature of cone-beam CT scanning trajectory, ensuring that the system can output effective data that meets diagnostic needs in all sea conditions.

[0097] S3. The scanning trajectory planning module obtains the temporal scanning information output by cone-beam CT based on the adaptive scanning trajectory. The data filtering module pairs and filters the temporal scanning information and motion state vectors to obtain the target projection sequence. The specific implementation details are as follows: First, the scan trajectory planning module obtains the temporal scan information output by cone-beam CT based on the adaptive scan trajectory.

[0098] The first step, the scanning trajectory planning module, is based on the adaptive scanning trajectory control of the rotating gantry turntable device of cone-beam CT.

[0099] In this embodiment, the scanning trajectory planning module sends motion control commands to the host computer of the cone-beam CT system based on the aforementioned determined adaptive scanning trajectory parameters, including but not limited to target angular velocity and acceleration time, to start the rotating gantry turntable device. After receiving the commands, the host computer drives the servo motor to control the turntable to accelerate. During this process, it is necessary to monitor the real-time rotation speed of the turntable until it reaches the preset stable angular velocity, and ensure that the rotation speed fluctuation is within the allowable error range, thereby establishing a stable motion reference for subsequent continuous scanning.

[0100] As the turntable accelerates and stabilizes, the scanning trajectory planning module sends a data acquisition start command to the detector through a synchronous triggering mechanism. This command sets the detector's fixed frame rate and exposure parameters, further ensuring that the detector completes initialization and is ready to receive X-ray signals before the turntable enters the uniform rotation stage. In this embodiment, the synchronous control mechanism ensures strict alignment of mechanical movement and data acquisition on the time axis, avoiding the loss of projection data or angle deviation caused by inconsistent start-up timing.

[0101] The second step, the scanning trajectory planning module, acquires the temporal scanning information output by controlling the X-ray tube and detector in the rotating gantry turntable device.

[0102] In this embodiment, after the rotating gantry turntable reaches a stable operating state, the scanning trajectory planning module controls the acquisition actions performed by the X-ray tube and detector according to the geometric constraints defined in the adaptive scanning trajectory. During the scanning process, the X-ray tube emits a cone beam of X-rays that penetrates the subject, and the detector receives the attenuated X-ray signal and converts it into an analog electrical signal, which is then processed by analog-to-digital conversion to form digital projection data.

[0103] As the turntable continues to rotate, the X-ray tube and detector strictly follow the planned path of the adaptive scanning trajectory at every moment. The scanning trajectory planning module can record the state of each frame of projection data in real time, including but not limited to the rotation angle of the turntable, the acquisition time of the detector, and the current scanning mode identifier. It mainly contains the raw data of spatial geometric information and time series characteristics, and finally forms the time-series scanning information containing timestamps, projection angles and grayscale information, providing basic data for subsequent data filtering and image reconstruction.

[0104] Then, the data filtering module pairs and filters the temporal scan information and motion state vectors to obtain the target projection sequence.

[0105] The first step, the data filtering module, determines the acquisition time of each frame of projection from the time-series scan information.

[0106] In one specific embodiment, the data filtering module parses the metadata in the time-series scan information. This metadata, synchronously written by the detector hardware upon exposure completion, includes a microsecond-level hardware trigger timestamp, further locating the acquisition time of each frame of the projected image and recording it as... ,time It corresponds to the physical time point when the detector completes X-ray exposure, photoelectric signal conversion and outputs a complete digital image. Using this as the time reference for subsequent data association can avoid time alignment errors introduced by data transmission delays and effectively meet the time alignment requirements in high-dynamic motion scenarios at sea.

[0107] The second step, the data filtering module, pairs data based on the acquisition time and the motion state vector to obtain the motion state vector corresponding to the acquisition time of each frame projection.

