Flat panel detector positioning method, device, system, equipment and medium

Abstract

The embodiment of the invention discloses a flat panel detector positioning method, device, system and equipment and a medium, and the method comprises the steps: obtaining an image, collected by an image collection device, of a flat panel detector, and determining the pixel coordinates of a feature mark and a feature point in the image through a positioning model; determining first position attitude information and confidence coefficient according to the pixel coordinates, the internal reference matrix and the physical coordinates of the feature points; according to the confidence coefficient and a preset threshold value, the first position and attitude information is determined to serve as positioning information, or the second position and attitude information is determined to serve as the positioning information, and the second position and attitude information is determined according to the motion state information obtained by the inertial sensor. According to the technical scheme of the embodiment of the invention, the problems of large size and high cost of the flat panel detector positioned by the existing magnetic positioning device are solved, the positioning cost of the flat panel detector can be reduced and the positioning reliability and portability can be improved by combining the positioning mode of visual detection and the inertial sensor, and the method can adapt to different scenes.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of computer technology, and in particular to a method, device, system, equipment, and medium for positioning a flat panel detector. Background Art

[0002] In some portable X-ray imaging systems, the flat-panel detector is often placed behind the patient, preventing the radiologist from directly observing its position. Improper detector placement can cause the imaging range to shift, necessitating a retake of the X-ray. Traditional solutions use magnetic positioning devices to track the detector's position, but these systems are bulky, susceptible to interference from surrounding magnetic fields, and expensive. Summary of the Invention

[0003] The embodiments of the present invention provide a flat panel detector positioning method, apparatus, system, device, and medium, which can reduce the positioning cost of the flat panel detector, improve positioning reliability and portability, and can adapt to different scenarios.

[0004] In a first aspect, an embodiment of the present invention provides a method for positioning a flat panel detector, the method comprising:

[0005] An image acquisition device acquires a first image of the flat panel detector;

[0006] Inputting the first image into the flat panel detector positioning model to obtain a feature marker of the flat panel detector in the first image and pixel coordinates of a feature point associated with the feature marker;

[0007] Determine first position and posture information and a confidence level of the flat-panel detector at a first moment corresponding to the first image based on the pixel coordinates, the intrinsic parameter matrix of the image acquisition device, and the physical coordinates of the flat-panel detector coordinate system corresponding to the feature points;

[0008] When the confidence level is greater than a preset threshold, the first position posture information is used as positioning information of the flat panel detector;

[0009] When the confidence level is less than or equal to a preset threshold, using the second position and posture information as the positioning information of the flat-panel detector, wherein the second position and posture information is determined based on the motion state information of the flat-panel detector obtained by the inertial sensor and the position and posture information of the flat-panel detector at a second moment, where the second moment is a moment before the first moment;

[0010] Among them, the flat panel detector positioning model is a neural network model.

[0011] In a second aspect, an embodiment of the present invention provides a flat panel detector positioning device, the device comprising:

[0012] An image acquisition module, configured to acquire a first image captured by the flat panel detector by the image acquisition device;

[0013] a pixel coordinate determination module, configured to input the first image into a flat panel detector positioning model, and obtain a feature mark of the flat panel detector in the first image, and pixel coordinates of feature points associated with the feature mark;

[0014] A first position and posture information determination module is configured to determine first position and posture information and a confidence level of the flat-panel detector at a first moment corresponding to the first image based on pixel coordinates, an intrinsic parameter matrix of the image acquisition device, and physical coordinates of the flat-panel detector coordinate system corresponding to the feature points;

[0015] A first positioning information determining module, configured to use the first position posture information as positioning information of the flat panel detector when the confidence level is greater than a preset threshold;

[0016] a second positioning information determining module, configured to use second position and posture information as positioning information of the flat-panel detector when the confidence level is less than or equal to a preset threshold, wherein the second position and posture information is determined based on motion state information of the flat-panel detector acquired by the inertial sensor and position and posture information of the flat-panel detector at a second moment, the second moment being a moment before the first moment;

[0017] Among them, the flat panel detector positioning model is a neural network model.

[0018] In a third aspect, an embodiment of the present invention provides a flat panel detector positioning system, the system comprising:

[0019] Image acquisition device, inertial measurement unit and computing module;

[0020] Among them, the image acquisition device is used to acquire images of the flat-panel detector; the inertial measurement device is arranged on the flat-panel detector, and is used to acquire the motion state information of the flat-panel detector; the calculation module is used to determine the positioning information of the flat-panel detector according to the flat-panel detector positioning method provided by any embodiment of the present invention.

[0021] In a fourth aspect, an embodiment of the present invention further provides a computer device, the computer device comprising:

[0022] one or more processors;

[0023] a memory for storing one or more programs;

[0024] When the one or more programs are executed by one or more processors, the one or more processors implement the flat panel detector positioning method provided by any embodiment of the present invention.

[0025] In a fifth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the flat panel detector positioning method provided by any embodiment of the present invention.

[0026] In a sixth aspect, an embodiment of the present invention further provides a computer program product, including a computer program. When the computer program is executed by a processor, the flat panel detector positioning method provided by any embodiment of the present invention is implemented.

