A method and system for fusing visible light image and fluorescence image

Abstract

The present invention discloses a method and system for fusing visible light images and fluorescence images, and relates to the field of medical image processing technology. The method comprises: performing image registration on a visible light camera and a near-infrared camera to obtain registration parameters; synchronously acquiring visible light images and fluorescence images, preprocessing the received fluorescence images, and aligning the pixels of the visible light images and the preprocessed fluorescence images according to the registration parameters; performing pixel fusion on the aligned visible light images and fluorescence images to generate a multispectral image; and sequentially merging the fused multispectral images to generate a corresponding video sequence. The present invention adopts offline camera registration, and utilizes the registration parameters to efficiently and accurately align visible light images and fluorescence images. By enhancing the preprocessing, a high-quality fluorescence image can be generated. The fusion weight is calculated based on the edge intensity value and the fluorescence intensity value, so that the weighted fused multispectral image can simultaneously maintain edge and fluorescence intensity information, and is suitable for medical equipment such as fluorescence microscopes / endoscopes.

Description

Technical Field

[0001] The present invention belongs to the technical field of medical image processing, and more specifically, relates to a visible light image and fluorescence image fusion method and system. Background Art

[0002] Currently, most medical imaging systems utilize color visible light sensors with CMOS / CCD structures. With the rapid advancement of photosensitive element manufacturing processes and imaging technologies, the resolution and quality of visible light images now meet the medical community's universal requirements for high-quality imaging. However, color visible light cameras only produce visible light images and cannot capture information outside the visible spectrum. In contrast, near-infrared cameras can capture and image signals outside the visible wavelength range. By combining them with dyes such as indocyanine green (ICG), which are absorbed by biological tissue and can be excited by lasers of specific wavelengths to generate invisible light, near-infrared imaging can reveal details that are difficult to clearly observe with visible light, such as blood vessels, polyps, tumors, and other lesions that are difficult to distinguish under visible light. Therefore, the fusion of visible light and fluorescence imaging has become a key development in the medical device industry in recent years. This technology can be widely applied in biological clinical surgery. Simultaneous multispectral imaging can provide doctors with a vast amount of pathological information, possessing significant practical value in assisting doctors in making accurate clinical diagnoses and improving surgical efficiency and success rates. However, the current visible light and fluorescence image fusion technology faces several difficulties: (1) The medical community has strict requirements on the authenticity and accuracy of imaging results. Therefore, image fusion must be based on strict pixel alignment. Visible light images and fluorescence images are usually collected from different cameras. If the images are not accurately aligned before fusion, it may lead to imaging errors and stroboscopic images. (2) The fluorescence signal collected by the near-infrared camera is very weak and has interference signals. It cannot be directly fused with the visible light image and requires appropriate enhancement preprocessing. (3) In the process of weighted image fusion, the setting of weights is very critical to the imaging quality. Inappropriate weight setting may lead to problems such as low brightness, blurred edges, and difficulty in distinguishing the strength of the fluorescence signal. Summary of the Invention

[0003] In response to the above-mentioned defects or improvement needs of the prior art, the present invention proposes a method and system for fusing visible light images and fluorescence images, which can generate high-quality multispectral images that can carry rich pathological information.

[0004] To achieve the above object, according to one aspect of the present invention, a method for fusing a visible light image and a fluorescent image is provided, comprising:

[0005] S1: Perform image registration on the visible light camera and the near-infrared camera to obtain registration parameters;

[0006] S2: Synchronously collect visible light images and fluorescence images using a visible light camera and a near-infrared camera, respectively;

[0007] S3: preprocessing the received fluorescence image, aligning the pixels of the visible light image and the preprocessed fluorescence image according to the registration parameters;

[0008] S4: Pixel fusion of the registered and aligned visible light image and fluorescence image to generate a multispectral image;

[0009] S5: Merge the fused multispectral images in sequence to generate corresponding video sequences.

[0010] In some optional embodiments, step S1 includes:

[0011] S11: During the registration process, the visible light camera and the near-infrared camera are placed at fixed positions, and the relative positions between the visible light camera and the near-infrared camera remain fixed;

[0012] S12: The target is placed on an adjustable metal slide rail, directly in front of the visible light camera and the near-infrared camera. The target can emit fluorescence within the near-infrared camera band.