[0108] In this embodiment, the data filtering module is based on the data collection time. The output of the motion sensing module Retrieve the six-degree-of-freedom motion state vector at the corresponding moment from the high-frequency motion data stream. Further, for different scenarios in the implementation, two matching algorithms can be selected: when the output frame rate of the motion state vector is more than 5 times higher than the detector's acquisition frame rate, the nearest neighbor matching algorithm should be used to directly select the nearest neighbor vector. The motion state vector with the smallest time difference is used as the pairing result; when the frame rate difference is less than 5 times, a linear interpolation algorithm is used. Using two adjacent motion state vectors as a reference, the result is obtained through time-weighted interpolation. The precise motion state vector corresponding to each moment. After pairing, the output motion state vector must fully contain the real-time displacement of the three translational degrees of freedom (sway, roll, heave) and the real-time rotation angle of the three rotational degrees of freedom (pitch, roll, yaw). All dimensions are uniformly adopted in the International System of Units (SI), with displacement in millimeters and angle in radians, to ensure the consistency of dimensions in subsequent data processing.

[0109] The third step involves combining the motion state vector corresponding to each frame of projection acquisition to obtain a paired data stream of projection and motion. In this embodiment, each frame of projection image is combined with its corresponding motion state vector to form a paired data stream of projection and motion. The data stream in this embodiment not only contains image grayscale information but also real-time pose perturbation information of the image acquisition platform, providing an information basis for the subsequent introduction of motion constraints.

[0110] The fourth step is to obtain a reference projection sequence based on the redundant angles and the paired data stream of projection and motion.

[0111] In a specific embodiment, a reference projection sequence is mainly generated based on the aforementioned redundant angles and paired data streams. First, the redundant angles are determined according to the preset scanning parameters of cone-beam CT. Redundancy angles in typical shipboard scenarios are set as follows: This ensures that the original acquired time-series scan information contains theoretical... Sufficient additional projection frames are retained after the scan endpoint to prevent the 360° closure point from shifting out of the original sequence range due to hull motion. Then, from the end of the paired data stream sequence, the theoretical... Projection after position As the starting point for backtracking, backtracking forward. Aspect projection to All projection frames within this interval are combined to form a reference projection sequence.

[0112] in The value is dynamically calculated based on the detector frame rate and the turntable's rated speed: , in, Indicates from theory The total number of projection frames traced back from the scan position. Indicates the detector frame rate. Indicates redundant angles. This indicates the scan time per revolution of the turntable. This further ensures that the reference sequence necessarily and completely covers the actual sequence. The projection frames corresponding to the closure points provide a sufficient candidate range for subsequent similarity matching.

[0113] The fifth step introduces a real-time motion perturbation weighting factor and improves the SSIM algorithm formula by referring to the reference projection sequence and the paired data stream, so as to obtain the improved SSIM algorithm formula.

[0114] This embodiment introduces a real-time motion disturbance weighting factor to improve the SSIM algorithm, which can solve the defect of the traditional SSIM algorithm in being unable to detect dynamic motion disturbances at sea. The embodiment calculates the real-time motion disturbance weighting factor for each frame based on the motion state vector in the paired data stream. First, the total translation disturbance value is obtained by summing the amplitudes of the three translation components of the current frame, and the total rotation disturbance value is obtained by summing the angular amplitudes of the three rotation components. Both types of disturbance values ​​are then normalized using a preset maximum permissible disturbance threshold. Finally, a weighted sum is obtained. Its value range is constrained by Within the range.

[0115] When the ship is in a stable state, both translational and rotational disturbances approach zero. The value automatically approaches 1, indicating that the image frame has high credibility; when the ship experiences severe rolling in high sea states, the disturbance value rises rapidly. Automatically reduce the risk of motion blur in the frame image, and subsequent matching weights will be adaptively suppressed.

[0116] Based on this, the traditional SSIM algorithm is modified to include real-time motion perturbation weighting factors. By introducing the formula, we obtain the improved SSIM algorithm formula with motion adaptive constraints.

[0117] The traditional SSIM algorithm is as follows: , in, This represents the structural similarity between image A and image B. , Let A and B represent the mean values ​​of image signals A and B, respectively. Represents the covariance of the image signal. , These represent small constants set to avoid the denominator being zero. , These represent the standard deviations of image signals A and B, respectively.

