A method and system for processing resolution imaging data of a gamma camera

Abstract

The application relates to a gamma camera resolution imaging data processing method and system, and relates to the technical field of image data processing.The application comprises the following steps: configuring image acquisition parameters of a target gamma camera; acquiring multiple initial acquisition images by performing image acquisition on a target acquisition region by the target gamma camera, and acquiring the spatial geometric relationship of the multiple initial acquisition images; improving the image resolution of the multiple initial acquisition images based on the image acquisition parameters, and acquiring multiple enhanced images; and performing spatial registration on the multiple enhanced images according to the multiple enhanced images and the spatial geometric relationship, and obtaining a spatially registered image.The application solves the problems of low resolution and insufficient multi-image registration accuracy in traditional gamma camera imaging, and realizes accurate processing of high-resolution imaging data through parameter customization configuration, multi-source image enhancement and spatial geometric registration, and improves the definition and spatial consistency of gamma camera imaging.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of image data processing, in particular to a gamma camera resolution imaging data processing method and system. BACKGROUND

[0002] With the development of image data processing technology, improving the imaging resolution of the gamma camera and the spatial registration accuracy of multiple images has become the key to optimizing data processing efficiency. At present, the traditional gamma camera imaging data processing mostly uses fixed parameter acquisition and basic image enhancement mode, which has the problems of insufficient resolution and low spatial registration accuracy, and it is difficult to meet the needs of high-definition imaging and accurate three-dimensional reconstruction in the fields of industrial detection and scientific imaging.

[0003] The existing processing method only relies on uniform parameters for image acquisition and resolution improvement, which leads to insufficient multi-view image fusion accuracy in complex scenes, and cannot adapt to the differentiated needs of different imaging scenes, nor can it meet the requirements of high resolution and spatial consistency in the intelligent development of image data processing. SUMMARY

[0004] In order to solve the above technical problems, the present application provides a gamma camera resolution imaging data processing method and system, which improves the imaging resolution of the gamma camera and the spatial registration accuracy of multiple images, and solves the problems of insufficient resolution and low spatial registration accuracy caused by fixed parameter acquisition in traditional processing, and insufficient multi-view image fusion accuracy in complex scenes.

[0005] The embodiments of the present application disclose the following technical solutions:

[0006] In a first aspect, the embodiments of the present application provide a gamma camera resolution imaging data processing method, which comprises:

[0007] configuring image acquisition parameters of a target gamma camera;

[0008] acquiring multiple initial acquisition images by the target gamma camera for image acquisition of a target acquisition area, and acquiring a spatial geometric relationship of the multiple initial acquisition images;

[0009] improving the image resolution of the multiple initial acquisition images based on the image acquisition parameters to obtain multiple enhanced images;

[0010] performing spatial registration on the multiple enhanced images according to the multiple enhanced images and the spatial geometric relationship to obtain spatially registered images.

[0011] In a second aspect, the embodiments of the present application provide a gamma camera resolution imaging data processing system, which comprises:

[0012] A parameter configuration module is configured to configure image acquisition parameters of a target gamma camera.

[0013] A multi-view acquisition module is configured to acquire a plurality of initial acquisition images by performing image acquisition on a target acquisition region by using the target gamma camera, and to acquire a spatial geometric relationship of the plurality of initial acquisition images.

[0014] An image enhancement module is configured to perform image resolution enhancement on the plurality of initial acquisition images based on the image acquisition parameters, and to acquire a plurality of enhanced images.

[0015] A spatial registration module is configured to perform spatial registration on the plurality of enhanced images based on the plurality of enhanced images and the spatial geometric relationship, and to obtain a spatially registered image.

[0016] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0017] The present application provides a gamma camera resolution imaging data processing method and system, which configures image acquisition parameters of a target gamma camera, performs multi-view image acquisition on a target region and acquires a spatial geometric relationship, enhances image resolution based on parameters, and finally realizes spatial registration of enhanced images. The method effectively solves the problems of low resolution and poor multi-view registration accuracy in traditional imaging by constructing a three-dimensional voxel grid, determining a registration position and image mapping. At the same time, through the steps of "parameter configuration-image acquisition-resolution enhancement-spatial registration", the adaptability of imaging to different scenes is significantly improved, and the drawbacks of poor imaging effect are overcome, providing a reliable data processing scheme for related fields.

[0018] The technical solutions of the present application realize the transformation of gamma camera imaging from low resolution to high resolution and from single view to multi-view accurate registration through the synergistic effect of multi-parameter collaborative configuration, image enhancer construction and spatial geometric registration, avoid imaging deviation caused by parameter fixation and single image enhancement mode, and improve the scientificity and accuracy of gamma camera imaging. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0020] Figure 1 A flowchart of a gamma camera resolution imaging data processing method provided by an embodiment of the present application is shown in the figure.

[0021] Figure 2A schematic diagram of the structure of a gamma camera resolution imaging data processing system provided in this application embodiment;

[0022] The components represented by each number in the attached diagram are explained below:

[0023] Parameter configuration module 01, multi-view acquisition module 02, image enhancement module 03, spatial registration module 04. Detailed Implementation

[0024] This application provides a method and system for processing gamma camera resolution imaging data, which solves the technical problems of low imaging resolution of gamma cameras, insufficient spatial registration accuracy of multiple images, poor adaptability to complex scenes under fixed parameter acquisition mode, and insufficient accuracy of multi-view image fusion in the prior art.