[0027] The embodiments of the above invention have the following advantages or beneficial effects:

[0028] In the embodiment of the present invention, the image acquisition device acquires a first image of the flat panel detector;

[0029] The first image is input into the flat-panel detector positioning model to obtain the feature mark of the flat-panel detector in the first image and the pixel coordinates of the feature points associated with the feature mark; based on the pixel coordinates, the intrinsic parameter matrix of the image acquisition device, and the physical coordinates of the flat-panel detector coordinate system corresponding to the feature points, the first position and posture information and confidence of the flat-panel detector at the first moment corresponding to the first image are determined; when the confidence is greater than a preset threshold, the first position and posture information is used as the positioning information of the flat-panel detector; when the confidence is less than or equal to the preset threshold, the second position and posture information is used as the positioning information of the flat-panel detector, wherein the second position and posture information is determined based on the motion state information of the flat-panel detector obtained by the inertial sensor and the position and posture information of the flat-panel detector at the second moment, the second moment being the moment before the first moment; wherein the flat-panel detector positioning model is a neural network model. The technical solution of the embodiment of the present invention solves the problem of large size and high cost of the current magnetic positioning device for positioning flat-panel detectors. By combining the positioning method of visual detection and inertial sensors, the positioning cost of the flat-panel detector can be reduced, the positioning reliability and portability can be improved, and it can adapt to different scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 is a flow chart of a flat panel detector positioning method provided by an embodiment of the present invention;

[0031] Figure 2 is a flow chart of a flat panel detector positioning method provided by an embodiment of the present invention;

[0032] Figure 3 This is a schematic diagram of characteristic markings of a flat panel detector provided by an embodiment of the present invention;

[0033] Figure 4 1 is a schematic structural diagram of a flat panel detector positioning device provided by an embodiment of the present invention;

[0034] Figure 5 1 is a schematic structural diagram of a flat panel detector positioning device provided by an embodiment of the present invention;

[0035] Figure 6 It is a structural diagram of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0036] The present invention will be further described in detail below with reference to the accompanying drawings and examples. It will be understood that the specific embodiments described herein are intended only to illustrate the present invention and are not intended to limit the present invention. It should also be noted that, for ease of description, the accompanying drawings only illustrate portions relevant to the present invention, not all structures.

[0037] Figure 1 This is a flowchart of a flat-panel detector positioning method provided in an embodiment of the present invention. This embodiment is applicable to scenarios involving flat-panel detector positioning. The method can be performed by a flat-panel detector positioning device, which can be implemented using software and / or hardware and integrated into a computer device with application development capabilities.

[0038] like Figure 1 As shown, the flat panel detector positioning method of this embodiment includes the following steps:

[0039] S110 , an image acquisition device acquires a first image of a flat panel detector.

[0040] The image acquisition device can be an RGB camera, which captures a color two-dimensional image in the video of the flat-panel detector during the positioning of the flat-panel detector. The first image can be an image of the flat-panel detector corresponding to a first moment, and the first moment can be any moment in the video.

[0041] S120 , inputting the first image into a flat panel detector positioning model to obtain characteristic marks of the flat panel detector in the first image and pixel coordinates of characteristic points associated with the characteristic marks.

[0042] The flat-panel detector positioning model is a neural network model, such as a neural network model based on a self-attention mechanism, that is, a neural network model based on a Transformer architecture. The neural network model is trained offline using a large amount of manually annotated sample flat-panel detector image data. Based on the network parameters obtained during the training phase, such as encoder weights, the model can perform image preprocessing, feature vector generation, and position encoding on the first image. The model extracts global feature relationships based on the Transformer encoder, focuses on the corresponding feature areas in the image through the attention mechanism to determine feature markers, and the decoder outputs the pixel coordinates of feature points associated with the feature markers. The flat-panel detector positioning model can also be a deep learning model such as a CNN (Convolutional Neural Network) or a YOLO (You Only Look Once, a single-stage target detection algorithm), which is not limited in this embodiment.

[0043] S130 , determining first position and posture information and a confidence level of the flat-panel detector at a first moment corresponding to the first image based on the pixel coordinates, the intrinsic parameter matrix of the image acquisition device, and the physical coordinates of the flat-panel detector coordinate system corresponding to the feature points.

[0044] Based on the pixel coordinates, the camera's intrinsic parameter matrix, and the physical coordinates of the feature points in the flat-panel detector coordinate system, a PnP (Perspective-n-Point) algorithm is used to determine the first position and attitude information and confidence level of the flat-panel detector at the first moment corresponding to the first image. The first position and attitude information is the position and attitude of the flat-panel detector relative to the camera head, and may include translation and rotation information.

[0045] S140: When the confidence level is greater than a preset threshold, use the first position posture information as positioning information of the flat panel detector.

[0046] The preset threshold can be set according to actual needs. For example, the preset threshold is 0.9. When the confidence level is greater than 0.9, the first position posture information meets the accuracy standard of the positioning information, and the first position posture information is used as the positioning information of the flat panel detector.