[0013] S13: Synchronously acquire visible light images and fluorescence images using a visible light camera and a near-infrared camera;

[0014] S14: Obtain corresponding feature points and their coordinate positions of the visible light image and the fluorescent image, wherein the coordinates of the corresponding feature points in the visible light image and the fluorescent image constitute feature pairs, and calculate a set of registration parameters based on multiple sets of feature pair coordinates, wherein the registration parameters are pixel offset values or affine transformation matrices;

[0015] S15: Adjust the distance between the target and the visible light camera and the near-infrared camera at a fixed length, and execute steps S13 to S14 in sequence to obtain multiple sets of registration parameters;

[0016] S16: performing data cleaning on the multiple groups of registration parameters obtained, discarding abnormal registration parameter groups that deviate from the overall average value and whose differences exceed the set value, and calculating the average value of each parameter in the remaining registration parameter groups as the final registration parameter group.

[0017] The registration process in step S1 only needs to be performed once, and the relevant image fusion module can realize real-time image alignment by reading and using the registration parameters before executing the visible light image and fluorescence image fusion stage.

[0018] In some optional embodiments, in step S3, preprocessing the received fluorescence image includes: performing denoising, intensity transformation, data nonlinear mapping, and pseudo-colorization preprocessing on the received fluorescence image in sequence.

[0019] In some optional embodiments, the image fusion process in step S4 is specifically a weighted summation of corresponding pixels of the visible light image and the fluorescence image in a specific color space, wherein the weight parameter is calculated and determined by the pixel edge intensity value of the visible light image and the pixel intensity value of the fluorescence image, and the color space includes but is not limited to RGB, YCbCr, HSV, CIELAB, etc.

[0020] In some optional embodiments, in step S3, preprocessing the received fluorescence image includes:

[0021] An edge-preserving filter is used to smooth and denoise the fluorescence image. A set fluorescence signal intensity threshold is used to distinguish between the fluorescence signal and the interference signal. The fluorescence signal is linearly enhanced according to the set intensity amplification factor, and the intensity value of the interference signal is set to 0. The intensity value of the fluorescence pixel is nonlinearly mapped to the pixel depth range of the visible light image. A three-channel pseudo-color image is generated from the single-channel fluorescence image, and the chromaticity of the pseudo-color image is specified by the user.

[0022] Specifically, the fluorescent signal and the interference signal are distinguished by setting the fluorescent signal intensity threshold, including:

[0023] A fluorescence signal intensity threshold is set. Pixels with signal intensity higher than the set fluorescence signal intensity threshold are considered as fluorescence signals, and pixels with signal intensity lower than the set fluorescence signal intensity threshold and not equal to 0 are considered as interference signals.

[0024] In some optional embodiments, I'' i,j =w′ i,j I′ i,j +w″ i,j I″ i,j Determine the pixel I″′ after image fusion i,j , where w′ i,j is the visible light image pixel I′ after registration and alignment i,j The fusion weight value, w″ i,j is the fluorescence image pixel I″ after registration and alignment i,j The fusion weight value of , (i, j) is the subscript of the pixel.

[0025]

[0026] In some optional embodiments, by w″ i,j =1-w′ i,j Make sure it is visible

[0027]

[0028] The fusion weight value w′ of the light image i,j and the fusion weight value w″ of the fluorescence image i,j, where Ω is the set of all pixel subscripts belonging to the fluorescent area, p i,j ∈[0,1] is the edge intensity value of the visible light image pixel, q i,j ∈[0,1] is the intensity value of the fluorescence image pixel. Optionally, the edge intensity value of the visible light pixel can be calculated by edge detection operators such as Laplace, Sobel and Canny.

[0029] According to another aspect of the present invention, a visible light image and fluorescence image fusion system is provided, comprising: an image acquisition system and a software system, wherein the image acquisition system comprises a visible light camera for acquiring visible light images and a near-infrared camera for acquiring fluorescence images; the software system comprises an image processing module, an image fusion module, and a video sequence generation module;

[0030] The image processing module is used to perform image registration on the visible light camera and the near-infrared camera of the image acquisition system to obtain registration parameters;

[0031] The image acquisition system is used to synchronously acquire visible light images and fluorescence images through a visible light camera and a near-infrared camera respectively, and synchronously input the acquired visible light images and fluorescence images into an image processing module of the software system;

[0032] The image processing module is further configured to pre-process the received fluorescence image and align pixels of the visible light image and the pre-processed fluorescence image according to the registration parameters;

[0033] The image fusion module is used to perform pixel fusion on the registered and aligned visible light and fluorescence images to generate a multispectral image;

[0034] The video sequence generation module is used to sequentially merge the fused multispectral images to generate corresponding video sequences.