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0027] In the description of this application, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0028] Example 1, as shown in the appendix Figure 1 As shown, this application provides a method for processing gamma camera resolution imaging data, the method comprising the following steps:

[0029] S100: configuring an image acquisition parameter of a target gamma camera;

[0030] In the embodiments of the present application, in the process of gamma camera resolution imaging data processing, in order to realize high-resolution imaging of the target region and accurate multi-view image registration, a multi-dimensional parameter system covering energy screening, time control and spatial positioning needs to be configured according to the imaging requirements.

[0031] Specifically, the energy window parameters are set according to the characteristic energy of the target radioactive material, including the center energy value, the window width, and the multi-window division of the main window and the scatter window, the image contrast is improved and the scattering interference is suppressed by using the energy resolution characteristics of the detector.

[0032] Further, the acquisition time parameters are determined in combination with the estimated activity level of the target region and the imaging quality requirements, the exposure time and the dynamic imaging frame rate of a single view are set, and the detector saturation is prevented by using the count rate limit, and the image count statistics and detection efficiency are balanced according to the static / dynamic acquisition mode.

[0033] Subsequently, the spatial positioning parameters are planned according to the spatial size of the target region, the three-dimensional coordinates, the pitch / yaw / roll angles of the camera are set, the acquisition distance and the multi-angle acquisition track are determined, the coordinate system is constructed by positioning the reference point to align the basis, and the three-dimensional spatial geometry basis is provided for subsequent image registration.

[0034] The process forms a parameter combination scheme for different imaging scenarios through the cooperative configuration logic of the energy window, acquisition time and spatial positioning parameters, and lays a parameter foundation for high-resolution imaging data processing.

[0035] The step S100 in the method provided by the embodiments of the present application includes:

[0036] The energy window parameters are set according to the characteristic energy of the target radionuclide;

[0037] The acquisition time parameters are determined based on the estimated activity level of the target acquisition region and the image quality requirements;

[0038] The spatial positioning parameters are planned according to the spatial size of the target acquisition region;

[0039] The energy window parameters, the acquisition time parameters and the spatial positioning parameters are summarized to form the image acquisition parameter.

[0040] In the embodiments of the present application, in the process of gamma camera resolution imaging data processing, in order to realize high-resolution imaging of the target region and accurate multi-view image registration, three types of parameters (energy window parameters, acquisition time parameters, spatial positioning parameters) need to be configured cooperatively.

[0041] Specifically, the setting of the energy window parameter needs to be based on the characteristic energy characteristics of the target radionuclide. Taking Cs-137 as an example, its characteristic peak energy is 662 keV, and accordingly the energy window center is set to 662 keV, and the window width is set in the range of ± 15%. The main signal window and the scattering window are divided by multi-window, and the gamma rays of a specific energy are screened by using the energy resolution of the detector, so as to improve the image contrast and suppress the scattering interference.

[0042] Further, the determination of the acquisition time parameter needs to take into account the estimated activity of the target region and the imaging quality. If the activity of the target region is low (such as in a sparse distribution of radioactive substances scene), in order to ensure sufficient statistical count, the single view acquisition time needs to be extended to 120 seconds; if the activity is high, it can be shortened to 60 seconds to balance the detection efficiency.

[0043] At the same time, the frame rate is set according to the dynamic imaging requirements (such as 5 frames per second), and the count rate limit (such as 10000 counts per second) is used to prevent the saturation of the detector.

[0044] Further, the planning of the spatial positioning parameter needs to be based on the spatial size of the target region. Taking a circular detection region with a diameter of 3 meters as an example, the camera is set at a distance of 2 meters from the target, and through the attitude of three-dimensional coordinates (X=0, Y=0, Z=2), pitch angle 45°, yaw angle 0°, 12 acquisition angle motion trajectories are planned at an interval of 30°, and the center of the region is taken as the reference point for positioning, so as to build a three-dimensional spatial coordinate system basis for subsequent multi-view image registration.

[0045] Further, the three types of parameters are integrated through a structured data format. For example, a parameter set containing "energy window (662 keV ± 15%), acquisition time (60 seconds per view), spatial positioning (distance 2 meters, 12 angles at an interval of 30°)" is generated, forming a standardized image acquisition parameter configuration scheme, which provides a quantitative basis for subsequent image acquisition and resolution improvement.

[0046] This process realizes the collaborative optimization of the parameter system through the collaborative configuration of the nuclide characteristics, activity evaluation and spatial size, and ensures the accuracy and consistency of the imaging data.

[0047] S200: acquiring multiple initial acquisition images by image acquisition of a target acquisition region by the target gamma camera, and acquiring the spatial geometric relationship of the multiple initial acquisition images;

[0048] In the embodiment of the present application, in the process of gamma camera resolution imaging data processing, in order to realize accurate registration and three-dimensional reconstruction of multi-view images, image acquisition and spatial parameter acquisition need to be carried out based on spatial positioning parameters.

[0049] Specifically, first, the multiple acquisition positions of the target gamma camera are planned according to the spatial positioning parameters, and each position corresponds to at least one acquisition angle.

[0050] Further, a first acquisition position is selected from the plurality of acquisition positions, and a corresponding first acquisition angle is obtained. A parameter instruction is sent through the camera control interface to trigger image acquisition at the position and angle to obtain a first initial acquisition image.

