Multi-dimensional image fusion method based on X-ray machine and industrial camera
By using hollow aluminum plates for image acquisition and calibration in X-ray machines and industrial cameras, the problems of image offset, telescopic and angular distortion are solved, and the fusion and registration of images are achieved, and the accuracy and efficiency of detection are improved.
Patent Information
- Application Number
- CN202311567982.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-11-23
- Publication Date
- 2024-12-06
AI Technical Summary
Traditional methods cannot achieve end-to-end comprehensive detection of X-ray machines and industrial camera images, and due to different imaging principles and parameters, the acquired images have problems with offset, scaling and angular distortion.
The hollow aluminum plate is used as the calibration plate, and images are collected under X-ray machines and industrial cameras, calibration parameters are obtained through analysis and processing, and images are preprocessed and weighted averaging algorithm are performed to realize image registration and fusion.
The geometric alignment of X-ray machines and industrial camera images is achieved, which improves detection accuracy, provides more comprehensive internal and external information of the object, reduces repeated operations, and improves detection efficiency.
Smart Images

Figure CN119090731A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of image fusion and multi-dimensional detection, and in particular to a multi-dimensional image fusion method based on an X-ray machine and an industrial camera. Background Art
[0002] In the industrial production process, it is often necessary to perform multi-dimensional inspection on the tested packages. The traditional method is generally to either detect foreign objects with different densities inside the package through X-ray images, or to detect defects on the surface of the package through industrial camera images R(x,y), which cannot achieve end-to-end comprehensive inspection. Therefore, a method is now designed to perform multi-dimensional fusion of images collected by X-ray machines and industrial cameras. However, the imaging principles and parameters of X-ray machines and industrial cameras are different, which will cause the actual images collected to have problems of offset, telescoping and angle distortion. Summary of the invention
[0003] In order to solve the above technical problems, a multi-dimensional image fusion method based on an X-ray machine and an industrial camera is provided. This technical solution solves the problem that the imaging principles and parameters of the X-ray machine and the industrial camera proposed in the above background technology are different, which will cause the actual image collected to have offset, scaling and angle distortion.
[0004] In order to achieve the above purpose, the technical solution adopted by the present invention is:
[0005] A multi-dimensional image fusion method based on an X-ray machine and an industrial camera, comprising:
[0006] A hollow aluminum plate is set as a calibration plate, and several groups of calibration plate images in different fields of view are collected under an X-ray machine and an industrial camera.
[0007] Analyze and process the calibration plate images in different fields of view of several groups of calibration plates to obtain calibration parameters between the X-ray machine and the industrial camera, wherein the calibration parameters include an offset x, an offset y, a telescopic ratio in the x direction, a telescopic ratio in the y direction, and an angle difference;
[0008] Obtain the internal image of the inspection object under the X-ray machine;
[0009] Obtain the appearance image of the detected object under the industrial camera;
[0010] Preprocessing the internal image and the appearance image to obtain a registration image of the internal image and a registration image of the appearance image;
[0011] A weighted average algorithm is performed according to the registered image of the internal image and the registered image of the appearance image to obtain a fused image.
[0012] Preferably, the step of analyzing and processing the calibration plate images in different fields of view of the plurality of calibration plates to obtain the calibration parameters between the X-ray machine and the industrial camera specifically comprises the following steps:
[0013] Select two images of the X-ray machine calibration plate and the industrial camera calibration plate with the same pixel information and environmental information from a plurality of groups of images;
[0014] Establish a rectangular coordinate system and place the two calibration plate images into different rectangular coordinate systems;
[0015] The images of the two calibration plates are extracted and processed to obtain the calibration parameters between the X-ray machine and the industrial camera.
[0016] Preferably, the extraction and calculation processing of the two calibration plate images to obtain the calibration parameters between the X-ray machine and the industrial camera specifically includes the following steps:
[0017] Perform edge threshold extraction processing on the two calibration plate images to obtain the coordinates of the corner points of the edges of the two calibration plate images;
[0018] The coordinates of the corner points of the two adjacent edges are calculated in the x-axis and y-axis directions respectively to obtain the Euclidean distance between the corner points of the edges in the two calibration plate images;
[0019] The Euclidean distances of the two calibration plate images are calculated and processed respectively to obtain the scaling ratios in the x direction and the scaling ratios in the y direction of the two calibration plate images;
[0020] The images of the two calibration plates are scaled according to the scaling ratio in the x direction and the scaling ratio in the y direction to obtain the angle difference;
[0021] According to the angle difference, the two calibration plate images are rotated and calculated to obtain the offset x and offset y.