[0118] The improved SSIM algorithm formula in this embodiment satisfies the following relationship: , in, This represents the structural similarity between image A and image B. This represents the weighting factor for real-time motion perturbations. , Let A and B represent the mean values ​​of image signals A and B, respectively. Represents the covariance of the image signal. , These represent small constants set to avoid the denominator being zero. , These represent the standard deviations of image signals A and B, respectively.

[0119] In the above example, A represents the reference image, and B represents the image to be compared. , This avoids setting small constants with denominators of zero, and the range of values ​​is... The improved algorithm achieves dimensional self-consistency and global normalization, and can automatically adjust the evaluation weight of image similarity according to real-time sea conditions, making it fully adaptable to dynamic imaging scenarios at sea.

[0120] The sixth step is to extract the image to be compared from the temporal scan information and determine the reference image from the reference projection sequence.

[0121] The embodiment extracts the first image projection of the sequence from the temporal scan information. As a baseline reference image, this frame corresponds to the first valid projection of the cone-beam CT rotating gantry as it formally enters the uniform scanning phase, which helps to eliminate blurry and invalid frames during the mechanical start-up and stop phases. Simultaneously, all images to be compared are extracted from the reference projection sequence, covering all candidate projection frames within a preset redundancy angle, to ensure that the set of images to be compared necessarily includes the actual projection. The target projection frame corresponding to the position needs to be pre-selected using the motion state vector during actual operation, ensuring that the first frame after the turntable starts satisfies the condition that the rotational angular velocity deviation is less than the rated value. The projection corresponding to the time is marked as... This can prevent non-uniform speed frames during the startup phase from being mistakenly identified as reference frames.

[0122] The seventh step is to analyze the structural similarity between the image to be compared and the reference image based on the reference projection sequence and the improved SSIM algorithm formula.

[0123] This embodiment further calculates the improved structural similarity between the image to be compared and the reference image, using the benchmark reference image as the benchmark. Using a fixed reference, all images to be compared in the reference projection sequence are traversed, and the structural similarity between the two images is calculated frame by frame by substituting them into the improved SSIM algorithm formula. The motion perturbation weighting factor of the corresponding frame is called in real time during the calculation process. The system automatically performs weight suppression on the similarity results of highly perturbated frames to avoid mismatches caused by motion blur. In order to improve computational efficiency, this embodiment can pre-normalize and crop all projected images to retain effective pixels in the detector imaging area to participate in similarity calculation, further meeting the real-time processing requirements in maritime emergency rescue scenarios.

[0124] The eighth step, the data filtering module, filters the images to be compared in the time-series scan information based on structural similarity to obtain the target projection sequence.

[0125] In this embodiment, the data filtering module iterates through all the images to be compared. The calculation results are obtained. The projection frame corresponding to the maximum value is the precise projection frame. Closed position projection Then, a segment is captured from the reference frame. arrive All continuous projection data between the points are processed, and invalid frames with mechanical start / stop at the beginning and end of the sequence, as well as high-perturbation distortion frames with SSIM' values ​​below a preset threshold (set to 0.6 in this embodiment), resulting in a complete, continuous target projection sequence without obvious motion artifacts. This sequence is then input into the subsequent image reconstruction module for image reconstruction. The above sequence completely covers... The effective imaging angle, the number of projections and the geometric distribution are fully adapted to the completeness requirements of cone-beam CT reconstruction algorithms, providing a reliable guarantee for outputting clear diagnostic images under sea conditions from the data source.

[0126] S4. The image reconstruction module reconstructs images based on the target projection sequence to achieve stable imaging and auxiliary diagnosis of cone-beam CT on medical ships under sea conditions.

[0127] In this embodiment, the image reconstruction module is designed to achieve stable imaging and rapid assisted diagnosis for cone-beam CT on a medical ship under high sea states. The image reconstruction module can output three-dimensional volume data and simultaneously generate two-dimensional transverse images for clinical interpretation. The specific implementation details are as follows: In this embodiment, the image reconstruction module can receive target projection sequences, multi-source data, and frame-by-frame timestamp aligned six-degree-of-freedom motion state vectors in real time.