[0051] At the same time, camera built-in sensor data is read in real time, and camera position coordinates, attitude angles, and acquisition distances of the image are recorded, and the space parameters of the first initial acquisition image are constructed based on the obtained data.

[0052] Further, image acquisition operations are sequentially performed at the remaining acquisition positions and corresponding angles according to a preset acquisition trajectory. The acquisition of all point positions is completed through loop iteration to form a data set containing a plurality of initial acquisition images and corresponding space parameters.

[0053] Finally, based on the space parameters of each initial acquisition image, the rotation and translation relationship between adjacent images is calculated through a three-dimensional coordinate transformation matrix. The coordinate and attitude angle difference in the two sets of space parameters are used to construct a rotation matrix (describing attitude change) and a translation vector (describing position change), so as to determine the spatial geometric transformation relationship between the plurality of initial acquisition images.

[0054] This process constructs a data set containing image data and geometric spatial relationship through multi-view ordered acquisition and accurate recording of space parameters, provides multi-dimensional input for subsequent resolution enhancement and spatial registration, and ensures the spatial consistency and reconstructability of imaging data.

[0055] The method provided in the embodiment of the application comprises the following steps S200:

[0056] According to the spatial positioning parameters, a plurality of acquisition positions of the target gamma camera are determined, and each acquisition position has at least one acquisition angle;

[0057] A first acquisition position is obtained from the plurality of acquisition positions, and a first acquisition angle is obtained in the corresponding acquisition angle of the first acquisition position;

[0058] In the first acquisition position and the first acquisition angle, the target gamma camera is controlled to perform image acquisition to obtain a first initial acquisition image;

[0059] Camera position coordinates, camera attitude angles, and acquisition distances of the first initial acquisition image are obtained, and the space parameters of the first initial acquisition image are determined based on the camera position coordinates, the camera attitude angles, and the acquisition distances;

[0060] Image acquisition is sequentially performed at the remaining acquisition positions and corresponding acquisition angles to obtain a plurality of initial acquisition images and corresponding space parameters;

[0061] determine a spatial geometric transformation relationship between the plurality of initial acquisition images based on spatial parameters of each initial acquisition image.

[0062] In the embodiments of the present application, in the process of multi-view imaging data processing of the gamma camera, in order to realize accurate registration of multi-view images and conversion of geometric space, image acquisition and spatial parameter acquisition need to be performed according to spatial positioning parameters.

[0063] Specifically, first, the plurality of acquisition positions of the camera are determined according to the three-dimensional coordinates, the acquisition trajectory and the view planning in the spatial positioning parameters.

[0064] Exemplarily, taking detection of a circular target region with a diameter of 3 meters as an example, the camera is arranged at a position 2 meters away from the target, a motion trajectory of 12 acquisition angle intervals is planned through a three-dimensional coordinate (X=0, Y=0, Z=2), a pitch angle of 45° and a yaw angle of 0°, thereby forming 12 sets of acquisition point position combinations.

[0065] Among them, the three-dimensional coordinates are fused and positioned by a laser radar and an inertial navigation system, the accuracy can reach ±0.03 meters, and the angle control error is not more than ±0.1°.

[0066] Further, a first acquisition position is selected from the plurality of acquisition point positions, such as a position with a three-dimensional coordinate (2, 0, 0) and a yaw angle of 0°, and a corresponding first acquisition angle (a pitch angle of 45°) is acquired. An acquisition instruction is sent through a camera control interface to trigger the gamma camera to perform image acquisition at the position and the angle.

[0067] Among them, in the acquisition process, the camera built-in sensor data is read synchronously, that is, the position coordinates (2.002, 0.001, 0.003) meters are acquired by using a laser range finder and a GPS, the attitude angle (a pitch angle of 44.98°, a yaw angle of 0.02° and a roll angle of 0.01°) is acquired by an inertial measurement system (IMU), the actual acquisition distance is 2.001 meters, and then the spatial parameters (including the coordinate position (X, Y, Z), the pitch angle, the yaw angle and the roll angle) of the first initial acquisition image are determined.

[0068] Further, image acquisition is sequentially performed at the remaining 11 sets of acquisition positions and corresponding angles according to the preset acquisition trajectory.

[0069] Exemplarily, from the yaw angle of 0° to the yaw angle of 30°, the acquisition distance of 2 meters is kept unchanged, the precise displacement and angle adjustment are realized by a motor-driven platform, the positioning error is controlled within ±0.01 meters, and the image with a pitch angle of 45° is acquired at each position, and the repeated positioning accuracy is ±0.05°.

[0070] Further, after each group of image acquisition, the spatial parameters including position coordinates, attitude angles, and acquisition distances are read and stored in real time through the API interface.

[0071] If the deviation of a certain group of parameters exceeds the threshold value (e.g., the acquisition distance deviation is greater than 0.05 meters or the attitude angle deviation is greater than 0.2°), repeated acquisition is automatically triggered to ensure the reliability of all image spatial parameters.

[0072] Finally, based on the obtained 12 groups of initial acquisition images and their spatial parameters, the spatial geometric transformation relationship between the images can be calculated by constructing a rotation matrix and a translation vector.

[0073] The rotation matrix is used to describe the attitude change of the image, and the translation vector is used to represent the position change.