[0022] Preferably, performing scaling processing on the images of the two calibration plates according to the scaling ratio in the x direction and the scaling ratio in the y direction to obtain the angle difference specifically comprises the following steps:
[0023] Performing scaling processing on the two calibration plate images according to the scaling ratio in the x direction and the scaling ratio in the y direction to obtain scaled images of the two calibration plates;
[0024] The distances between the upper edge of the calibration plate and the x-axis of the image in the two telescopic images are calculated and processed respectively to obtain two sets of angles;
[0025] The two sets of included angles are calculated and processed to obtain the angle difference.
[0026] Preferably, the calculation formula of the Euclidean distance is:
[0027]
[0028]
[0029] Where, d x1 is the Euclidean distance in the x-axis direction, (x1, y1) is the coordinate of the corner point of the edge in the calibration plate image under the X-ray machine, and (x2, y2) is the coordinate of the corner point of the adjacent edge in the calibration plate image under the X-ray machine; d y1 is the Euclidean distance in the y-axis direction, (x3, y3) is the coordinate of the corner point of the edge in the X-ray machine calibration plate image, (x4, y4) is the coordinate of the corner point of the adjacent edge in the y-axis direction in the X-ray machine calibration plate image; similarly, d x2 d y2 .
[0030] Preferably, the calculation formulas for the expansion ratio in the x direction and the expansion ratio in the y direction are:
[0031] scale_x=d x1 / d x2 ;
[0032] scale_y=d y1 / d y2 ;
[0033] Where scale_x is the scaling ratio in the x-axis direction, d x1 and d x2 is the Euclidean distance in the x-axis direction; scale_y is the scaling ratio in the y-axis direction, d y1 and d y2 is the Euclidean distance in the y-axis direction.
[0034] Preferably, the calculation formula of the angle difference is:
[0035] angle_x=atan((x1-x2) / (y1-y2));
[0036] angle-2d=atan((x1-x2) / (y1-y2));
[0037] angle=angle x -angle-2d;
[0038] Where angle_x is the angle between the upper edge of the calibration plate and the x-axis of the image in the telescopic image under the X-ray camera, angle_2d is the angle between the upper edge of the calibration plate and the x-axis of the image in the telescopic image under the industrial camera, and angle is the angle difference between angle_x and angle-2d.
[0039] Preferably, the preprocessing of the internal image and the appearance image to obtain the edge features of the internal image and the edge features of the appearance image specifically comprises the following steps:
[0040] Perform dynamic threshold segmentation on the internal image and the external image respectively to obtain low-pass filtered images of the internal image and the external image;
[0041] Perform grayscale processing on the low-pass filtered images of the internal image and the external image to obtain two sets of detection object mask images;
[0042] The two sets of detection object mask images are registered through calibration parameters to obtain the registered image.
[0043] Preferably, the step of performing dynamic threshold segmentation on the internal image and the external image respectively to obtain low-pass filtered images of the internal image and the external image specifically comprises the following steps:
[0044] Smoothing the internal image and the external image to obtain an internal smoothed image and an external smoothed image;
[0045] Screening the pixel values of the internal smooth image and the appearance smooth image;
[0046] A grayscale threshold is set. If the pixel value of the internal smooth image and the appearance smooth image is greater than the grayscale threshold, the pixel is eliminated. If the pixel value of the internal smooth image and the appearance smooth image is less than the grayscale threshold, the pixel value is retained and the retained pixel values are integrated to obtain a low-pass filtered image.
