Depth camera based skin dose calculation method

Abstract

The application discloses a skin dose calculation method based on a depth camera, comprising the following steps: obtaining the transformation relationship between the beam coordinate system of a three-dimensional perspective device and the camera coordinate system of the depth camera by calibrating a calibration phantom with a plurality of marker balls; collecting corresponding images by the depth camera and calculating the coordinates of each pixel point in the image under the beam coordinate system according to the transformation relationship; determining the beam region according to the beam angle of the ray source, combining the coordinates of each pixel point to determine the actual incident position and the incident depth of the patient, and calculating the corresponding skin dose accordingly. The application can locate the actual incident position of the patient by the depth camera, calculate the skin dose at the actual incident position, and display accurate skin dose indication.

Description

Technical Field

[0001] This invention relates to the field of X-ray technology, and more particularly to a method for calculating skin dose based on a depth camera. Background Technology

[0002] Skin dose is a measure of the ionizing radiation energy received or “absorbed” per unit mass of skin tissue. Currently, cone-beam computed tomography (CBCT) systems indicate the patient's incident dose (skin dose) by the manufacturer using a fixed incident position set empirically. The dose rate and dose at this unobstructed position are calculated as the patient's incident dose indication. However, this method is only an estimate. The patient's distance from the probe and their posture can significantly affect the actual radiation dose received, causing a discrepancy between the actual dose received and the CT-indicated dose. Therefore, a new skin dose calculation method is needed to monitor the actual radiation dose received by the patient. Summary of the Invention

[0003] Purpose of the invention: To address the above-mentioned shortcomings, this invention proposes a skin dose calculation method based on a depth camera, which uses a depth camera to locate the patient's actual incident position, thereby providing an accurate patient incident dose.

[0004] Technical solution:

[0005] Skin dose calculation methods based on depth cameras include:

[0006] The transformation relationship between the beam coordinate system of the 3D perspective device and the camera coordinate system of the depth camera is obtained by calibrating a calibration phantom with several marker balls inside.

[0007] The patient's lesion image is acquired by a depth camera, and the coordinates of each pixel in the lesion image in the beam coordinate system are calculated according to the aforementioned transformation relationship.

[0008] The beam area is determined based on the beam angle of the radiation source, and the actual incident position and depth of the patient are determined by combining the coordinates of each pixel point mentioned above. The corresponding skin dose is then calculated accordingly.

[0009] The calibration model contains two layers of calibration balls, one above the other. Each calibration ball in each layer is the same height, but the spacing between the calibration balls in each layer is different.

[0010] The specific transformation relationship between the beam coordinate system of the 3D perspective device and the camera coordinate system of the depth camera obtained through calibration is as follows:

[0011] A three-dimensional perspective device scans the perspective image of the calibration model, identifies and extracts the coordinates of the calibration sphere and its centroid.

[0012] The layer number of each calibration ball is obtained by identifying the spacing between calibration balls, and the height of the calibration ball is obtained according to the design parameters of the calibration phantom, thereby obtaining the coordinates of the centroid of the calibration ball in the beam coordinate system;

[0013] Images of the calibration phantom are acquired using a depth camera, and the coordinates of the centroid of the calibration sphere in the camera coordinate system are obtained by extracting the images.

[0014] Calculate the transformation relationship between the beam coordinate system and the camera coordinate system.

[0015] The threshold method was used to identify and extract the coordinates of the calibration sphere and its centroid in the perspective image.

[0016] The specific definition of the beam region is as follows:

[0017]

[0018] in, Let P2.x be the beam angle of the X-ray source, and P2.y and P2.z be the coordinates of each pixel on the x-axis, y-axis, and z-axis of the beam coordinate system, respectively.

[0019] By combining the coordinates of each pixel mentioned above, pixels that are not within the beam area are eliminated to determine the patient's actual incident position.

[0020] The specific steps to determine the patient's actual injection location and depth are as follows:

[0021] Within the beam area, acquire the pixel closest to the ray source and determine whether it is an anomaly.

[0022] If it is not an abnormal point, then its distance from the radiation source is taken as the patient's incidence depth.

[0023] If it is an outlier, it will be removed from the beam region;

[0024] Repeat the above process until the patient's injection depth is obtained.

[0025] The specific steps for determining whether a pixel within the beam region is an anomaly are as follows:

[0026] Calculate the distance h0 between the target pixel and the ray source on the perspective image, and the distance h between each pixel within a set distance range and the ray source. Based on this, calculate the corresponding distance difference Δh = h - h0, and calculate the average value and standard deviation of the corresponding distance difference Δh. Determine whether the average value exceeds the first threshold H1 and the standard deviation exceeds the second threshold H2. If neither exceeds, the target point is considered not an anomaly; otherwise, the target point is considered an anomaly.

[0027] The corresponding skin dose calculated is as follows:

[0028] Obtain the fixed incident depth H that varies with time. C radiation dose rate R C Based on this, the patient's actual incident dose rate R is calculated. O =R C *(H C / H O ) 2 Thus, the corresponding skin dose can be calculated. Where T is the total exposure time, and R is... Ot Let be the incident dose at time t.