[0074] Specifically, the translation vector is directly calculated from the difference between the camera position coordinates of two images, reflecting the displacement of the camera in three-dimensional space; the rotation matrix is constructed according to the difference in attitude angle, and the matrix elements are calculated through trigonometric functions to realize the quantization of the rotation attitude. The combined transformation matrix formed by the two can accurately map the spatial point coordinates in one image to the corresponding position in another image, providing a geometric transformation basis for the spatial registration of multi-view images.

[0075] Taking two groups of images with yaw angles of 0° and 30° as an example, after Kalman filtering denoising of the spatial parameters, the standardized parameters are obtained, i.e., the camera position coordinates of the image with yaw angle of 0° are (2.00, 0.00, 0.00) meters, and the attitude angles are pitch angle 45°, yaw angle 0°, and roll angle 0°; the camera position coordinates of the image with yaw angle of 30° are (1.73, 1.00, 0.00) meters, and the attitude angles are pitch angle 45°, yaw angle 30°, and roll angle 0°.

[0076] Further, the translation relationship is determined by the difference between the camera position coordinates of the two images. That is, the X-axis difference is -0.27 meters, the Y-axis difference is 1.00 meters, and the Z-axis difference is 0 meters, so the translation vector is 0.27 meters in the X-Y plane to the negative direction of the X-axis, 1.00 meters to the positive direction of the Y-axis, and the Z-axis position remains unchanged.

[0077] Among them, since the two images only have a 30° difference in yaw angle, the rotation matrix is a matrix that rotates 30° clockwise around the Z-axis, with the first row values being 0.866, -0.5, and 0, the second row values being 0.5, 0.866, and 0, and the third row values being 0, 0, and 1. This matrix represents the rotation transformation of the image around the Z-axis.

[0078] Further, the spatial geometric transformation relationship is constructed by a translation vector and a rotation matrix. Taking the coordinate (2.00, 0.00, 0.00) of a feature point in the image with a yaw angle of 0° as an example, the coordinate (1.462, 2.00, 0.00) is obtained after rotation transformation and superposition of the translation amount, which is consistent with the coordinate trend of the image with a yaw angle of 30°.

[0079] Further, the edge points (such as pixel coordinates (512, 512)) of the same name radioactive source in the two images are selected for verification. After coordinate transformation, the pixel deviation of the projection to the image with a yaw angle of 30° is less than 1 pixel, which meets the registration accuracy requirement under the resolution of 1024*1024.

[0080] The spatial geometric transformation relationship of the 12 groups of images constructs a multi-angle mapping network in the three-dimensional space, provides accurate geometric mapping basis for subsequent registration of the enhanced images to a unified three-dimensional voxel grid, ensures the consistency of the multi-view images in space, and thus improves the resolution of the final imaging and the accuracy of the three-dimensional geometric construction.

[0081] S300 performs image resolution enhancement on the plurality of initial acquisition images based on the image acquisition parameters to obtain a plurality of enhanced images;

[0082] In the embodiments of the present application, in the imaging data processing of the gamma camera, in order to compensate for the resolution limitation of the initial acquisition images caused by insufficient radioactive counting, noise interference, etc., the image details need to be intelligently optimized according to the image acquisition parameters combined with the machine learning model to obtain enhanced images with higher resolution, and to provide high-quality data basis for subsequent spatial registration.

[0083] Specifically, first, an image enhancer is constructed combined with the image acquisition parameters. This process constructs image acquisition constraints based on energy window parameters, acquisition time parameters and spatial positioning parameters.

[0084] Further, the retrieved sample images are labeled for image enhancement, and sample enhanced images are generated through noise suppression, edge sharpening, etc., forming a training pair of "initial image-enhanced image". The image enhancer is trained according to the plurality of sample acquisition images and the sample enhanced images.

[0085] Further, a first initial acquisition image (such as an image with a yaw angle of 0°) is extracted from the plurality of initial acquisition images. The first initial acquisition image is input into the image enhancer to obtain a first enhanced image, which is added to the plurality of enhanced images.

[0086] Subsequently, the remaining initial acquisition images are sequentially processed for resolution enhancement according to the above step procedure to obtain a plurality of enhanced images.

[0087] The process realizes the improvement of image resolution by the combination of image acquisition parameters and a machine learning model, provides support for subsequent image registration based on spatial geometric relationship, and ensures the accuracy during multi-view image fusion.

[0088] The step S300 in the method provided in the embodiment of the application comprises:

[0089] constructing an image enhancer for the image acquisition parameters;

[0090] extracting a first initial acquisition image from the plurality of initial acquisition images;

[0091] inputting the first initial acquisition image into the image enhancer to obtain a first enhanced image;

[0092] adding the first enhanced image to the plurality of enhanced images.

[0093] In the embodiment of the application, in the gamma camera resolution imaging data processing process, in order to improve the resolution of the initial acquisition image, an image enhancer needs to be constructed based on the image acquisition parameters to solve the problem of image detail blur caused by insufficient radioactive counting, noise interference and the like, and to provide a high-definition data basis for subsequent spatial registration.