[0047] Compared with the prior art, the present invention provides a multi-dimensional image fusion method based on an X-ray machine and an industrial camera, which has the following beneficial effects:
[0048] Firstly, by fusing X-ray images and industrial camera images, the present invention can obtain internal and external information of the object to be measured in a comprehensive image, thereby realizing more comprehensive detection. Secondly, by means of a calibration and calibration process, the present invention geometrically aligns the X-ray machine image and the industrial camera image, thereby reducing the difference between the images and improving the accuracy of detection. Furthermore, the present invention can comprehensively utilize the information of the X-ray machine image and the industrial camera image, thereby making the fused image more comprehensive and detailed, and providing more information on the characteristics and status of the object to be measured. Furthermore, by fusing the two images, the present invention realizes one-time processing and analysis, thereby reducing repeated operations and improving detection efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0049] Figure 1 This is a flow chart of steps S100-S600 in a multi-dimensional image fusion method based on an X-ray machine and an industrial camera proposed by the present invention;
[0050] Figure 2 This is a flow chart of steps S201-S203 in a multi-dimensional image fusion method based on an X-ray machine and an industrial camera proposed in the present invention;
[0051] Figure 3 This is a flow chart of steps S2031-S2035 in a multi-dimensional image fusion method based on an X-ray machine and an industrial camera proposed in the present invention;
[0052] Figure 4 This is a flow chart of steps S20341-S20343 in a multi-dimensional image fusion method based on an X-ray machine and an industrial camera proposed in the present invention;
[0053] Figure 5 This is a flow chart of steps S501-S503 in a multi-dimensional image fusion method based on an X-ray machine and an industrial camera proposed in the present invention;
[0054] Figure 6 This is a flow chart of steps S5011-S5013 in a multi-dimensional image fusion method based on an X-ray machine and an industrial camera proposed in the present invention. DETAILED DESCRIPTION
[0055] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.
[0056] Reference Figure 1 As shown, a multi-dimensional image fusion method based on an X-ray machine and an industrial camera includes:
[0057] S100, setting a hollow aluminum plate as a calibration plate, and collecting several groups of calibration plate images in different fields of view under an X-ray machine and an industrial camera respectively;
[0058] S200, analyzing and processing the calibration plate images in different fields of view of several groups of calibration plates to obtain calibration parameters between the X-ray machine and the industrial camera, wherein the calibration parameters include an offset x, an offset y, a telescopic ratio in the x direction, a telescopic ratio in the y direction, and an angle difference;
[0059] S300, obtaining an internal image of the inspection object under an X-ray machine;
[0060] S400, obtaining an appearance image of the inspection object under the industrial camera;
[0061] S500, preprocessing the internal image and the appearance image to obtain a registered image of the internal image and a registered image of the appearance image;
[0062] S600: Perform a weighted average algorithm on the registered image of the internal image and the registered image of the external image to obtain a fused image.
[0063] In this embodiment, a hollow aluminum plate with an outer contour length of 300mm, a width of 200mm, an inner contour length of 150mm, and a width of 100mm is used as a calibration plate to ensure clear imaging under both the X-ray machine and the industrial camera. In addition, in the industrial production process, the internal image and the appearance image of the object to be measured are collected at the same time, and the camera trigger interval is determined according to the number of encoder pulses, the camera installation position and the corresponding field of view distance. It is determined that the photographing field of the X-ray machine and the industrial camera is at the same position to ensure that the calibration data when the image is fused is consistent in space. Assume that the X-ray camera is placed in front of the industrial camera, that is, the X-ray machine first photographs the incoming material and the encoder outputs a pulse corresponding to the distance traveled by the belt line is P. The distance between the X-ray machine and the industrial camera is measured as S. It can be seen that the image pulse difference δ between the X-ray and the industrial camera is S / P. When the belt line starts to move, the X-ray camera starts to shoot, and the industrial camera triggers the shooting after delaying δ pulses. In this way, the position photographed by the industrial camera and the X-ray camera is the same position of the belt line.
[0064] Reference Figure 2 As shown, analyzing and processing the calibration plate images in different fields of view of several groups of calibration plates to obtain the calibration parameters between the X-ray machine and the industrial camera specifically includes the following steps:
[0065] S201, selecting two images of an X-ray machine calibration plate and an industrial camera calibration plate having the same pixel information and environmental information from a plurality of groups of images;
[0066] S202, establishing a rectangular coordinate system, and placing the two calibration plate images into different rectangular coordinate systems;
[0067] S203, extracting and calculating the images of the two calibration plates to obtain calibration parameters between the X-ray machine and the industrial camera.