[0035] In some optional embodiments, the registration of the visible light camera and the near infrared camera includes:

[0036] During the registration process, the visible light camera and the near-infrared camera of the image acquisition system are placed in fixed positions, and the relative positions between the visible light camera and the near-infrared camera remain fixed;

[0037] The target is placed on an adjustable metal slide rail, directly in front of the camera. The target can emit fluorescence within the near-infrared camera band.

[0038] Visible light images and fluorescence images are synchronously acquired through an image acquisition system, and the acquired visible light images and fluorescence images are synchronously input into an image processing module of a software system;

[0039] The feature detection algorithm of the image processing module is used to obtain the corresponding feature points and their coordinate positions of the visible light image and the fluorescence image. The coordinates of the corresponding feature points in the visible light image and the fluorescence image form feature pairs. A set of registration parameters is calculated based on the coordinates of multiple feature pairs.

[0040] Adjust the distance between the target and the visible light camera and the fluorescence camera by a fixed length to obtain multiple sets of registration parameters;

[0041] The obtained parameter groups are cleaned, and abnormal parameter groups that deviate from the overall average value and have large differences are discarded. The average values of the parameters in the remaining parameter groups are calculated as the final registration parameter group and written into the register of the image fusion module for table lookup during the registration process.

[0042] Among them, the sensor of the visible light camera is CMOS or CCD, the color type is color or monochrome, and the near-infrared camera has good responsiveness in the near-infrared band. For targets with geometric shapes, Harris, Surf, Sift and other detection algorithms can be used to obtain feature points. The specific detection algorithm varies according to the characteristics of the target and will not be listed one by one.

[0043] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0044] The present invention comprises the following steps: S1: performing image registration on the visible light camera and near-infrared camera of the image acquisition system to obtain registration parameters; S2: synchronously acquiring visible light images and fluorescence images through the image acquisition system and synchronously inputting them into the image processing module of the software system; S3: the image processing module preprocessing the received fluorescence image and aligning the pixels of the visible light image and the preprocessed fluorescence image according to the registration parameters; S4: performing pixel fusion on the aligned visible light image and fluorescence image to generate a multispectral image; and S5: sequentially merging the fused multispectral images to generate a corresponding video sequence. The present invention employs offline camera registration and utilizes the registration parameters to efficiently and accurately align the visible light image and fluorescence image. Enhanced preprocessing can generate a high-quality fluorescence image. Fusion weights are calculated based on edge intensity values and fluorescence intensity values, so that the weighted fused multispectral image can simultaneously maintain edge and fluorescence intensity information, making it suitable for medical equipment such as fluorescence microscopes and endoscopes. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 This is a flowchart of white light image and fluorescence image registration of an endoscopic image acquisition system provided by an embodiment of the present invention;

[0046] Figure 2 This is a flowchart of a white light image and fluorescence image fusion provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0047] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0048] In the examples of the present invention, “first”, “second”, etc. are used to distinguish different objects rather than to describe a specific order or sequence.

[0049] The embodiment of the present invention provides a method for fusing white light images and fluorescence images in a fluorescence endoscope. The specific process is as follows: Figure 1 As shown, this embodiment includes the following steps:

[0050] S1: Perform image registration on the white light camera and the fluorescence camera in the endoscope image acquisition system to obtain registration parameters;

[0051] S2: Synchronously capture white light images and fluorescence images through the white light camera and near-infrared camera of the image acquisition system, and input them into the image processing module of the software system;

[0052] S3: The image processing module preprocesses the fluorescence image and aligns the pixels of the white light image and the preprocessed fluorescence image according to the registration parameters;

[0053] S4: Pixel fusion of the registered and aligned white light image and fluorescence image to generate a multispectral image;

[0054] S5: Merge the fused multispectral images in sequence to generate corresponding video sequences.