[0094] The step of "constructing an image enhancer for the image acquisition parameters" in the method provided in the embodiment of the application comprises:

[0095] constructing an image acquisition constraint based on the image acquisition parameters;

[0096] performing image retrieval in a historical gamma camera acquisition image database according to the image acquisition constraint to obtain a plurality of sample acquisition images;

[0097] performing image enhancement annotation on the plurality of sample acquisition images to obtain a plurality of sample enhanced images;

[0098] training the image enhancer according to the plurality of sample acquisition images and the plurality of sample enhanced images.

[0099] In the embodiment of the application, in the gamma camera resolution imaging data processing process, in order to realize effective enhancement of the low-resolution image, a special image enhancer needs to be constructed according to the image acquisition parameters, and the model needs to be optimized in combination with historical data.

[0100] Specifically, first, the image acquisition constraints are constructed based on the image acquisition parameters. For example, when the image acquisition parameters are an energy window of 662 keV ± 15%, a single-view acquisition time of 60 seconds, and a camera distance to the target of 2 meters, the parameters need to be converted into threshold conditions for database retrieval: the energy window range is limited to 662 keV ± 20% (tolerance reserved), the acquisition time interval is set to 50-70 seconds, and the distance in the spatial positioning parameter is set to 1.8-2.2 meters.

[0101] The constraint conditions are formed by quantifying the imaging physical parameters to form a screening rule for historical samples, ensuring that the retrieved samples are consistent with the current imaging conditions, and providing effective data support for subsequent training.

[0102] Further, the historical gamma camera acquisition image database is retrieved according to the image acquisition constraints.

[0103] The database stores gamma images under different nuclides (such as Cs-137, Co-60) and different acquisition conditions, and by matching the energy window, acquisition time, and spatial positioning parameter, sample acquisition images with a matching degree higher than 80% are screened out.

[0104] For example, for the constraint of the 662 keV energy window, images using the Cs-137 nuclide and with the energy window set in the range of 662 keV ± 20% are preferentially retrieved, while samples with too short acquisition time (< 30 seconds) or too far distance (> 3 meters) are filtered out, and finally an initial sample set of multiple groups of images is formed.

[0105] Further, the multiple sample acquisition images are image-enhanced and labeled. The labeling process combines the gamma imaging physical characteristics and the resolution enhancement requirements, and through professional image processing software (such as OsiriX), the sample images are subjected to noise suppression, edge sharpening, and radioactive signal normalization processing.

[0106] For example, for images with insufficient count statistics due to short acquisition time (such as 40 seconds), a maximum likelihood denoising algorithm is used to reduce noise; for images with blurred edges due to a long distance between the camera and the target (such as 2.2 meters), a sharpening algorithm is used to enhance the outline.

[0107] The labeled sample enhanced images need to retain the radioactive distribution characteristics of the original images, while the resolution is increased by more than 30%, forming a training pair corresponding to the sample acquisition image.

[0108] Finally, an image enhancer is trained and generated according to the sample acquisition image and the sample enhanced image. A convolutional neural network (such as U-Net) is used as the basic architecture, and the image acquisition parameters (such as the energy window center value, the acquisition time, and the camera distance) are used as the conditional input model, so that the model can perceive the image enhancement rules under different parameters.

[0109] Specifically, the training data set is split into a training set and a validation set in a ratio of 8:2, supervised training is performed using the batch gradient descent method, 32 groups of images are selected each time to iteratively optimize the model parameters, and the mapping relationship from low resolution images to high resolution images is learned by minimizing the loss function of the peak signal-to-noise ratio and the structural similarity index.

[0110] Further, after completing 10 training cycles, the model performance is evaluated using the validation set. If the peak signal-to-noise ratio on the validation set shows an insufficient improvement of less than 0.5 dB for 5 consecutive evaluations, the learning rate decay mechanism is triggered, and the learning rate is reduced to 50% of the current value, making the parameter adjustment more precise.

[0111] To avoid overfitting, an early stopping mechanism is introduced, i.e., when the structural similarity index of the validation set does not improve for 15 consecutive cycles, the training is immediately terminated and the optimal model parameters are saved.

[0112] After about 200 training cycles of iteration, the peak signal-to-noise ratio of the model on the validation set is improved by more than 15 dB, the radioactive hotspot positioning error is less than 1 pixel, and the structural similarity index of the predicted enhanced image and the sample enhanced image is greater than 0.9, indicating that the image enhancer training is complete and has the ability to improve the resolution of gamma images under different acquisition parameters.

[0113] Finally, the target initial acquisition image and its corresponding image acquisition parameters are input into the trained enhancer, and based on the parameter-driven image enhancement model, a high-resolution enhanced image is output, providing high-quality data support for subsequent spatial registration.

[0114] For example, input a set of initial acquisition images with an energy window of 662 keV ± 15%, an acquisition time of 40 seconds, and a camera distance of 2.2 meters from the target. After processing by the constructed image enhancer, the output enhanced image has significantly reduced noise, clearer edge profile, more accurate radioactive hotspot positioning, and resolution improved by more than 30%, meeting the requirements of subsequent spatial registration.

[0115] Further, after the image enhancer is constructed, a first initial acquisition image is extracted from a plurality of initial acquisition images.

[0116] Specifically, first, a first initial acquisition image is extracted from a plurality of initial acquisition images. That is, according to the acquisition trajectory planned by the spatial positioning parameters, a standard view image with a yaw angle of 0° and a pitch angle of 45° is usually selected as the reference.