[0128] After receiving the data, the image reconstruction module performs a data verification process as follows: It verifies the time difference between the timestamps of all projected frames and the corresponding motion state vectors, ensuring that the time difference does not exceed [the specified value]. And remove abnormal frames with excessive time alignment errors; verify the total number of frames in the target projection sequence to cover. The completeness of the scanning angle is ensured to guarantee that the number of projections is not less than a preset threshold. The rationality of motion state vectors is verified by cross-validating data from multiple sources. If the motion disturbance amplitude of a frame exceeds the preset maximum allowable threshold, the frame is automatically marked to enter the subsequent key correction queue, thus preventing invalid data from entering the reconstruction process from the data source.

[0129] In one embodiment, the image reconstruction module converts the six-degree-of-freedom motion state vectors bound to each frame in the aforementioned target projection sequence into the cone-beam CT imaging coordinate system. The homogeneous rigid body transformation matrix has translational components that directly map the real-time displacement values ​​of the ship's pitch, sway, and heave, while rotational components correspond to the real-time angle values ​​of pitch, roll, and yaw. In this embodiment, all parameters are uniformly expressed in international standard units of millimeters and radians to avoid dimensional errors.

[0130] Based on the rigid body transformation matrix mentioned above, the three-dimensional spatial offset of the X-ray tube focus and detector panel relative to the rotation center of the test bed at the moment of projection acquisition can be derived in reverse. In this embodiment, a trilinear interpolation algorithm is used to perform sub-pixel level remapping of the pixel coordinates of the projected image, so as to accurately align the actual acquisition pose affected by the ship's motion to the theoretical ideal scanning trajectory under the steady-state scenario on land, thereby eliminating the projection geometric misalignment problem caused by the rigid body motion of the ship from the root.

[0131] In one embodiment, dynamic weight matching and 3D volume data reconstruction are performed. To break the assumption of absolute stillness in traditional scanning platforms, this embodiment uses frame-by-frame dynamic weight update logic to automatically match the back-projection weights of each frame's projection with the corrected real-time geometric pose. Specifically, for residual small-amplitude perturbations in sea conditions of level 3 and above, at least three rounds of low-constraint SART iterative optimization are introduced. Regularization constraints on the motion state vector are added during the iteration process to avoid artifacts introduced by excessive iteration. Without significantly increasing reconstruction time, slight motion artifacts are further suppressed, ultimately resulting in an output resolution of [resolution value missing]. Stable 3D CT volumetric data, with provisions for rapid response to meet the needs of maritime emergency rescue.

[0132] In one embodiment, image quality closed-loop verification and image generation are performed. Based on the output 3D CT volume data, the image reconstruction module can automatically perform adaptive adjustment of window width and window level, and, for emergency scenarios frequently occurring at sea such as fractures, foreign bodies in the body, and closed organ injuries, preset specific edge enhancement parameters to automatically enhance the contrast of the lesion area. Simultaneously, by calculating the modulation transfer function (MTF) of the reconstructed image, the spatial resolution of the image is verified to be no less than [a certain value]. The module calculates the percentage of motion artifacts remaining in the entire sequence. If the relevant indicators do not meet the standards, the module automatically triggers local iterative recalibration. Without manual intervention, it can output three-dimensional volume data that meets the quality requirements and two-dimensional transverse images directly sliced ​​from it.

[0133] The image reconstruction module in this embodiment further converts the verified 3D CT images and 2D transverse images into a medical format conforming to the DICOM 3.0 standard. It directly connects to the shipboard medical workstation and can automatically annotate and provide preliminary diagnostic prompts for suspicious areas after image output. This provides technical support for rapid diagnosis for ship doctors who lack experience in imaging diagnosis at sea, and further ensures the high-quality imaging and auxiliary diagnostic functions of cone-beam CT on maritime medical ships.