[0047] S150: When the confidence level is less than or equal to a preset threshold, use the second position posture information as positioning information of the flat panel detector.

[0048] The second position and posture information is determined based on the motion state information of the flat panel detector acquired by the inertial sensor and the position and posture information of the flat panel detector at a second moment, where the second moment is a moment before the first moment.

[0049] When the confidence level is less than or equal to the preset threshold, it indicates that the flat-panel detector may be severely occluded in the first image, resulting in very inaccurate results output by the flat-panel detector positioning model. In this case, it is necessary to use the second position posture information as the positioning information of the flat-panel detector to avoid interruption of the positioning process.

[0050] Inertial sensors can be six-axis or nine-axis. For example, a nine-axis inertial sensor combines different types of sensing elements, including a three-axis accelerometer, a three-axis gyroscope, and a three-axis magnetometer, to accurately measure the motion state and attitude of an object in three-dimensional space. Motion state information includes angular velocity and linear acceleration. By integrating and filtering this data, the relative displacement and rotation change of the detector from the second moment to the first moment can be estimated. Based on the position and attitude information at the second moment, the relative displacement and rotation change, the second position and attitude information of the flat-panel detector at the first moment can be obtained.

[0051] In actual applications, the position and posture information determined by the data from the inertial sensor can remain accurate for a certain period of time. For example, within one minute, the flat-panel detector will not be blocked for more than one minute. Therefore, in actual applications, this embodiment collects data through the inertial sensor to determine the position and posture to supplement the situation where the positioning model cannot output accurate positioning information. It can determine the accurate and continuous motion trajectory and ensure the accuracy, stability and robustness of positioning.

[0052] The technical solution of this embodiment is to acquire a first image of a flat-panel detector by an image acquisition device; input the first image into a flat-panel detector positioning model to obtain a feature mark of the flat-panel detector in the first image and the pixel coordinates of the feature points associated with the feature mark; determine first position and posture information and confidence of the flat-panel detector at a first moment corresponding to the first image based on the pixel coordinates, the intrinsic parameter matrix of the image acquisition device, and the physical coordinates of the flat-panel detector coordinate system corresponding to the feature points; when the confidence is greater than a preset threshold, use the first position and posture information as the positioning information of the flat-panel detector; when the confidence is less than or equal to the preset threshold, use second position and posture information as the positioning information of the flat-panel detector, wherein the second position and posture information is determined based on the motion state information of the flat-panel detector acquired by the inertial sensor and the position and posture information of the flat-panel detector at a second moment, the second moment being the moment before the first moment; wherein the flat-panel detector positioning model is a neural network model. The technical solution of the embodiment of the present invention solves the problem of large size and high cost of current magnetic positioning devices for positioning flat-panel detectors. By combining visual detection and inertial sensor positioning, the positioning cost of the flat-panel detector can be reduced, the positioning reliability and portability can be improved, and the positioning can be adapted to different scenarios.

[0053] Figure 2This is a flowchart of a flat-panel detector positioning method provided in an embodiment of the present invention. This embodiment, which shares the same inventive concept as the flat-panel detector positioning method described in the previous embodiment, further describes the process of determining the second position and posture information. This method can be performed by a flat-panel detector positioning device, which can be implemented using software and / or hardware and integrated into a computer device with application development capabilities.

[0054] like Figure 2 As shown, the flat panel detector positioning method of this embodiment includes the following steps:

[0055] S210 , an image acquisition device acquires a first image of a flat panel detector.

[0056] S220 , input the first image into a flat panel detector positioning model to obtain a feature mark of the flat panel detector in the first image and pixel coordinates of feature points associated with the feature mark.

[0057] The characteristic points may include characteristic points at at least three different positions. In an optional embodiment, the characteristic points include four corner points and midpoints of four sides of the flat panel detector.

[0058] The four corner points form the basic topological framework of the flat-panel detector, directly determining its shape and orientation. The edge midpoints, located at the geometric center of each edge of the flat-panel detector, assist in verifying edge symmetry and equal length. Corner points provide vertex constraints, while edge midpoints provide edge constraints. Combining these two creates a more complete set of geometric constraints, reducing false positives.

[0059] By combining the feature points of corner points and edge midpoints, when subsequently calculating the pose information through pixel coordinates, the rotation angle can be calculated by the direction of the line connecting the edge midpoints. This makes the midpoint distribution more even and has stronger noise resistance.

[0060] In an image, when a corner point of a flat-panel detector is blocked, the midpoint of the edge can still be used as a valid feature point to maintain positioning.

[0061] In this embodiment, eight feature points are set so that when the pose information is subsequently calculated through the pose algorithm and pixel coordinates, only at least three feature points are needed to determine the pose information, thereby avoiding positioning failure caused by obstruction of the flat-panel detector.

[0062] In an optional embodiment, different feature points are associated with different patterns of feature marks.