[0117] Exemplarily, when 12-view collection is performed on a circular target area with a diameter of 3 meters, an initial image corresponding to a first collection position with a coordinate of (2.00, 0.00, 0.00) meters and a yaw angle of 0° is preferentially extracted, and the image is suitable as a reference sample for subsequent enhancement processing due to a positive view angle and a low noise level.

[0118] Further, the first initial collection image is input into an image enhancer to obtain a first enhanced image.

[0119] Specifically, the image enhancer automatically invokes pre-trained model parameters for targeted processing (such as noise suppression, edge sharpening, and radioactive signal normalization) according to the collection parameters (such as an energy window of 662 keV ± 15%, a collection time of 60 seconds, and a camera distance of 2 meters) of the image, and finally outputs a first enhanced image with a resolution improved by more than 30%.

[0120] Further, the first enhanced image is added to an enhanced image set according to a preset index rule. When stored, the index position is determined according to spatial parameters such as the collection position and angle, and the enhanced image together with other enhanced images forms a complete data set, providing a high-quality data basis for subsequent spatial registration.

[0121] Exemplarily, if the first enhanced image is collected at a yaw angle of 0° and a pitch angle of 45°, it is stored in the first index unit of the data set according to the spatial positioning parameters, and is associated with parameters such as the energy window and the collection time. In this way, the image is integrated with other enhanced images in order, forming a complete data set.

[0122] The data set not only retains the radioactive distribution characteristics, but also improves the detail clarity by more than 30% through unified processing, thereby providing high-quality standardized data for subsequent three-dimensional voxel grid registration and ensuring the spatial consistency and geometric mapping accuracy of multi-view images.

[0123] S400: Spatially registering the plurality of enhanced images according to the plurality of enhanced images and the spatial geometric relationship to obtain a spatially registered image.

[0124] In the embodiments of the present application, in the process of gamma camera imaging data processing, in order to realize three-dimensional spatial fusion of multi-view enhanced images, a three-dimensional voxel grid needs to be constructed based on the spatial geometric relationship, and precise mapping of the images is realized through registration position calculation, so as to obtain a spatially consistent registered image.

[0125] Specifically, first, a three-dimensional voxel grid is established, the target collection area is divided into a plurality of voxel units, and a target area grid model is obtained. The grid model takes voxels as the basic unit, spatially discretizes the target area, and ensures that the spatial resolution matches the pixel accuracy of the enhanced image.

[0126] Further, the registration positions of each enhanced image are determined in the plurality of voxel units according to the spatial geometric relationship to obtain a plurality of registration positions. The spatial geometric relationship is derived from the spatial parameters (such as camera position coordinates, attitude angles) of each initial acquisition image, and reflects the transformation relationship between different view images through a rotation matrix and a translation vector.

[0127] Finally, the plurality of enhanced images are configured in the target region grid model according to the plurality of registration positions to obtain the spatially registered images. That is, each enhanced image is mapped into the three-dimensional voxel grid according to the registration position, and the pixel values of the overlapping regions are processed by taking an arithmetic mean value, so that the spatially registered images with spatial consistency and accurate radioactivity distribution of each view image are finally obtained.

[0128] This step realizes the spatial registration of the multi-view enhanced images through the construction of the three-dimensional voxel grid, the spatial geometric transformation calculation and the image mapping configuration, and ensures the accurate reconstruction of the radioactivity distribution in the three-dimensional space.

[0129] The step S400 in the method provided in the embodiments of the present application includes:

[0130] A three-dimensional voxel grid is established, the target acquisition region is divided into a plurality of voxel units, and a target region grid model is obtained;

[0131] The registration positions of each enhanced image are determined in the plurality of voxel units according to the spatial geometric relationship to obtain a plurality of registration positions;

[0132] The plurality of enhanced images are configured in the target region grid model according to the plurality of registration positions to obtain the spatially registered images.

[0133] In the process of gamma camera imaging data processing in the embodiments of the present application, in order to realize the spatial fusion of the multi-view enhanced images, a three-dimensional voxel grid model needs to be constructed and combined with the spatial geometric transformation relationship to realize the accurate registration of the images.

[0134] Specifically, first, a three-dimensional voxel grid is established, the target acquisition region is divided into a plurality of voxel units, and a target region grid model is obtained.

[0135] The three-dimensional voxel grid uses voxels as the basic unit of spatial discretization, for example, a cylindrical target region with a radius of 2 meters is divided into a 80x80x100 voxel matrix, and the spatial size of each voxel is 2.5cmx2.5cmx2cm, so as to ensure that the voxel resolution matches the pixel density of the enhanced image and avoid registration errors caused by spatial quantization errors.

[0136] Further, the registration positions of each enhanced image are determined in the plurality of voxel units according to the spatial geometric relationship to obtain a plurality of registration positions.

[0137] The spatial geometric relationship is calculated from the spatial parameters of each initial acquisition image, including constructing a rotation matrix through the camera position coordinates (X, Y, Z), the pitch angle, the yaw angle and the roll angle, determining a translation vector in combination with the acquisition distance, and thus establishing the transformation relationship of different view images to the three-dimensional voxel grid.

[0138] Exemplarily, for an enhanced image acquired at the coordinates (3, 0, 0) meters and the yaw angle 90°, the projection reference plane of the enhanced image in the voxel grid is determined by calculating the rotation matrix of rotating 90° around the Z axis and the translation vector (-3, 0, 0).