[0068] Reference Figure 3 As shown, extracting and calculating the two calibration plate images to obtain the calibration parameters between the X-ray machine and the industrial camera specifically includes the following steps:
[0069] S2031, performing edge threshold extraction processing on the two calibration plate images to obtain the coordinates of the corner points of the edges of the two calibration plate images;
[0070] S2032, respectively calculating the coordinates of the corner points of the edges adjacent to each other in the x-axis and y-axis directions to obtain the Euclidean distance between the corner points of the edges in the two calibration plate images;
[0071] S2033, respectively calculating and processing the Euclidean distances of the two calibration plate images to obtain the scaling ratios in the x direction and the scaling ratios in the y direction of the two calibration plate images;
[0072] S2034, performing scaling processing on the images of the two calibration plates according to the scaling ratio in the x direction and the scaling ratio in the y direction to obtain an angle difference;
[0073] S2035. Perform rotation transformation and calculation on the two calibration plate images according to the angle difference to obtain the offset x and the offset y.
[0074] The calculation formula of Euclidean distance is:
[0075]
[0076]
[0077] Where, d x1 is the Euclidean distance in the x-axis direction, (x1, y1) is the coordinate of the corner point of the edge in the calibration plate image under the X-ray machine, and (x2, y2) is the coordinate of the corner point of the adjacent edge in the calibration plate image under the X-ray machine; d y1 is the Euclidean distance in the y-axis direction, (x3, y3) is the coordinate of the corner point of the edge in the X-ray machine calibration plate image, (x4, y4) is the coordinate of the corner point of the adjacent edge in the y-axis direction in the X-ray machine calibration plate image; similarly, d x2 d y2 .
[0078] The calculation formulas for the expansion ratio in the x direction and the expansion ratio in the y direction are:
[0079] scale_x=d x1 / d x2 ;
[0080] scale_y=d y1 / d y2 ;
[0081] Where scale_x is the scaling ratio in the x-axis direction, d x1 and d x2 is the Euclidean distance in the x-axis direction; scale_y is the scaling ratio in the y-axis direction, d y1 and d y2 is the Euclidean distance in the y-axis direction.
[0082] Reference Figure 4 As shown, the images of the two calibration plates are scaled according to the scaling ratio in the x direction and the scaling ratio in the y direction, and obtaining the angle difference specifically includes the following steps:
[0083] S20341, performing scaling processing on the two calibration plate images according to the scaling ratio in the x direction and the scaling ratio in the y direction to obtain scaled images of the two calibration plates;
[0084] S20342, respectively calculating the distance between the upper edge of the calibration plate and the x-axis of the image in the two telescopic images to obtain two sets of angles;
[0085] S20343. Calculate and process the two sets of included angles to obtain the angle difference.
[0086] The angle difference is calculated as:
[0087] angle_x=atan((x1-x2) / (y1-y2));
[0088] angle_2d=atan((x1-x2) / (y1-y2));
[0089] angle=angle x -angle-2d;
[0090] Where angle_x is the angle between the upper edge of the calibration plate and the x-axis of the image in the telescopic image under the X-ray camera, angle_2d is the angle between the upper edge of the calibration plate and the x-axis of the image in the telescopic image under the industrial camera, and angle is the angle difference between angle_x and angle-2d.
[0091] In this embodiment, the four edges of the two calibration plates are extracted by the calibration plate edge threshold, and the corner points p1, p2, p3, and p4 of the adjacent edges are calculated to respectively calculate the Euclidean distance between the two images p1 and p2, and the Euclidean distance between p2 and p3. The scaling ratios scale_x and scale_y in the x and y directions are obtained by the ratio between the distances. The image is scaled according to the obtained scaling ratio, and the angle angle between the upper edge of the calibration plate and the X-axis of the image is calculated. The calibration plate image is then rotated according to the obtained angle, and the upper left corner p1 of the calibration plate image is used as the reference point. The difference between the transformed corner point and the original image p1 is calculated to obtain the offset x and the offset y.
[0092] Reference Figure 5 As shown, preprocessing the internal image and the appearance image to obtain the registered image of the internal image and the registered image of the appearance image specifically includes the following steps:
[0093] S501, performing dynamic threshold segmentation on the internal image and the external image respectively to obtain low-pass filtered images of the internal image and the external image;
[0094] S502, performing grayscale processing on the low-pass filtered images of the internal image and the external image to obtain two sets of detection object mask images;
[0095] S503 , registering the two sets of detection object mask images using calibration parameters to obtain a registered image.