[0029] Beneficial effects: This invention uses a depth camera to locate the patient's actual injection point and calculates the skin dose at the actual injection point, which can display an accurate skin dose indication. Attached Figure Description

[0030] Figure 1 This is a flowchart of the skin dosage calculation method of the present invention;

[0031] Figure 2 This is a schematic diagram of the calibration phantom.

[0032] Figure 3 This is a schematic diagram of the beam coordinate system. Detailed Implementation

[0033] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments.

[0034] like Figure 1 As shown, the skin dose calculation method based on a depth camera of the present invention includes the following steps:

[0035] (1) Parameter calibration;

[0036] The parameter calibration of this invention adopts the following method: Figure 2 The calibration phantom shown is a cubic structure containing two layers of calibration balls. Specifically, the calibration balls are colored metal balls, or more specifically, steel balls. All calibration balls in each layer have the same height, which is h. u and h d The spacing between the calibration balls in each layer is different to identify the height of the calibration balls. The specific parameter calibration steps are as follows:

[0037] (11) The C-arm of the CBCT is placed vertically with the high-voltage X-ray tube (i.e., the X-ray source) above it. The depth camera is mounted on the side of the high-voltage X-ray tube. A horizontal base is set between the high-voltage X-ray tube and the detector (the base will not significantly obstruct the X-ray). The calibration phantom is placed on the base and positioned within the field of view of the C-arm and the depth camera. The height of the base surface is measured as h. b The source-image distance between the X-ray source and the detector is obtained as sid;

[0038] (12) Scan the perspective image of the calibration phantom, identify and extract the coordinates of the calibration sphere and its centroid;

[0039] Specifically, the calibration sphere is obtained by using the threshold method (i.e., extracting pixels with pixel values ​​less than a set threshold on the perspective image) and calculating its centroid coordinate Z in the image coordinate system.

[0040] (13) The layer number of each calibration ball is obtained by identifying the spacing between calibration balls, and the height h of a certain calibration ball is obtained according to the design parameters of the calibration phantom. The coordinates Z1 of its centroid in the beam coordinate system are calculated by geometric projection based on the centroid coordinates of the ball in the image coordinate system obtained in step (12), where the beam coordinate system is as follows: Figure 3 As shown, a coordinate system is established with the ray source as the origin and the z-axis perpendicular to the imaging plane.

[0041] Specifically, the coordinate values ​​Z1.x, Z1.y, and Z1.z of a certain calibration ball in the beam coordinate system along the x-axis, y-axis, and z-axis are respectively:

[0042] Z1.x = Zx*(h+h) b ) / sid

[0043] Z1.y=Zy*(h+h b ) / sid

[0044] Z1.z=h+h b

[0045] Where Zx and Zy are the x-axis and y-axis coordinates of the calibration sphere in the image coordinate system, respectively;

[0046] (14) Acquire images of the calibration phantom using a depth camera and extract the coordinates Z2 of the centroid of the calibration sphere in the camera coordinate system;

[0047] In this invention, the depth camera is an RGB binocular camera, and the coordinates of the centroid of the calibration sphere in the camera coordinate system are obtained by recognizing the acquired RGB image and depth image.

[0048] (15) Based on the coordinates of the centroids of each marker sphere in the beam coordinate system obtained in step (13) and the coordinates of the centroids of each marker sphere in the camera coordinate system obtained in step (14), the linear transformation matrix R between the beam coordinate system and the camera coordinate system is calculated:

[0049] R = (Z2) T *Z2) -1 *Z2 T *Z1

[0050] Among them, Z2 T This represents the transpose operation on Z2;

[0051] (2) Locate the patient's actual injection position and injection depth;

[0052] (21) Acquire RGB and depth images of the patient’s lesion using a depth camera, and identify the coordinates P1 of each pixel in the image in the camera coordinate system. Calculate the coordinates P2 = R * P1 in the beam coordinate system based on the linear transformation matrix R obtained in step (1).

[0053] (22) Based on the beam angle of the CBCT X-ray source Determine the beam region C, such as Figure 3 As shown, the details are as follows:

[0054]

[0055] Wherein, P2.x, P2.y, and P2.z are the coordinate values ​​of each pixel on the x-axis, y-axis, and z-axis in the beam coordinate system, respectively;

[0056] Based on whether each pixel obtained in the above judgment step (21) is within the beam region C, remove the pixels that are not within the beam region C;

[0057] (23) Obtain the pixel closest to the X-ray source (i.e., high-voltage X-ray tube) within the beam region C and determine whether it is an abnormal point;

[0058] If it is not an abnormal point, then its distance from the radiation source is taken as the patient's incidence depth H. O ;

[0059] If it is an abnormal point, remove it from the beam region C, and repeat this step until the patient's incident depth H is obtained. O ;

[0060] Specifically, determining whether a pixel within the beam region C is an outlier involves:

[0061] Calculate the distance h0 between the target pixel and the ray source and the distance h between each pixel around the target pixel and the ray source in the perspective image. Calculate the distance difference Δh = h - h0 between h and h0 for each pixel around the target pixel. Calculate the average and standard deviation of each distance difference. Based on this, determine whether the average exceeds the first threshold H1 and whether the standard deviation exceeds the second threshold H2. If neither exceeds the threshold, the target point is considered not an anomaly; otherwise, the target point is considered an anomaly.