[0139] Finally, the plurality of enhanced images are configured in the target region grid model according to the plurality of registration positions, and a spatially registered image is obtained.

[0140] Specifically, each enhanced image is mapped into the three-dimensional voxel grid according to the transformation matrix, and for the pixel values of the overlapping region, an arithmetic mean value method is used for fusion processing, so as to obtain a spatially registered image.

[0141] Exemplarily, when a voxel unit is covered by two enhanced images at the same time, the arithmetic mean value of the corresponding pixel values of the two images is taken as the final value of the voxel, and if there are multiple images overlapping, the mean value is calculated according to the number of images participating in the fusion.

[0142] This processing method ensures the consistency of the radioactivity signals of the multi-view images in space, avoids the signal superposition error caused by the view angle difference, and finally forms a three-dimensional voxel image with spatial consistency and accurate radioactivity distribution of each view image.

[0143] This step realizes the spatial registration of the multi-view enhanced images by constructing a three-dimensional voxel grid, determining the spatial geometric registration position and image pixel fusion, not only retains the radioactivity distribution characteristics of each image, but also eliminates the view angle deviation through a unified spatial coordinate system, and has application value in the fields of nuclear medicine and industrial detection.

[0144] The above specific embodiments achieve the following technical effects:

[0145] The embodiment of the application provides a kind of gamma camera resolution imaging data processing method, first, the image acquisition parameter of target gamma camera is configured, including energy window parameter, acquisition time parameter and spatial positioning parameter, lay foundation for high-resolution imaging.Second, the image acquisition of target acquisition area is carried out by target gamma camera, obtains multiple initial acquisition images and its spatial geometric relationship, provides data support for subsequent processing.Then, based on the resolution of initial acquisition image is promoted based on image acquisition parameter, obtains multiple enhanced images, improves image quality.Finally, according to enhanced image and spatial geometric relationship, spatial registration is carried out, and spatial registration image is obtained, and the fusion of multi-view image is realized.The method is through the standardization process of multi-parameter collaborative configuration, multi-view acquisition, image enhancement and spatial registration, effectively reduce the interference of noise and other factors in imaging process;With the aid of machine learning algorithm, image enhancer is constructed, and the accuracy of image resolution promotion is improved;By three-dimensional voxel grid construction and spatial geometric transformation calculation, multi-view image is registered to unified spatial coordinate system, so that imaging result is more in line with actual scene.

[0146] The method provided by the embodiment of the application solves the problems of poor imaging effect and poor adaptability caused by low resolution and poor multi-view image registration accuracy in traditional gamma camera imaging. The efficiency and accuracy of imaging are effectively improved, and reliable data support is provided for related fields.

[0147] Embodiment two, as shown in the figure Figure 2 Based on the inventive concept of the gamma camera resolution imaging data processing method provided in embodiment one, the application further provides a gamma camera resolution imaging data processing system, which specifically comprises:

[0148] The parameter configuration module 01 is configured to configure the image acquisition parameters of the target gamma camera.

[0149] The multi-view acquisition module 02 is configured to acquire multiple initial acquisition images by acquiring images of the target acquisition area by the target gamma camera, and acquire the spatial geometric relationship of the multiple initial acquisition images.

[0150] The image enhancement module 03 is configured to enhance the image resolution of the multiple initial acquisition images based on the image acquisition parameters, and obtain multiple enhanced images.

[0151] The spatial registration module 04 is configured to register the multiple enhanced images based on the multiple enhanced images and the spatial geometric relationship, and obtain spatial registration images.

[0152] In one embodiment, the parameter configuration module 01 is further configured to:

[0153] The energy window parameter is set according to the characteristic energy of the target radionuclide.

[0154] determining the acquisition time parameter based on the estimated activity level of the target acquisition region and the image quality requirement;

[0155] planning the spatial positioning parameter according to the spatial size of the target acquisition region;

[0156] summarizing the energy window parameter, the acquisition time parameter and the spatial positioning parameter to form the image acquisition parameter.

[0157] In one embodiment, the multi-view acquisition module 02 is further configured to:

[0158] determining a plurality of acquisition positions of the target gamma camera according to the spatial positioning parameter, each of the acquisition positions having at least one acquisition angle;

[0159] acquiring a first acquisition position from the plurality of acquisition positions and a first acquisition angle corresponding to the first acquisition position;

[0160] controlling the target gamma camera to perform image acquisition at the first acquisition position and the first acquisition angle to obtain a first initial acquisition image;

[0161] obtaining camera position coordinates, camera attitude angle and acquisition distance of the first initial acquisition image, and determining a spatial parameter of the first initial acquisition image based on the camera position coordinates, the camera attitude angle and the acquisition distance;

[0162] sequentially performing image acquisition at the remaining acquisition positions and corresponding acquisition angles to obtain a plurality of initial acquisition images and corresponding spatial parameters;

[0163] determining a spatial geometric transformation relationship between the plurality of initial acquisition images based on the spatial parameters of the initial acquisition images.

[0164] In one embodiment, the image enhancement module 03 is further configured to:

[0165] constructing an image enhancer for the image acquisition parameter;

[0166] extracting a first initial acquisition image from the plurality of initial acquisition images;

[0167] inputting the first initial acquisition image into the image enhancer to obtain a first enhanced image;

[0168] adding the first enhanced image to the plurality of enhanced images.