[0055] In this embodiment, the image acquisition system in step S1 includes: a white light camera and a fluorescence camera for respectively acquiring white light signals and fluorescence signals. The white light camera is a color CMOS camera. The fluorescence camera has good responsivity in the near-infrared band, covering the 700-900 nm band of fluorescence excited by fluorescent reagents; a specially designed registration target that emits invisible light within the near-infrared camera band; multiple groups of LEDs emitting specified fluorescence wavelengths, each within the fluorescence camera band, are evenly spaced on the target. The wavelengths are within the fluorescence camera band, and selecting a band with a high fluorescence camera responsivity is a preferred embodiment; and a software system with an algorithm for detecting feature points in the target's fluorescent area and calculating calibration parameters.

[0056] Furthermore, the registration of the white light camera image and the fluorescence camera image in step S1 includes the following steps:

[0057] S11: During the registration process, the cameras of the endoscope image acquisition system are placed in a fixed position as a whole, and the relative positions between the white light camera and the near-infrared camera are kept fixed;

[0058] S12: Place the target on an adjustable metal slide rail, directly in front of the camera;

[0059] S13: Synchronously acquiring white light images and fluorescence images through an image acquisition system, and synchronously inputting the acquired images into an image processing module of a software system;

[0060] S14: Using the feature detection algorithm of the image processing module to obtain the coordinate positions of the LED feature points in the white light and fluorescent images, the corresponding feature point coordinates of the two images form feature pairs, and a set of registration parameters is calculated based on the multiple sets of feature pair coordinates using the random sample consensus algorithm (RANSAC);

[0061] S15: Adjust the distance between the target and the camera at a fixed length, and execute steps S13-S14 in sequence to obtain multiple sets of registration parameters;

[0062] S16: Perform data cleaning on the obtained parameter groups, discard abnormal parameter groups that deviate from the overall average value and have large differences, calculate the average value of each parameter in the remaining parameter groups as the final registration parameter group, and write it into the register of the image fusion module for table lookup during the registration process.

[0063] Furthermore, the registration process in the above step S1 only needs to be performed once, and the relevant image fusion module can read and use the registration parameters before performing the visible light and fluorescence image fusion stage to achieve real-time image alignment.

[0064] Furthermore, the fluorescence image preprocessing in the above step S3 includes: performing denoising, intensity transformation, data nonlinear mapping and pseudo-colorization processing on the collected fluorescence image in sequence.

[0065] Furthermore, the fluorescence image preprocessing process in step S3 above specifically includes the following steps:

[0066] S31: Anisotropic diffusion filtering is used to smooth and denoise the fluorescence image. The filtering formula is as follows:

[0067]

[0068] Among them, cN x,y ,cS x,y ,cE x,y ,cW x,y is the diffusion coefficient in each direction, is the image gradient value in each direction, K is the thermal conductivity coefficient, and t is the number of iterations. Users can adjust the parameters K, t, and λ to control the intensity and smoothness of fluorescence image denoising.

[0069] S32: Thresholding and enhancing the fluorescence image intensity; the user sets the fluorescence signal intensity threshold. Pixels with signal intensities higher than the set fluorescence signal intensity threshold are considered fluorescence signals. The fluorescence signal is linearly enhanced according to the intensity amplification factor set by the user. Pixels with signal intensities lower than the set fluorescence signal intensity threshold and not equal to 0 are considered interference signals, and the interference signal intensity value is set to 0.

[0070] S33: nonlinearly mapping the intensity value of the fluorescence pixel to the pixel bit depth range of the white light image. Taking a white light image with a bit depth of 8 bits as an example, the fluorescence image pixel value needs to be mapped to the interval [0, 255]. The maximum fluorescence intensity value can be preset as a reference value and a gamma formula can be used to nonlinearly map the fluorescence pixel to the target interval.

[0071] S34: Generate a three-channel pseudo-color image from a single-channel fluorescence image, and the chromaticity of the pseudo-color image is specified by the user; specifically, the user can specify that the pixels of a certain channel RGB of the pseudo-color image are generated by the pixel intensity of the fluorescence image, and the pixels of the remaining two channels are set to zero, so as to obtain a red, blue, and green pure color pseudo-color image, in which case the chromaticity value corresponds to the fluorescence intensity value one-to-one; in addition, the user can also set a composite color pseudo-color image, that is, the color of the target area presents a gradual change as the fluorescence signal intensity changes, and the fluorescence intensity value corresponds to the chromaticity values of multiple colors.