[0062] Among them, the surrounding pixels are defined as pixels on the perspective image whose distance from the target pixel is less than a set distance range.

[0063] (3) Calculate the incident dose;

[0064] A fixed incident depth H is obtained by using an external dose detection device or by calculating exposure parameters such as voltage and current. C radiation dose rate R C Based on this, the patient's actual incident dose rate R is calculated. O =R C *(H C / H O ) 2 While the incident depth H changes over time C radiation dose rate R C It may also change, thus allowing the incident dose to be calculated. Where T is the total exposure time, and R is... Ot Let be the incident dose at time t.

[0065] This invention uses a calibration phantom to calibrate parameters, thereby obtaining the transformation relationship between the beam coordinate system and the camera coordinate system. Then, while obtaining a fluoroscopic image of the patient's lesion through CBCT scanning, the RGB image and depth image of the patient's lesion are acquired through a depth camera to identify the patient's lesion. Thus, the actual incident position of the patient can be located according to the aforementioned transformation relationship, and the skin dose at the actual incident position can be calculated accordingly. Accurate dose rate indication, cumulative dose indication, and dose area indication can be displayed.

[0066] The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the specific details of the above embodiments. Within the scope of the technical concept of the present invention, various equivalent transformations (such as quantity, shape, position, etc.) can be made to the technical solution of the present invention, and these equivalent transformations all fall within the protection scope of the present invention.

Claims

1. A skin dose calculation method based on a depth camera, characterized in that: include: The transformation relationship between the beam coordinate system of the 3D perspective device and the camera coordinate system of the depth camera is obtained by calibrating a calibration phantom with several marker balls inside. The patient's lesion image is acquired by a depth camera, and the coordinates of each pixel in the lesion image in the beam coordinate system are calculated according to the aforementioned transformation relationship. The beam region is determined based on the beam angle of the X-ray source. Within the beam region, the first pixel point closest to the X-ray source is obtained. The distance h0 between the first pixel point and the X-ray source and the distance h between each pixel point within a set distance range from the first pixel point and the X-ray source are calculated. Based on this, the corresponding distance difference Δh = h - h0 is calculated, and the average value and standard deviation of the corresponding distance difference Δh are calculated. Determine whether the average value exceeds the first threshold and whether the standard deviation exceeds the second threshold. If none of the above conditions are met, then the first pixel is considered not an anomaly, and its distance from the ray source is taken as the incident depth H. O ; Otherwise, the first pixel is considered an anomaly and is removed from the beam region; Repeat the above process until the incident depth is obtained; This determines the patient's actual injection location and depth, and the corresponding skin dose is calculated accordingly.

2. The skin dosage calculation method according to claim 1, characterized in that: The calibration mold contains two layers of calibration balls, each with the same height. The spacing between the calibration balls in the first layer is different from that in the second layer.

3. The skin dosage calculation method according to claim 2, characterized in that: The specific transformation relationship between the beam coordinate system of the 3D perspective device and the camera coordinate system of the depth camera obtained through calibration is as follows: A three-dimensional perspective device scans the perspective image of the calibration model, identifies and extracts the coordinates of the calibration sphere and its centroid. The layer number of each calibration ball is obtained by identifying the spacing between calibration balls, and the height of the calibration ball is obtained according to the design parameters of the calibration phantom, thereby obtaining the coordinates of the centroid of the calibration ball in the beam coordinate system; Images of the calibration phantom are acquired using a depth camera, and the coordinates of the centroid of the calibration sphere in the camera coordinate system are obtained by extracting the images. Calculate the transformation relationship between the beam coordinate system and the camera coordinate system.

4. The skin dosage calculation method according to claim 3, characterized in that: The threshold method was used to identify and extract the coordinates of the calibration sphere and its centroid in the perspective image.

5. The skin dosage calculation method according to claim 1, characterized in that: The specific definition of the beam region is as follows: P2.x / P2.z<tanφ && P2.y / P2.z<tanφ; Where φ is the beam angle of the ray source, and P2.x, P2.y, and P2.z are the coordinates of each pixel on the x-axis, y-axis, and z-axis in the beam coordinate system, respectively.

6. The skin dosage calculation method according to claim 1, characterized in that: By combining the coordinates of each pixel mentioned above, pixels that are not within the beam area are eliminated to determine the patient's actual incident position.

7. The skin dosage calculation method according to claim 1, characterized in that: The corresponding skin dose calculated is as follows: Obtain the fixed incident depth H that varies with time. C radiation dose rate R C Based on this, the patient's actual incident dose rate R is calculated. O =R C (H C / H O ) 2 Thus, the corresponding skin dose can be calculated. Where T is the total exposure time. Let be the incident dose at time t.

Images