[0169] In one embodiment, the spatial registration module 04 is further configured to:

[0170] A three-dimensional voxel grid is established, the target acquisition region is divided into a plurality of voxel units, and a target region grid model is acquired;

[0171] According to the spatial geometric relationship, a registration position of each of the enhanced images is determined in the plurality of voxel units, and a plurality of registration positions are acquired;

[0172] According to the plurality of registration positions, the plurality of enhanced images are configured on the target region grid model, and the spatial registration image is acquired.

[0173] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned describes a specific embodiment of the present application. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0174] The above only describes the preferred embodiments of the present application, and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0175] The specification and drawings of the present application are only exemplary descriptions of the present application, and are considered to cover any and all modifications, changes, combinations or equivalents within the scope of the present application. Obviously, those skilled in the art can make various modifications and changes to the present application without departing from the scope of the present application. Thus, if these modifications and changes of the present application belong to the scope of the present application and its equivalent technology, the present application intends to include these modifications and changes.

Claims

1. A method of processing gamma camera resolution imaging data, the method comprising: The method comprises: configuring image acquisition parameters of a target gamma camera; acquiring a plurality of initial acquisition images by image acquisition of a target acquisition region by the target gamma camera, and acquiring a spatial geometric relationship of the plurality of initial acquisition images; image resolution enhancement of the plurality of initial acquisition images based on the image acquisition parameters, to acquire a plurality of enhanced images; spatial registration of the plurality of enhanced images according to the plurality of enhanced images and the spatial geometric relationship, to obtain a spatially registered image; the image acquisition parameters comprise energy window parameters, acquisition time parameters, and spatial positioning parameters; acquiring a plurality of initial acquisition images by image acquisition of a target acquisition region by the target gamma camera, and acquiring a spatial geometric relationship of the plurality of initial acquisition images, comprising: determining a plurality of acquisition positions of the target gamma camera according to the spatial positioning parameters, each of the acquisition positions having at least one acquisition angle; acquiring a first acquisition position among the plurality of acquisition positions, and acquiring a first acquisition angle among the acquisition angles corresponding to the first acquisition position; controlling the target gamma camera to perform image acquisition at the first acquisition position and the first acquisition angle, to acquire a first initial acquisition image; acquiring camera position coordinates, camera attitude angles, and acquisition distances of the first initial acquisition image, and determining spatial parameters of the first initial acquisition image based on the camera position coordinates, the camera attitude angles, and the acquisition distances; sequentially performing image acquisition at the remaining acquisition positions and corresponding acquisition angles to acquire a plurality of initial acquisition images and corresponding spatial parameters; determining a spatial geometric transformation relationship between the plurality of initial acquisition images based on the spatial parameters of each initial acquisition image; image resolution enhancement of the plurality of initial acquisition images based on the image acquisition parameters, to acquire a plurality of enhanced images, comprising: constructing an image enhancer for the image acquisition parameters; extracting a first initial acquisition image from the plurality of initial acquisition images; inputting the first initial acquisition image into the image enhancer to acquire a first enhanced image; adding the first enhanced image to the plurality of enhanced images; constructing an image enhancer for the image acquisition parameters, comprising: constructing image acquisition constraints based on the image acquisition parameters; performing image retrieval in a historical gamma camera acquisition image database according to the image acquisition constraints, to acquire a plurality of sample acquisition images; performing image enhancement annotation on the plurality of sample acquisition images, to acquire a plurality of sample enhanced images; training the image enhancer according to the plurality of sample acquisition images and the plurality of sample enhanced images.

2. The method of claim 1, wherein, configuring image acquisition parameters of a target gamma camera, comprising: setting the energy window parameters according to the characteristic energy of the target radionuclide; determining the acquisition time parameters based on the estimated activity level and image quality requirements of the target acquisition region; planning the spatial positioning parameters according to the spatial dimensions of the target acquisition region; summarizing the energy window parameters, the acquisition time parameters, and the spatial positioning parameters to form the image acquisition parameters.

3. The method of claim 1, wherein, According to the plurality of enhanced images and the spatial geometric relationship, the plurality of enhanced images are spatially registered to obtain a spatially registered image, comprising: A three-dimensional voxel grid is established, the target acquisition region is divided into a plurality of voxel units, and a target region grid model is obtained; According to the spatial geometric relationship, the registration position of each enhanced image in the plurality of voxel units is determined to obtain a plurality of registration positions; According to the plurality of registration positions, the plurality of enhanced images are configured in the target region grid model to obtain the spatially registered image.

4. A gamma camera resolution imaging data processing system characterized by, The system is used to execute the gamma camera resolution imaging data processing method in any one of claims 1-3, and the system comprises: A parameter configuration module is configured to configure the image acquisition parameters of the target gamma camera; A multi-view acquisition module is configured to acquire a plurality of initial acquisition images by performing image acquisition on a target acquisition region through the target gamma camera, and to acquire a spatial geometric relationship of the plurality of initial acquisition images; An image enhancement module is configured to improve the image resolution of the plurality of initial acquisition images based on the image acquisition parameters to obtain a plurality of enhanced images; A spatial registration module is configured to spatially register the plurality of enhanced images according to the plurality of enhanced images and the spatial geometric relationship to obtain a spatially registered image.

Images