[0134] Please see Figure 2 In an optional embodiment, the present invention also provides a cone-beam CT adaptive stabilization and rapid diagnostic system for marine medical ships. This system includes a motion sensing module, a scan trajectory planning module, a data filtering module, and an image reconstruction module. These modules are interconnected, implementing the specific steps of the embodiments of the cone-beam CT adaptive stabilization and rapid diagnostic system for marine medical ships provided by the present invention. The cone-beam CT adaptive stabilization and rapid diagnostic system for marine medical ships of the present invention has a complete structure, is objectively stable, and enhances the overall applicability and practical application capabilities of the present invention.

[0135] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the present invention.

Claims

1. A cone-beam CT adaptive image stabilization and rapid diagnostic system for maritime medical ships, characterized in that, The system includes a motion sensing module, a scanning trajectory planning module, a data filtering module, and an image reconstruction module; The motion sensing module is designed and installed, and multi-source data is collected using the motion sensing module. A motion state vector is constructed based on the multi-source data. The scanning trajectory planning module receives the multi-source data and the motion state vector, and sets up a dual-mode scanning mechanism in combination with historical diagnostic information, and plans the adaptive scanning trajectory of cone-beam CT according to the dual-mode scanning mechanism. The scanning trajectory planning module obtains the temporal scanning information output by cone-beam CT based on the adaptive scanning trajectory, and the data filtering module pairs and filters the temporal scanning information and the motion state vector to obtain the target projection sequence. The image reconstruction module reconstructs the image based on the target projection sequence to achieve stable imaging and auxiliary diagnosis of cone-beam CT on a medical ship under sea conditions.

2. The adaptive image stabilization and rapid diagnostic system for cone-beam CT on a marine medical ship according to claim 1, characterized in that, The design and installation of the motion sensing module, and the collection of multi-source data using the motion sensing module, include: An inertial measurement unit and a laser displacement sensor are configured in the motion sensing module; The inertial measurement unit and the laser displacement sensor are aligned and calibrated to complete the design and installation of the motion sensing module. The inertial measurement unit in the motion sensing module is used to collect inertial-based hull attitude data; Laser ranging data is acquired through the laser displacement sensor in the motion sensing module; The motion sensing module performs spatiotemporal synchronization alignment processing on the ship attitude data and the laser ranging data to output multi-source data.

3. The adaptive image stabilization and rapid diagnostic system for cone-beam CT on a marine medical ship according to claim 2, characterized in that, The construction of the motion state vector based on the multi-source data includes: The motion sensing module sets the inertial component fusion weight coefficient and the laser component fusion weight coefficient based on the ship attitude data and laser ranging data in the multi-source data. The motion sensing module performs frequency domain complementary weighting and fusion processing on the multi-source data based on the inertial component fusion weighting coefficient and the laser component fusion weighting coefficient to construct a standardized six-degree-of-freedom motion state vector.

4. The adaptive image stabilization and rapid diagnostic system for cone-beam CT on a marine medical ship according to claim 1, characterized in that, The scanning trajectory planning module receives the multi-source data and the motion state vector, and sets up a dual-mode scanning mechanism in conjunction with historical diagnostic information, including: The scanning trajectory planning module receives historical diagnostic information; The scanning trajectory planning module classifies sea state levels; The scanning trajectory planning module analyzes the magnitude and rate of change of the translation component based on the multi-source data and the motion state vector; The scanning trajectory planning module sets up a dual-mode scanning mechanism based on the sea state level, the historical diagnostic information, the translation component amplitude, and the rate of change. The dual-mode scanning mechanism includes a motion-compensated scanning mechanism and a non-isocentric large FOV scanning mechanism.

5. The adaptive image stabilization and rapid diagnostic system for cone-beam CT on a marine medical ship according to claim 4, characterized in that, The scanning trajectory planning module sets up a dual-mode scanning mechanism based on the sea state level, the historical diagnostic information, the translation component amplitude, and the rate of change, including: When the sea state level is below level three, the dual-mode scanning mechanism adopts a motion compensation scanning mechanism. A spatial coordinate update model is established in the motion compensation scanning mechanism; The real-time spatial coordinate information of the X-ray tube focal point and the detector center is obtained by using the spatial coordinate update model. Based on the real-time spatial coordinate information, a relative position constant constraint function is constructed in the motion compensation scanning mechanism; The scanning trajectory planning module constrains the distance between the X-ray tube and the detector through the relative position constant constraint function to maintain the constant projection geometry.