[0072] Further, such as Figure 2 As shown, the image fusion in the above step S4 is a weighted summation of corresponding pixels of the white light image and the fluorescence image in a specific color space, wherein the weight parameter is calculated and determined by the edge intensity value of the white light image pixel and the pixel intensity value of the fluorescence image. The color space includes but is not limited to RGB, YCbCr, HSV, CIELAB, etc.

[0073] Furthermore, the fusion pixel calculation method in the white light image and fluorescence image fusion method in step S4 above takes the YCbCr color space as an example, and the calculation formula is:

[0074]

[0075] Among them, {Y′ i,j ,Cb′ i,j ,Cr′ i,j} and {Y″ i,j ,Cb″′ i,j ,C″r i,j} represent the pixel values of white light image and fluorescence image in YCbCr color space, respectively, {Y″′ i,j,Cb″′ i,j ,Cr″′ i,j} represents the pixel value after fusion, w′ i,j and w″′ i,j are the fusion weight values of white light image pixels and fluorescence image pixels respectively, and (i, j) is the subscript of the pixel.

[0076] Furthermore, in the white light image and fluorescence image fusion algorithm in step S4 above, the fusion weight parameter of the white light image and the fluorescence image is calculated by the following formula:

[0077]

[0078] Where Ω is the set of all pixel subscripts belonging to the fluorescent area, p i,j ∈[0,1] is the edge intensity value of the white light pixel, q i,j ∈[0,1] is the intensity value of the fluorescent pixel. The edge intensity value of the white light pixel can be obtained by convolving the grayscale image of the white light image with the following Laplace edge detection operator template:

[0079] or

[0080] The above-described embodiment takes the fusion of white light images and fluorescence images in a fluorescence endoscope as an example, but the technical solution described in the present invention is not limited to endoscopic images. Other scenarios that require fusion of visible light images and fluorescence images can also be image fused through the technical solution described in the present invention as long as they meet the characteristics of the images processed by this technical solution.

[0081] It should be pointed out that, according to the needs of implementation, the various steps / components described in this application can be split into more steps / components, or two or more steps / components or partial operations of steps / components can be combined into new steps / components to achieve the purpose of the present invention.

[0082] It will be easily understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for fusing visible light images and fluorescent images, characterized in that: include: S1: Perform image registration on the visible light camera and the near-infrared camera to obtain registration parameters; S2: Synchronously collect visible light images and fluorescence images using a visible light camera and a near-infrared camera, respectively; S3: preprocessing the received fluorescence image, aligning the pixels of the visible light image and the preprocessed fluorescence image according to the registration parameters; S4: Pixel fusion of the registered and aligned visible light image and fluorescence image to generate a multispectral image; S5: Merge the fused multispectral images in sequence to generate corresponding video sequences; The image fusion process in step S4 is specifically to perform weighted summation of corresponding pixels of the visible light image and the fluorescence image in a specific color space, wherein the weight parameter is calculated and determined by the pixel edge intensity value of the visible light image and the pixel intensity value of the fluorescence image; by I″′ i,j =w′ i,j I′ i,j +w″ i,j I″ i,j Determine the pixel I″′ after image fusion i,j , where w′ i,j is the visible light image pixel I′ after registration and alignment i,j The fusion weight value, w″ i,j is the fluorescence image pixel I″ after registration and alignment i,j The fusion weight value, (i, j) is the subscript of the pixel; Depend on Determine the fusion weight value w′ of the visible light image i,j and the fusion weight value w″ of the fluorescence image i,j , where Ω is the set of all pixel subscripts belonging to the fluorescent area, p i,j ∈[0,1] is the edge intensity value of the visible light image pixel, q i,j ∈[0,1] is the intensity value of the fluorescence image pixel.

2. The method according to claim 1, characterized in that Step S1 includes: S11: During the registration process, the visible light camera and the near-infrared camera are placed at fixed positions, and the relative positions between the visible light camera and the near-infrared camera remain fixed; S12: The target is placed on an adjustable metal slide rail, directly in front of the visible light camera and the near-infrared camera. The target can emit fluorescence within the near-infrared camera band. S13: Synchronously acquire visible light images and fluorescence images using a visible light camera and a near-infrared camera; S14: Obtain corresponding feature points and their coordinate positions of the visible light image and the fluorescent image, wherein the coordinates of the corresponding feature points in the visible light image and the fluorescent image constitute feature pairs, and calculate a set of registration parameters based on multiple sets of feature pair coordinates, wherein the registration parameters are pixel offset values or affine transformation matrices; S15: Adjust the distance between the target and the visible light camera and the near-infrared camera at a fixed length, and execute steps S13 to S14 in sequence to obtain multiple sets of registration parameters; S16: performing data cleaning on the multiple groups of registration parameters obtained, discarding abnormal registration parameter groups that deviate from the overall average value and whose differences exceed the set value, and calculating the average value of each parameter in the remaining registration parameter groups as the final registration parameter group.