[0063] like Figure 3As shown, a specific marker is placed at each of the four corners and the midpoint of each of the four sides of the flat-panel detector, making each feature point highly recognizable. The specific marker can be a different color, shape, or coding pattern. For example, different shapes of markers, such as a "+" sign, circle, square, or diamond, are used at the four corners to provide sufficient information for subsequent pose calculations. Through this specific pattern design, no matter the angle the camera is shooting from, the image can clearly distinguish the detector. Even if the detector is partially obscured, as long as the camera can see some of the markers, such as at least three markers, the system can use the prior pattern layout to infer the complete pose of the detector relative to the camera.

[0064] S230 , determining first position and posture information and a confidence level of the flat-panel detector at a first moment corresponding to the first image based on the pixel coordinates, the intrinsic parameter matrix of the image acquisition device, and the physical coordinates of the flat-panel detector coordinate system corresponding to the feature points.

[0065] S240: When the confidence level is greater than a preset threshold, use the first position posture information as positioning information of the flat panel detector.

[0066] S250 : Acquire motion state information of the flat panel detector between the second moment and the first moment through an inertial sensor, where the motion state information includes angular velocity and linear acceleration.

[0067] The inertial sensor is set on the flat-panel detector and can obtain the motion state information of the flat-panel detector during the positioning process. The motion state information includes angular velocity and linear acceleration.

[0068] For example, the angular velocity may be acquired by a gyroscope of the inertial sensor, and the linear acceleration may be acquired by an accelerometer of the inertial sensor.

[0069] S260 , integrating and filtering the motion state information to obtain displacement change information and rotation change information of the flat panel detector from the second moment to the first moment.

[0070] Specifically, the process of determining the rotation change information can be to perform discrete integration of the angular velocity to obtain the angle, suppress high-frequency noise through low-pass filtering, or correct the attitude through complementary filtering to fuse accelerometer and magnetometer data, or Kalman filtering to use the state equation to optimally estimate the angle to obtain the rotation change information of the flat-panel detector from the second moment to the first moment.

[0071] The displacement change information can be determined by removing the gravity component from the linear acceleration to obtain the motion acceleration, integrating it once to obtain the velocity, and then integrating it again to obtain the displacement. Due to accelerometer noise and zero bias, high-pass filtering can be used to remove the trend term. This is then combined with the gyroscope attitude solution to calculate the gravity projection to obtain the displacement change information of the flat panel detector from the second moment to the first moment.

[0072] S270 , determining second position and posture information of the flat panel detector at the first moment based on the displacement change information, the rotation change information, and the position and posture information at the second moment.

[0073] The position information and the displacement change information in the position and attitude information at the second moment are superimposed, and the attitude information and the rotation change information in the position and attitude information at the second moment are synthesized to determine the second position and attitude information at the first moment.

[0074] The position and attitude information at the second moment is determined by a flat-panel detector model or calculated based on motion state information acquired by an inertial sensor. Specifically, if the confidence level of the position and attitude information output by the flat-panel detector positioning model for the second image corresponding to the second moment is greater than a preset threshold, the position and attitude information output by the flat-panel detector positioning model is used as the position and attitude information at the second moment. If the confidence level of the position and attitude information output by the flat-panel detector positioning model for the second image corresponding to the second moment is less than or equal to a preset threshold, the position and attitude information calculated based on motion state information acquired by the inertial sensor is used as the position and attitude information at the second moment.

[0075] S280: When the confidence level is less than or equal to a preset threshold, use the second position posture information as positioning information of the flat panel detector.

[0076] After the second position and posture information at the first moment is used as the positioning information of the flat panel detector, at each moment after the first moment, it is determined whether the first position and posture information or the second position and posture information at each moment after the first moment is used as the positioning information based on the confidence level.

[0077] In this embodiment, when the visual recognition module is temporarily unable to provide position information due to reasons such as pattern occlusion, the IMU module continues to track the motion state of the detector to ensure that the positioning process is not interrupted. In order to make full use of both visual and inertial information sources, this embodiment adopts a visual-inertial fusion strategy. When the first position posture information is greater than a preset confidence threshold, the first position posture information is used as positioning information. If image recognition fails or the pattern is severely occluded, resulting in the visual module being unable to provide reliable positioning information, it automatically switches to relying on the inertial data, i.e., motion state data, of the inertial sensor IMU (Inertial Measurement Unit) to determine the position posture information of the detector, and uses the second position posture information as the positioning information of the flat-panel detector. In this embodiment, through the vision and inertial fusion strategy, the system exhibits a high level of stability and continuity during the positioning process. Even if the detector is temporarily blocked or violent movements occur during movement, it can still continue to output reliable position information.

[0078] To improve the accuracy of positioning information determination, in an optional implementation, the first position and posture information is corrected according to the motion state information to obtain third position and posture information, and the third position and posture information is used as the positioning information of the flat panel detector.

[0079] Specifically, if the confidence level is greater than a preset threshold, the second position and posture information of the flat-panel detector can be calculated using the method described in steps S250-S270. The first position and posture information are then fused with the second position and posture information using a filtering algorithm, such as a Kalman filter or a Bayesian filter, to obtain third position and posture information, which is then used as the positioning information for the flat-panel detector. Correcting the first position and posture information using the second position and posture information to obtain the third position and posture information can improve positioning accuracy, reduce jitter, and achieve a smoother motion trajectory.