6. The adaptive image stabilization and rapid diagnostic system for cone-beam CT on a marine medical ship according to claim 4, characterized in that, The scanning trajectory planning module sets up a dual-mode scanning mechanism based on the sea state level, the historical diagnostic information, the translation component amplitude, and the rate of change, including: When the sea state level is greater than or equal to level three, the dual-mode scanning mechanism adopts a non-isocentric large FOV scanning mechanism. In the non-isocentric large FOV scanning mechanism, a non-isocentric geometric constraint function is set; An effective imaging volume expansion model is constructed in the motion compensation scanning mechanism by combining the similar triangle imaging principle with the non-isocentric geometric constraint function. The scanning trajectory planning module uses the effective imaging volume expansion model to analyze the effective imaging range of the target area within the detector, so as to ensure the imaging integrity under sea conditions.

7. The adaptive image stabilization and rapid diagnostic system for cone-beam CT on a marine medical ship according to claim 4, characterized in that, The adaptive scanning trajectory for cone-beam CT planned based on the dual-mode scanning mechanism includes: The operation mechanism based on the dual-mode scanning mechanism sets the joint judgment conditions of the dual-mode scanning mechanism in the scanning trajectory planning module. The scanning trajectory planning module adaptively switches between the motion compensation scanning mechanism and the non-isocentric large FOV scanning mechanism based on the joint judgment conditions, and plans the adaptive scanning trajectory of cone-beam CT.

8. The adaptive image stabilization and rapid diagnostic system for cone-beam CT on a marine medical ship according to claim 1, characterized in that, The scanning trajectory planning module obtains the temporal scanning information output by cone-beam CT based on the adaptive scanning trajectory, including: The scanning trajectory planning module adjusts the rotating gantry turntable device of cone-beam CT based on the adaptive scanning trajectory. The scanning trajectory planning module acquires temporal scanning information output by controlling the X-ray tube and detector in the rotating gantry turntable device.

9. The adaptive image stabilization and rapid diagnostic system for cone-beam CT on a marine medical ship according to claim 1, characterized in that, The data filtering module pairs and filters the temporal scan information and the motion state vector to obtain the target projection sequence, including: The data filtering module determines the acquisition time of each frame projection from the time-series scan information; The data filtering module performs data pairing based on the acquisition time and the motion state vector to obtain the motion state vector corresponding to the projection acquisition time of each frame. By combining the motion state vector corresponding to each frame of projection acquisition time, a paired data stream of projection and motion is obtained; A reference projection sequence is obtained based on the redundant angles and the paired data stream of the projection and motion; A real-time motion perturbation weighting factor is introduced, and the SSIM algorithm formula is improved by referring to the reference projection sequence and the paired data stream to obtain the improved SSIM algorithm formula. The image to be compared is extracted from the temporal scan information, and the reference image is determined from the reference projection sequence; The structural similarity between the image to be compared and the reference image is analyzed based on the reference projection sequence and the improved SSIM algorithm formula. The data filtering module filters the images to be compared in the time-series scanning information based on the structural similarity to obtain the target projection sequence.

10. The adaptive image stabilization and rapid diagnostic system for cone-beam CT on a marine medical ship according to claim 1, characterized in that, The image reconstruction module reconstructs the image based on the target projection sequence to achieve stable imaging and auxiliary diagnosis of cone-beam CT on a medical ship under sea conditions, including: The image reconstruction module receives the target projection sequence, the multi-source data, and the motion state vector; The image reconstruction module reconstructs the CT image based on the target projection sequence, the multi-source data, and the motion state vector to obtain the reconstructed CT image. Based on the reconstructed CT images, stable imaging and auxiliary diagnosis of cone-beam CT on a medical ship under sea conditions are achieved.

Images