A full-slide hyperspectral scanning imaging device and method

By introducing three-dimensional mobile components and grating spectroscopy components in hyperspectral microscopy, automated data acquisition and splicing of full-slide hyperspectral scanning imaging devices is achieved, and the problems of slow imaging speed and lack of full-slide scanning capabilities in the prior art are solved. , improves data acquisition efficiency and image quality.

CN118655082BActive Publication Date: 2024-11-01SHANDONG UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411139365.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-20
Publication Date
2024-11-01
Estimated Expiration
2044-08-20

AI Technical Summary

Technical Problem

The existing hyperspectral microscopes have slow imaging speeds, which cannot meet the real-time imaging requirements, and do not have full-slide scanning capabilities. Data accuracy and stability are affected by light source and detector factors.

Method used

A full-slide hyperspectral scanning imaging device is designed, including three-dimensional moving components, grating spectroscopic components, imaging components and control components. Through the three-dimensional moving components, the grating spectroscopic components achieve rapid spectral dispersion, and the imaging component acquisition single Dimensional spectral image, control components perform image stitching, realizing automatic acquisition and stitching of spectral images of full-slide samples.

Benefits of technology

Automatic and efficient full-slide hyperspectral imaging is realized, data acquisition efficiency is improved, manual participation is reduced, image quality and stability is ensured, and multi-dimensional data cubes can be quickly acquired.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118655082B_ABST
    Figure CN118655082B_ABST
Patent Text Reader

Abstract

The present invention discloses a full-slide hyperspectral scanning imaging device and method, which relates to the field of hyperspectral imaging. In the device, a three-dimensional mobile component is used to accommodate a sample to be tested; a light source is used to irradiate the sample to be tested; an objective lens of a hyperspectral microscope is used to collect transmitted light carrying sample information and transmit it to a grating spectroscopic component; the grating spectroscopic component is used to spectrally disperse the received light to form monochromatic light of different wavelengths; an imaging component is used to image the monochromatic light of different wavelengths to obtain a single-dimensional spectral image of the sample and then send it to a control component; the three-dimensional mobile component is also used to receive a mobile scanning instruction issued by the control component, and drive the sample to be tested to move according to the mobile scanning instruction; the control component is used to store the single-dimensional spectral image of the sample and the coordinates of the moving position, and splice all the single-dimensional spectral images of the sample to obtain a full-slide sample spectral image. The present invention can automatically and efficiently realize full-slide hyperspectral imaging.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of hyperspectral imaging, and in particular to a full-slide hyperspectral scanning imaging device and method. Background Art

[0002] Hyperspectral Image (HSI) has higher spectral resolution than traditional color digital images, and usually contains dozens or hundreds of bands. Rich spectral information can provide a basis for accurate identification of targets, so it has been widely used in many fields. For example, in the field of life sciences, hyperspectral microscopes are widely used in cell imaging, histological research, neuroscience and medical diagnosis; hyperspectral microscopes can be used to image and quantitatively analyze different biological molecules in cells and tissues, which helps to understand cell functions, cell-to-cell interactions and disease development mechanisms. For example, in the medical field, Medical Hyperspectral Image (MHSI) can not only provide two-dimensional spatial distribution information of various tissue structures, but also obtain the complete spectrum of a certain point on a biological tissue sample in the wavelength range of interest. These characteristics determine the unique advantages of hyperspectral imaging technology in disease diagnosis.

[0003] However, the quality of hyperspectral microscopes currently available on the market varies greatly. Common problems include: slow imaging speed, unsuitable for application scenarios that require real-time imaging or rapid data collection; the performance of hyperspectral microscopes may be affected by factors such as the stability of the light source and the sensitivity of the detector, posing challenges to the accuracy and stability of the data, and other system hardware issues. At the same time, common imaging systems support the acquisition of hyperspectral data in a manually selected partial area range, and do not meet the requirements for full slide scanning. Summary of the invention

[0004] The object of the present invention is to provide a whole-slide hyperspectral scanning imaging device and method, which can automatically and efficiently realize whole-slide hyperspectral imaging.

[0005] To achieve the above object, the present invention provides the following solution.