[0080] In an optional embodiment, the training process of the flat-panel detector positioning model can be to obtain a sample image of the flat-panel detector obtained by an image acquisition device; use the sample image as model input, use the sample feature mark of the flat-panel detector in the sample image and the sample pixel coordinates of the sample feature point associated with the sample feature mark as label information, train the initial flat-panel detector positioning model, and obtain the flat-panel detector positioning model when the loss function reaches a preset convergence threshold.

[0081] The sample images include images captured by the image capture device under at least one shooting angle, shooting distance, shooting background and lighting condition.

[0082] The camera captures RGB video images containing the detector in real time. The image is analyzed using a flat-panel detector localization initialization model to extract the locations of the detector pattern's feature points. Taking a neural network model based on the Transformer architecture as an example, the model is trained offline using a large amount of manually annotated data. During training, a large number of images captured at various angles, distances, backgrounds, and lighting conditions are collected. The locations and pixel coordinates of the feature point markers in the patterns in the images are manually annotated as label information to supervise model learning. The model parameters are adjusted based on a loss function. When the loss function reaches a preset convergence threshold, a flat-panel detector localization model is derived. The flat-panel detector localization model automatically identifies the feature markers of the flat-panel detector in the image and accurately outputs the pixel coordinates corresponding to the associated feature points. Compared to traditional image processing or simple machine vision algorithms, deep learning models are more robust and can reliably detect the detector's position even in the presence of changing lighting or partial occlusion of the pattern.

[0083] The technical solution of this embodiment is to acquire a first image of a flat-panel detector by an image acquisition device; input the first image into a flat-panel detector positioning model to obtain a feature mark of the flat-panel detector in the first image, and the pixel coordinates of a feature point associated with the feature mark; determine first position and posture information and a confidence level of the flat-panel detector at a first moment corresponding to the first image based on the pixel coordinates, an internal parameter matrix of the image acquisition device, and the physical coordinates of the flat-panel detector coordinate system corresponding to the feature point; when the confidence level is greater than a preset threshold, use the first position and posture information as positioning information of the flat-panel detector; acquire motion state information of the flat-panel detector between the second moment and the first moment through an inertial sensor, the motion state information including angular velocity and linear acceleration; integrate and filter the motion state information to obtain a first position and posture information of the flat-panel detector from the second moment to the first moment; The method comprises the following steps: obtaining displacement change information and rotation change information from the second moment to the first moment; determining the second position and attitude information of the flat-panel detector at the first moment based on the displacement change information, the rotation change information, and the position and attitude information at the second moment; obtaining the motion state information of the flat-panel detector between the second moment and the first moment through an inertial sensor, the motion state information including angular velocity and linear acceleration; integrating and filtering the motion state information to obtain the displacement change information and rotation change information of the flat-panel detector from the second moment to the first moment; determining the second position and attitude information of the flat-panel detector at the first moment based on the displacement change information, the rotation change information, and the position and attitude information at the second moment; wherein the position and attitude information at the second moment is determined by a flat-panel detector model or calculated based on the motion state information obtained by the inertial sensor. When the confidence level is less than or equal to a preset threshold, the second position and attitude information is used as the positioning information of the flat-panel detector, wherein the second position and attitude information is determined based on the motion state information of the flat-panel detector obtained by the inertial sensor and the position and attitude information of the flat-panel detector at the second moment, the second moment being the moment before the first moment; and wherein the positioning model of the flat-panel detector is a neural network model. The technical solution of the embodiment of the present invention solves the problem of large size and high cost of the current magnetic positioning device for positioning flat-panel detectors. By combining visual detection and inertial sensor positioning methods, the positioning cost of the flat-panel detector can be reduced, the positioning reliability and portability can be improved, and it can adapt to different scenarios. By obtaining motion data and determining posture changes through inertial sensors, the positioning accuracy can be further improved.

[0084] Figure 4 This is a schematic diagram of a flat-panel detector positioning device according to an embodiment of the present invention. This embodiment is applicable to scenarios where flat-panel detector positioning is required. The flat-panel detector positioning device can be implemented using software and / or hardware and integrated into a computer terminal device with application development capabilities.

[0085] like Figure 4As shown, the flat panel detector positioning device includes: an image acquisition module 310 , a pixel coordinate determination module 320 , a first position and posture information determination module 330 , a first positioning information determination module 340 and a second positioning information determination module 350 .