[0096] Reference Figure 6 As shown, dynamic threshold segmentation is performed on the internal image and the external image respectively, and the low-pass filtered images of the internal image and the external image are obtained, which specifically includes the following steps:
[0097] S5011, smoothing the internal image and the external image to obtain an internal smoothed image and an external smoothed image;
[0098] S5012, filtering the pixel values of the internal smooth image and the appearance smooth image;
[0099] S5013, setting a grayscale threshold. If the pixel value of the internal smooth image and the appearance smooth image is greater than the grayscale threshold, the pixel is eliminated. If the pixel value of the internal smooth image and the appearance smooth image is less than the grayscale threshold, the pixel value is retained, and the retained pixel values are integrated to obtain a low-pass filtered image.
[0100] In this embodiment, the internal image captured by the X-ray machine and the appearance image captured by the industrial camera are preprocessed, mainly including: 2D image edge extraction, acquiring the edge features of the target by collecting the 2D image background image and the image with the target object. Since the background color may be light or dark, a dynamic threshold segmentation method is proposed to extract the target object from the non-uniform background, perform a smoothing process on the appearance image, and use the appearance smoothed image obtained after the smoothing process as the reference image. By setting the grayscale threshold t, a low-pass filtered image of the appearance image and the appearance smoothed image is obtained, wherein the size (diameter) of the object is determined by the mask size of the low-pass filter and the amplitude of the edge of the object. The grayscale of the internal connected area of the target object is set to 0, and the grayscale outside the area is set to 255, to obtain a mask map of the target object. The calibration parameters are used to register the two images, specifically including: generating a unit transformation matrix, rotating, translating and scaling the unit matrix according to the calibration parameters to obtain a new transformation matrix, and then transforming the internal image according to the transformation matrix to obtain the registered internal image, and applying a weighted average algorithm to the registered internal image to fuse the internal image with the appearance image to obtain a fused image M(x,y), which is specifically characterized by: first setting the fusion transparency α, then M(x,y)=α*X(x,y)+(1-α)*offset(x,y). The fusion algorithm can use the internal density information of the X-ray image and the surface defect information of the industrial camera image to comprehensively analyze and generate a fused image, providing more comprehensive internal and external information.
[0101] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.
Claims
1. A multi-dimensional image fusion method based on X-ray machine and industrial camera, It is characterized in that include: A hollow aluminum plate is set as a calibration plate, and several groups of calibration plate images in different fields of view are collected under an X-ray machine and an industrial camera. Analyze and process the calibration plate images in different fields of view of several groups of calibration plates to obtain calibration parameters between the X-ray machine and the industrial camera, wherein the calibration parameters include an offset x, an offset y, a telescopic ratio in the x direction, a telescopic ratio in the y direction, and an angle difference; Obtain the internal image of the inspection object under the X-ray machine; Obtain the appearance image of the detected object under the industrial camera; Preprocessing the internal image and the appearance image to obtain a registration image of the internal image and a registration image of the appearance image; A weighted average algorithm is performed according to the registered image of the internal image and the registered image of the appearance image to obtain a fused image.
2. According to claim 1, a multi-dimensional image fusion method based on an X-ray machine and an industrial camera, It is characterized in that The step of analyzing and processing the calibration plate images in different fields of view of the calibration plates to obtain the calibration parameters between the X-ray machine and the industrial camera specifically includes the following steps: Select two images of the X-ray machine calibration plate and the industrial camera calibration plate with the same pixel information and environmental information from a plurality of groups of images; Establish a rectangular coordinate system and place the two calibration plate images into different rectangular coordinate systems; The images of the two calibration plates are extracted and processed to obtain the calibration parameters between the X-ray machine and the industrial camera.
3. According to claim 2, a multi-dimensional image fusion method based on an X-ray machine and an industrial camera, It is characterized in that The extraction and calculation of the two calibration plate images to obtain the calibration parameters between the X-ray machine and the industrial camera specifically includes the following steps: Perform edge threshold extraction processing on the two calibration plate images to obtain the coordinates of the corner points of the edges of the two calibration plate images; The coordinates of the corner points of the two adjacent edges are calculated in the x-axis and y-axis directions respectively to obtain the Euclidean distance between the corner points of the edges in the two calibration plate images; The Euclidean distances of the two calibration plate images are calculated and processed respectively to obtain the scaling ratios in the x direction and the scaling ratios in the y direction of the two calibration plate images; The images of the two calibration plates are scaled according to the scaling ratio in the x direction and the scaling ratio in the y direction to obtain the angle difference; According to the angle difference, the two calibration plate images are rotated and calculated to obtain the offset x and offset y.