[0006] In a first aspect, the present invention provides a full-slide hyperspectral scanning imaging device, comprising a light source, a three-dimensional moving component, a grating spectroscopic component, an imaging component, a hyperspectral microscope, and a control component.

[0007] The three-dimensional moving component is located below the objective lens of the hyperspectral microscope, and the three-dimensional moving component is used to accommodate the sample to be tested.

[0008] The light source is used to illuminate the sample to be tested on the three-dimensional moving component; the objective lens of the hyperspectral microscope is used to collect the transmitted light carrying the sample information and transmit it to the grating spectrometer component; the grating spectrometer component is used to spectrally disperse the received light to form monochromatic light of different wavelengths; the imaging component is used to image the monochromatic light of different wavelengths to obtain a single-dimensional spectral image of the sample, and then send the single-dimensional spectral image of the sample to the control component.

[0009] The three-dimensional moving component is also used to receive the moving scanning instruction issued by the control component, drive the sample to be tested to move according to the moving scanning instruction, and record the coordinates of each moving position.

[0010] The control component is used to store all the single-dimensional spectral images of the samples and all the moving position coordinates, and to splice all the single-dimensional spectral images of the samples to obtain a full-slide sample spectral image.

[0011] In a second aspect, the present invention provides a whole-slide hyperspectral scanning imaging method, comprising the following steps.

[0012] Acquire single-dimensional spectral images of all samples collected by a full-slide hyperspectral scanning imaging device.

[0013] A reference single-dimensional spectral image is determined, and a four-neighborhood search is performed on all the sample single-dimensional spectral images using a breadth-first traversal method to obtain a sequence of images to be spliced ​​corresponding to the reference single-dimensional spectral image.

[0014] A target image to be stitched is selected from the sequence of images to be stitched, and the target image to be stitched is registered and stitched with the reference single-dimensional spectral image to obtain a stitched image.

[0015] The first image in the sequence of images to be stitched is used as the reference single-dimensional spectral image, and then the process returns to the step of obtaining the corresponding sequence of images to be stitched, until all the single-dimensional spectral images of the sample are searched and stitched, and a full-slide sample spectral image is obtained.

[0016] According to the specific embodiment provided by the present invention, the present invention discloses the following technical effects: in the device of the present invention, the three-dimensional moving component accommodates the sample to be tested, the light source irradiates the sample to be tested, the objective lens of the hyperspectral microscope collects the transmitted light carrying the sample information and transmits it to the grating spectroscopic component, the grating spectroscopic component performs spectral dispersion on the received light to form monochromatic light of different wavelengths, the imaging component images the monochromatic light of different wavelengths to obtain a single-dimensional spectral image of the sample, and then sends it to the control component; the three-dimensional moving component also receives the mobile scanning instruction issued by the control component, and drives the sample to be tested to move according to the mobile scanning instruction; the control component stores all the single-dimensional spectral images of the samples and all the moving position coordinates, and splices all the single-dimensional spectral images of the samples to obtain a full-slide sample spectral image. In summary, the device of the present invention can realize automatic data acquisition, and combined with the three-dimensional moving component to drive the sample to be tested to move, it can obtain a data cube composed of the single-dimensional spectral image of the sample, that is, multi-dimensional data, and finally obtain the full-slide sample spectral image by splicing. Since no manual participation is required throughout the process, the efficiency is higher. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0018] Figure 1 It is a schematic diagram of the structure of the full-slide hyperspectral scanning imaging device of the present invention.

[0019] Figure 2 It is a schematic structural diagram of the grating spectrometer component of the present invention.

[0020] Figure 3 Schematic diagram of the functional architecture of the control component of the present invention.

[0021] Figure 4 Schematic diagram of the image transformation method of the present invention.

[0022] Figure 5 This is a flow chart of the SURF algorithm of the present invention.

[0023] Figure 6 This is a block diagram of the software control system architecture in an example of the present invention. DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] In order to make the purpose, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0026] Example 1.

[0027] like Figure 1 As shown, the present invention provides a full-slide hyperspectral scanning imaging device, including a light source, a three-dimensional moving component, a grating spectroscopic component, an imaging component, a hyperspectral microscope and a control component.