[0086] Among them, the image acquisition module 310 is used to acquire a first image of a flat-panel detector acquired by an image acquisition device; the pixel coordinate determination module 320 is used to input the first image into the flat-panel detector positioning model to obtain a feature mark of the flat-panel detector in the first image, and the pixel coordinates of the feature points associated with the feature mark; the first position and posture information determination module 330 is used to determine the first position and posture information and confidence level of the flat-panel detector at a first moment corresponding to the first image based on the pixel coordinates, the intrinsic parameter matrix of the image acquisition device, and the physical coordinates of the flat-panel detector coordinate system corresponding to the feature points; the first positioning information determination module 340 is used to use the first position and posture information as the positioning information of the flat-panel detector when the confidence level is greater than a preset threshold; the second positioning information determination module 350 is used to use the second position and posture information as the positioning information of the flat-panel detector when the confidence level is less than or equal to a preset threshold, wherein the second position and posture information is determined based on the motion state information of the flat-panel detector acquired by the inertial sensor and the position and posture information of the flat-panel detector at a second moment, where the second moment is a moment before the first moment; wherein the flat-panel detector positioning model is a neural network model.

[0087] The technical solution of this embodiment is to acquire a first image of the flat panel detector by acquiring an image acquisition device;

[0088] The first image is input into the flat-panel detector positioning model to obtain the feature mark of the flat-panel detector in the first image and the pixel coordinates of the feature points associated with the feature mark; based on the pixel coordinates, the intrinsic parameter matrix of the image acquisition device, and the physical coordinates of the flat-panel detector coordinate system corresponding to the feature points, the first position and posture information and confidence of the flat-panel detector at the first moment corresponding to the first image are determined; when the confidence is greater than a preset threshold, the first position and posture information is used as the positioning information of the flat-panel detector; when the confidence is less than or equal to the preset threshold, the second position and posture information is used as the positioning information of the flat-panel detector, wherein the second position and posture information is determined based on the motion state information of the flat-panel detector obtained by the inertial sensor and the position and posture information of the flat-panel detector at the second moment, the second moment being the moment before the first moment; wherein the flat-panel detector positioning model is a neural network model. The technical solution of the embodiment of the present invention solves the problem of large size and high cost of the current magnetic positioning device for positioning flat-panel detectors. By combining the positioning method of visual detection and inertial sensors, the positioning cost of the flat-panel detector can be reduced, the positioning reliability and portability can be improved, and it can adapt to different scenarios.

[0089] In an optional embodiment, the feature points include four corner points and midpoints of four sides of the flat panel detector.

[0090] In an optional embodiment, different feature points are associated with different patterns of feature marks.

[0091] In an optional embodiment, the device further comprises:

[0092] The second position and attitude information determination unit is configured to obtain motion state information of the flat-panel detector between a second moment and a first moment through an inertial sensor, the motion state information including angular velocity and linear acceleration; integrate and filter the motion state information to obtain displacement change information and rotation change information of the flat-panel detector from the second moment to the first moment; and determine second position and attitude information of the flat-panel detector at the first moment based on the displacement change information, the rotation change information, and the position and attitude information at the second moment; wherein the position and attitude information at the second moment is determined by a flat-panel detector model or calculated using the motion state information obtained by the inertial sensor.

[0093] The third positioning information determination module is used to correct the first position and posture information according to the motion state information to obtain third position and posture information, and use the third position and posture information as the positioning information of the flat panel detector.

[0094] The model training module is used to obtain sample images of a flat-panel detector captured by an image acquisition device; use the sample images as model input, use the sample feature marks of the flat-panel detector in the sample images and the sample pixel coordinates of the sample feature points associated with the sample feature marks as label information, train the initial model for positioning the flat-panel detector, and obtain the flat-panel detector positioning model when the loss function reaches a preset convergence threshold; wherein the sample images include images captured by the image acquisition device under at least one shooting angle, shooting distance, shooting background and lighting condition.

[0095] The flat panel detector positioning device provided in the embodiment of the present invention can execute the flat panel detector positioning method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0096] Figure 5 A schematic structural diagram of a flat panel detector positioning system provided by an embodiment of the present invention.

[0097] like Figure 5 As shown, the flat panel detector positioning system includes: an image acquisition device 410 , an inertial measurement device 430 and a calculation module 420 .

[0098] Among them, the image acquisition device 410 is used to acquire images of the flat-panel detector; the inertial measurement device 430 is set on the flat-panel detector and is used to acquire the motion state information of the flat-panel detector; the calculation module 420 is used to determine the positioning information of the flat-panel detector according to the flat-panel detector positioning method provided by any embodiment of the present invention.

[0099] Image acquisition device 410 can be an RGB camera. Inertial measurement units (IMUs) are highly integrated and compact, suitable for embedded devices and offering higher measurement accuracy. They include accelerometers and gyroscopes and can be nine-axis or six-axis IMUs. Computing module 420 can be an embedded platform. The flat-panel detector model has a small size and efficient inference speed, enabling real-time execution on an embedded computing platform. Computing module 420 can be embedded in a mobile X-ray machine or portable terminal, eliminating the need for bulky, high-performance computers to perform calculations, thereby improving the system's practicality and portability.

[0100] The technical solution of this embodiment utilizes an image acquisition device for capturing images from a flat-panel detector; an inertial measurement device disposed on the flat-panel detector for collecting information about the flat-panel detector's motion state; and a computing module for determining the flat-panel detector's positioning information based on the flat-panel detector positioning method provided in any embodiment of the present invention. This technical solution addresses the issues of large size and high cost associated with current magnetic positioning devices for locating flat-panel detectors. By combining visual detection with inertial sensor positioning, this solution reduces the cost of flat-panel detector positioning, improves positioning reliability and portability, and enables adaptation to diverse scenarios.