3. The method according to claim 1 or 2, characterized in that In step S3, preprocessing the received fluorescence image includes: performing denoising, intensity transformation, data nonlinear mapping and pseudo-colorization preprocessing on the received fluorescence image in sequence.

4. The method according to claim 3, characterized in that In step S3, preprocessing the received fluorescence image includes: An edge-preserving filter is used to smooth and denoise the fluorescence image. A set fluorescence signal intensity threshold is used to distinguish between the fluorescence signal and the interference signal. The fluorescence signal is linearly enhanced according to the set intensity amplification factor, and the intensity value of the interference signal is set to 0. The intensity value of the fluorescence pixel is nonlinearly mapped to the pixel depth range of the visible light image. A three-channel pseudo-color image is generated from the single-channel fluorescence image, and the chromaticity of the pseudo-color image is specified by the user.

5. A visible light image and fluorescence image fusion system, characterized in that: include: An image acquisition system and a software system, wherein the image acquisition system includes a visible light camera for acquiring visible light images and a near-infrared camera for acquiring fluorescence images; the software system includes an image processing module, an image fusion module, and a video sequence generation module; The image processing module is used to perform image registration on the visible light camera and the near-infrared camera of the image acquisition system to obtain registration parameters; The image acquisition system is used to synchronously acquire visible light images and fluorescence images through a visible light camera and a near-infrared camera respectively, and synchronously input the acquired visible light images and fluorescence images into an image processing module of the software system; The image processing module is further configured to pre-process the received fluorescence image and align pixels of the visible light image and the pre-processed fluorescence image according to the registration parameters; The image fusion module is used to perform pixel fusion on the registered and aligned visible light and fluorescence images to generate a multispectral image, wherein the image fusion process specifically comprises weighted summation of corresponding pixels of the visible light image and the fluorescence image in a specific color space, wherein the weight parameter is calculated and determined by the pixel edge intensity value of the visible light image and the pixel intensity value of the fluorescence image; The video sequence generation module is used to sequentially merge the fused multispectral images to generate corresponding video sequences; by I″′ i,j =w′ i,j I′ i,j +w″ i,j I″ i,j Determine the pixel I″″ after image fusion i,j , where w′ i,j is the visible light image pixel I′ after registration and alignment i,j The fusion weight value, w″ i,j is the fluorescence image pixel I″ after registration and alignment i,j The fusion weight value, (i, j) is the subscript of the pixel; Depend on Determine the fusion weight value w′ of the visible light image i,j and the fusion weight value w″ of the fluorescence image i,j , where Ω is the set of all pixel subscripts belonging to the fluorescent area, p i,j ∈[0,1] is the edge intensity value of the visible light image pixel, q i,j ∈[0,1] is the intensity value of the fluorescence image pixel.

6. The system according to claim 5, characterized in that The registration of the visible light camera and the near infrared camera includes: During the registration process, the visible light camera and the near-infrared camera of the image acquisition system are placed in fixed positions, and the relative positions between the visible light camera and the near-infrared camera remain fixed; The target is placed on an adjustable metal slide rail, directly in front of the camera. The target can emit fluorescence within the near-infrared camera band. Visible light images and fluorescence images are synchronously acquired through an image acquisition system, and the acquired visible light images and fluorescence images are synchronously input into an image processing module of a software system; The feature detection algorithm of the image processing module is used to obtain the corresponding feature points and their coordinate positions of the visible light image and the fluorescence image. The coordinates of the corresponding feature points in the visible light image and the fluorescence image form feature pairs. A set of registration parameters is calculated based on the coordinates of multiple feature pairs. Adjust the distance between the target and the visible light camera and the fluorescence camera by a fixed length to obtain multiple sets of registration parameters; The obtained parameter groups are cleaned, and abnormal parameter groups that deviate from the overall average value and have large differences are discarded. The average values of the parameters in the remaining parameter groups are calculated as the final registration parameter group and written into the register of the image fusion module for table lookup during the registration process.

Images