[0028] The three-dimensional moving component is located below the objective lens of the hyperspectral microscope, and is used to accommodate the sample to be tested. The sample to be tested is a biological tissue pathological section, food, plant section or soil, which can be changed as needed. When the sample to be tested is a biological tissue pathological section, the spectral characteristics of cells, tissues and biological specimens can be observed and analyzed by the device of the present invention.

[0029] The light source is used to illuminate the sample to be tested on the three-dimensional moving component; wherein the light source can be a light source with a wide spectral range and a relatively flat spectrum in the target band, and is fixed to the lower base. In order to increase the intensity of the incident light, the device also includes a condenser, which is arranged between the light source and the three-dimensional moving component, and is used to focus the light emitted by the light source on the sample to be tested on the three-dimensional moving component.

[0030] The objective lens of the hyperspectral microscope is used to collect the transmitted light carrying the sample information and transmit it to the grating spectroscopic component; wherein the hyperspectral microscope is a high-performance biological microscope, generally a commercial trinocular microscope is selected, which is equipped with a manual adjustment button, which is more stable than the manually constructed microscope system, and the button can be adjusted according to actual needs to manually switch between human eye observation and imaging components. In an application example, the device also includes a dual-frequency damping vibration isolation platform, and the hyperspectral microscope is placed on the dual-frequency damping vibration isolation platform to reduce the impact of physical jitter on optical imaging.

[0031] The grating light splitting component is used to perform spectral dispersion on the received light to form monochromatic light of different wavelengths; Figure 2As shown, the grating spectroscopic component includes a slit, a collimating objective lens, a plane grating and a converging objective lens arranged in sequence. After being imaged at the slit by a high-power microscope objective lens, and processed by a collimating objective lens, a plane grating and a converging objective lens, spectral dispersion is achieved along the vertical slit direction, and spatial information is retained in the horizontal direction. That is, a diffraction grating is used in a collimated light beam to achieve the purpose of spectroscopic analysis. Specifically, a large number of tiny slits of the same size and equal spacing are arranged to form a transmission grating, and each slit produces a diffraction fringe. The transmitted light waves will interfere with each other to form diffraction-interference fringes. The maximum position of the fringes is determined by the wavelength, so the spectral spectrum of the target can be obtained. The grating spectroscopic component of the present invention can quickly perform spectroscopic analysis without the need for mechanical moving parts, has good reliability and stability and a very fast response speed, and the spectral dispersion obtained by the grating is linear, and can be dispersed in the full band, and can also simplify the system structure.

[0032] The imaging component is used to image monochromatic light of different wavelengths to obtain a single-dimensional spectral image of the sample, and then send the single-dimensional spectral image of the sample to the control component. Specifically, the imaging component is an array camera detector; when the hyperspectral microscope is a trinocular microscope, the array camera detector is set at the vertical eyepiece position of the hyperspectral microscope to image tiny tissues. The selection of a high-sensitivity, low-noise, stable and reliable array camera detector can capture images with a wider field of view in real time and process them quickly.

[0033] In a specific application example, the device also includes a data transmission component; the data transmission component is used to transmit the single-dimensional spectral image of the sample obtained by the imaging component to the control component, thereby realizing data collection and transmission.

[0034] The three-dimensional moving component is also used to receive the moving scanning instruction issued by the control component, drive the sample to be tested to move according to the moving scanning instruction, and record the coordinates of each moving position. Specifically, the three-dimensional moving component includes an XYZ moving stage and a stage controller; the XYZ moving stage is used to accommodate the sample to be tested, and specifically, the XYZ moving stage is an electric three-axis precision stage, which can achieve coordinate positioning through a dedicated stage controller.

[0035] The stage controller is connected to the XYZ mobile stage and the control component respectively, and the stage controller is used to: receive the mobile scanning instruction issued by the control component, control the movement of the XYZ mobile stage according to the mobile scanning instruction, and drive the sample to be tested to move when the XYZ mobile stage moves; record the moving position coordinates of the XYZ mobile stage after movement and send them to the control component for storage. In addition, the stage controller can also control the switch and brightness of the preview light source and the transmission light source, which are uniformly regulated through the serial communication port and docked with the control component. The stage controller is also equipped with a memory module that can store the XYZ three-axis coordinate information of the stage. The advantage of introducing the memory module is that as long as the origin coordinates in the memory remain unchanged, the coordinate system for each operation can be the same coordinate system, which saves the initialization time of hardware restart.