[0101] Figure 6 A schematic structural diagram of a computer device provided in an embodiment of the present invention. Figure 6 A block diagram of an exemplary computer device 12 suitable for use in implementing embodiments of the present invention is shown. Figure 6 The computer device 12 shown is only an example and should not limit the functionality and scope of use of the embodiments of the present invention. The computer device 12 can be any terminal device with computing capabilities, such as an intelligent controller, a server, a mobile phone, or other terminal devices.

[0102] like Figure 6 As shown, computer device 12 is implemented as a general-purpose computing device. Components of computer device 12 may include, but are not limited to, one or more processors or processing units 16, system memory 28, and a bus 18 that connects various system components (including system memory 28 and processing unit 16).

[0103] Bus 18 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, or a local bus using any of a variety of bus architectures. Examples of these architectures include, but are not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MAC) bus, an Enhanced ISA bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnect (PCI) bus.

[0104] The computer device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the computer device 12, including volatile and non-volatile media, removable and non-removable media.

[0105] System memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache 32. Computer device 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 34 may be configured to read and write non-removable, non-volatile magnetic media ( Figure 6 Not shown, often called a "hard drive"). Although Figure 6 Not shown, a magnetic disk drive for reading and writing to a removable non-volatile magnetic disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 18 via one or more data media interfaces. System memory 28 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of various embodiments of the present invention.

[0106] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in system memory 28. Such program modules 42 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may include an implementation of a network environment. Program modules 42 generally perform the functions and / or methods of the embodiments described herein.

[0107] The computer device 12 may also communicate with one or more external devices 14 (e.g., a keyboard, a pointing device, a display 24, etc.), one or more devices that enable a user to interact with the computer device 12, and / or any device that enables the computer device 12 to communicate with one or more other computing devices (e.g., a network card, a modem, etc.). Such communication may be performed via an input / output (I / O) interface 22. Furthermore, the computer device 12 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 20. As shown, the network adapter 20 communicates with the other modules of the computer device 12 via the bus 18. It should be understood that although Figure 6 Not shown, other hardware and / or software modules may be used in conjunction with the computer device 12, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RFID systems, tape drives, and data backup storage systems.

[0108] The processing unit 16 executes various functional applications and data processing by running programs stored in the system memory 28, such as implementing the flat panel detector positioning method provided in the embodiment of the present invention, which includes:

[0109] An image acquisition device acquires a first image of the flat panel detector;

[0110] Inputting the first image into the flat panel detector positioning model to obtain a feature marker of the flat panel detector in the first image and pixel coordinates of a feature point associated with the feature marker;

[0111] Determine first position and posture information and a confidence level of the flat-panel detector at a first moment corresponding to the first image based on the pixel coordinates, the intrinsic parameter matrix of the image acquisition device, and the physical coordinates of the flat-panel detector coordinate system corresponding to the feature points;

[0112] When the confidence level is greater than a preset threshold, the first position posture information is used as positioning information of the flat panel detector;

[0113] When the confidence level is less than or equal to a preset threshold, using the second position and posture information as the positioning information of the flat-panel detector, wherein the second position and posture information is determined based on the motion state information of the flat-panel detector obtained by the inertial sensor and the position and posture information of the flat-panel detector at a second moment, where the second moment is a moment before the first moment;

[0114] Among them, the flat panel detector positioning model is a neural network model.

[0115] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the method for positioning a flat panel detector provided in any embodiment of the present invention is implemented. The method includes:

[0116] An image acquisition device acquires a first image of the flat panel detector;

[0117] Inputting the first image into the flat panel detector positioning model to obtain a feature marker of the flat panel detector in the first image and pixel coordinates of a feature point associated with the feature marker;

[0118] Determine first position and posture information and a confidence level of the flat-panel detector at a first moment corresponding to the first image based on the pixel coordinates, the intrinsic parameter matrix of the image acquisition device, and the physical coordinates of the flat-panel detector coordinate system corresponding to the feature points;

[0119] When the confidence level is greater than a preset threshold, the first position posture information is used as positioning information of the flat panel detector;

[0120] When the confidence level is less than or equal to a preset threshold, using the second position and posture information as the positioning information of the flat-panel detector, wherein the second position and posture information is determined based on the motion state information of the flat-panel detector obtained by the inertial sensor and the position and posture information of the flat-panel detector at a second moment, where the second moment is a moment before the first moment;

[0121] Among them, the flat panel detector positioning model is a neural network model.

[0122] The computer storage medium of the embodiment of the present invention can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to: an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or component, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device.

[0123] A computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0124] Program code embodied on a computer-readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0125] The computer program code for performing the operations of the present invention can be written in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, Python, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0126] An embodiment of the present invention further provides a computer program product, including a computer program, which, when executed by a processor, implements the flat panel detector positioning method provided in any embodiment of the present application.