4. According to claim 3, a multi-dimensional image fusion method based on an X-ray machine and an industrial camera, It is characterized in that The step of performing scaling processing on the images of the two calibration plates according to the scaling ratio in the x direction and the scaling ratio in the y direction to obtain the angle difference specifically includes the following steps: Performing scaling processing on the two calibration plate images according to the scaling ratio in the x direction and the scaling ratio in the y direction to obtain scaled images of the two calibration plates; The distances between the upper edge of the calibration plate and the x-axis of the image in the two telescopic images are calculated and processed respectively to obtain two sets of angles; The two sets of included angles are calculated and processed to obtain the angle difference.
5. According to claim 3, a multi-dimensional image fusion method based on an X-ray machine and an industrial camera, It is characterized in that The calculation formula of the Euclidean distance is: Where, d x1 is the Euclidean distance in the x-axis direction, (x 1 ,y 1 ) is the coordinate of the corner point of the edge in the image of the calibration plate under the X-ray machine, (x 2 ,y 2 ) is the coordinate of the corner points of the adjacent edges in the x-axis direction in the calibration plate image under the X-ray machine; d y1 for y Euclidean distance in the axis direction, (x 3 ,y 3 ) is the coordinate of the corner point of the edge in the image of the calibration plate under the X-ray machine, (x 4 ,y 4 ) is the image of the calibration plate under the X-ray machine y The coordinates of the corner points of the adjacent sides in the axis direction; similarly, d x2 d y2 .
6. The multi-dimensional image fusion method based on an X-ray machine and an industrial camera according to claim 3, It is characterized in that The calculation formulas for the expansion ratio in the x direction and the expansion ratio in the y direction are: scale_x=d x1 / d x2 ; scale_y=d y1 / d y2 ; Where scale_x is the scaling ratio in the x-axis direction, d x1 and d x2 is the Euclidean distance in the x-axis direction; scale_y is the scaling ratio in the y-axis direction, d y1 and d y2 is the Euclidean distance in the y-axis direction.
7. The multi-dimensional image fusion method based on an X-ray machine and an industrial camera according to claim 3, It is characterized in that The calculation formula of the angle difference is: angle_x=atan((x 1 -x 2 ) / (and 1 -and 2 )); angle_2d=atan((x 1 -x 2 ) / (and 1 -and 2 )); angle=angle x -angle_2d; Where angle_x is the angle between the upper edge of the calibration plate and the x-axis of the image in the telescopic image under the X-ray camera, angle_2d is the angle between the upper edge of the calibration plate and the x-axis of the image in the telescopic image under the industrial camera, and angle is the angular difference between angle_x and angle_2d.
8. The multi-dimensional image fusion method based on an X-ray machine and an industrial camera according to claim 1, It is characterized in that The preprocessing of the internal image and the appearance image to obtain the edge features of the internal image and the edge features of the appearance image specifically includes the following steps: Perform dynamic threshold segmentation on the internal image and the external image respectively to obtain low-pass filtered images of the internal image and the external image; Perform grayscale processing on the low-pass filtered images of the internal image and the external image to obtain two sets of detection object mask images; The two sets of detection object mask images are registered through calibration parameters to obtain the registered image.
9. The multi-dimensional image fusion method based on an X-ray machine and an industrial camera according to claim 3, It is characterized in that The step of performing dynamic threshold segmentation on the internal image and the external image to obtain low-pass filtered images of the internal image and the external image specifically includes the following steps: Smoothing the internal image and the external image to obtain an internal smoothed image and an external smoothed image; Screening the pixel values of the internal smooth image and the appearance smooth image; A grayscale threshold is set. If the pixel value of the internal smooth image and the appearance smooth image is greater than the grayscale threshold, the pixel is eliminated. If the pixel value of the internal smooth image and the appearance smooth image is less than the grayscale threshold, the pixel value is retained and the retained pixel values are integrated to obtain a low-pass filtered image.