[0127] The computer program product, during implementation, may be written in one or more programming languages, or a combination thereof, for performing the operations of the present invention. The programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, Python, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0128] Those skilled in the art will appreciate that the modules or steps of the present invention described above can be implemented using a general-purpose computing device. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Alternatively, they can be implemented using program code executable by a computer device, which can then be stored in a storage device and executed by the computing device. Alternatively, they can be fabricated into separate integrated circuit modules, or multiple modules or steps can be fabricated into a single integrated circuit module. Thus, the present invention is not limited to any specific combination of hardware and software.

[0129] Note that the above are only preferred embodiments of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments, and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments and may include many other equivalent embodiments without departing from the concept of the present invention. The scope of the present invention is determined by the scope of the appended claims.

Claims

1. A method for positioning a flat panel detector, characterized in that: include: An image acquisition device acquires a first image of the flat panel detector; Inputting the first image into a flat panel detector positioning model to obtain a feature marker of the flat panel detector in the first image and pixel coordinates of a feature point associated with the feature marker; Determine first position and posture information and a confidence level of the flat-panel detector at a first moment corresponding to the first image based on the pixel coordinates, the intrinsic parameter matrix of the image acquisition device, and the physical coordinates of the flat-panel detector coordinate system corresponding to the feature points; When the confidence level is greater than a preset threshold, using the first position and posture information as positioning information of the flat panel detector; When the confidence level is less than or equal to a preset threshold, using second position and posture information as positioning information of the flat-panel detector, wherein the second position and posture information is determined based on motion state information of the flat-panel detector acquired by an inertial sensor and position and posture information of the flat-panel detector at a second moment, where the second moment is a moment before the first moment; Wherein, the flat panel detector positioning model is a neural network model.

2. The method according to claim 1, characterized in that The feature points include four corner points and midpoints of four sides of the flat panel detector.

3. The method according to claim 1 or 2, characterized in that Different feature points are associated with different patterns of the feature marks.

4. The method according to claim 1, wherein The process of determining the second position and posture information includes: Acquiring motion state information of the flat panel detector between the second moment and the first moment through an inertial sensor, wherein the motion state information includes angular velocity and linear acceleration; Integrating and filtering the motion state information to obtain displacement change information and rotation change information of the flat panel detector from the second moment to the first moment; determining second position and posture information of the flat panel detector at the first moment according to the displacement change information, the rotation change information, and the position and posture information at the second moment; The position and posture information at the second moment is determined by the flat panel detector model, or is calculated and determined by motion state information acquired by the inertial sensor.

5. The method according to claim 1, wherein The method also includes: The first position and posture information is corrected according to the motion state information to obtain third position and posture information, and the third position and posture information is used as positioning information of the flat panel detector.

6. The method according to claim 1, wherein The training process of the flat panel detector positioning model includes: Acquiring a sample image of the flat panel detector acquired by an image acquisition device; The sample image is used as a model input, and the sample feature mark of the flat panel detector in the sample image and the sample pixel coordinates of the sample feature point associated with the sample feature mark are used as label information to train the flat panel detector positioning initial model, and obtain the flat panel detector positioning model when the loss function reaches a preset convergence threshold; The sample images include images captured by the image acquisition device under at least one shooting angle, shooting distance, shooting background and lighting condition.

7. A flat panel detector positioning device, characterized in that: include: An image acquisition module, configured to acquire a first image captured by the flat panel detector by the image acquisition device; a pixel coordinate determination module, configured to input the first image into a flat panel detector positioning model to obtain a feature marker of the flat panel detector in the first image and pixel coordinates of a feature point associated with the feature marker; a first position and posture information determination module, configured to determine first position and posture information and a confidence level of the flat-panel detector at a first moment corresponding to the first image based on the pixel coordinates, an intrinsic parameter matrix of the image acquisition device, and the physical coordinates of the flat-panel detector coordinate system corresponding to the feature points; a first positioning information determining module, configured to use the first position posture information as the positioning information of the flat panel detector when the confidence level is greater than a preset threshold; a second positioning information determining module, configured to use second position and posture information as positioning information of the flat-panel detector if the confidence level is less than or equal to a preset threshold, wherein the second position and posture information is determined based on motion state information of the flat-panel detector acquired by an inertial sensor and position and posture information of the flat-panel detector at a second moment, the second moment being a moment before the first moment; Wherein, the flat panel detector positioning model is a neural network model.

8. A flat panel detector positioning system, characterized in that: include: Image acquisition device, inertial measurement unit and computing module; Wherein, the image acquisition device is used to acquire images from a flat panel detector; The inertial measurement device is provided on the flat panel detector and is used to collect motion state information of the flat panel detector; The calculation module is configured to determine the positioning information of the flat panel detector according to the flat panel detector positioning method according to any one of claims 1 to 6.

9. A computer device, characterized in that: The computer device comprises: one or more processors; a memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the flat panel detector positioning method according to any one of claims 1 to 6.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the flat panel detector positioning method according to any one of claims 1 to 6 is implemented.

Images