[0036] The control component is used to store all the sample single-dimensional spectral images and all the moving position coordinates, and to splice all the sample single-dimensional spectral images to obtain a full-slide sample spectral image. Figure 3 As shown, the control component is generally a computer, and the software platform inside the computer can realize functions such as motion control, preview, scanning, splicing, and storage.

[0037] In summary, the device of the present invention controls the XYZ moving stage through a computer to achieve coordinate positioning, and stores the XYZ three-axis coordinate information of the stage at the same time, eliminating the initialization time of hardware restart. The computer is used for motion control to realize a push-sweep scanning method, that is, the grating spectroscopic component is imaged onto a highly sensitive array camera detector, and the one-dimensional spatial information and spectral information of the biological sample are obtained through photoelectric conversion; when the stage is pushed and scanned along the vertical slit direction, the other-dimensional spatial information of the biological sample can be obtained, thereby obtaining a data cube containing the two-dimensional spatial information and spectral information of the biological sample; and due to the use of the grating spectroscopic component, dispersion within the entire band can be achieved; the image data formed on the array camera detector is transmitted to the computer via the data transmission interface, and finally the data is collected and stored through the computer's built-in data acquisition software.

[0038] Example 2.

[0039] The present invention also provides a whole-slide hyperspectral scanning imaging method, which includes the following steps.

[0040] Step 100, obtaining all single-dimensional spectral images of samples acquired by the full-slide hyperspectral scanning imaging device described in Example 1. Specifically, after the full-slide hyperspectral scanning imaging device described in Example 1 completes data acquisition, it is necessary to splice the multiple single-dimensional spectral images of samples.

[0041] Step 200 , determining a reference single-dimensional spectral image, and performing a four-neighborhood search on all the sample single-dimensional spectral images using a breadth-first traversal method to obtain a sequence of images to be spliced ​​corresponding to the reference single-dimensional spectral image.

[0042] In practical applications, in order to quickly stitch large-scale images, a large-scale image stitching algorithm based on motion state context is used. The algorithm first establishes the motion state context, and the first sample single-dimensional spectral image received by the control component is used as the reference single-dimensional spectral image, such as Figure 4 As shown in the figure, during the transformation, the row transformation is used first, and then the column transformation (referring to registration and splicing) is calculated, that is, from O1 to O2 to O4. To realize the registration of all acquired images, it is necessary to traverse and search for images with the same area in a large-scale image for registration and splicing. Based on this requirement, the breadth-first traversal is used.

[0043] In the breadth-first traversal algorithm, the four-neighborhood images of the benchmark single-dimensional spectral image are first accessed for registration: first, the nearest image after the benchmark single-dimensional spectral image search is placed in the queue for registration and splicing with the benchmark single-dimensional spectral image, and then the registered image is released. After completing the above steps, a reorganization search is performed to find the image closest to the benchmark single-dimensional spectral image, and as the new benchmark single-dimensional spectral image, the four-neighborhood image search and subsequent registration and splicing are performed again. By adopting the four-neighborhood search mentioned above in the present invention, except for the initial and terminal scan rows or columns, each image needs to undergo four searches and matches, which increases the accuracy of splicing. This cycle is repeated until all sample single-dimensional spectral images are searched and traversed. The influence of errors on image splicing is reduced by aligning the motion state context and the image space. The process of registration, splicing, and iterative cycles that occurs in the above process is specifically as follows.

[0044] Step 300 : Select a target image to be stitched from the sequence of images to be stitched, and register and stitch the target image to be stitched with the reference single-dimensional spectral image to obtain a stitched image.

[0045] The target image to be stitched is registered and stitched with the reference single-dimensional spectral image to obtain a stitched image, which specifically includes the following steps.

[0046] (1) Extracting image features in the overlapping area between the target image to be stitched and the reference unidimensional spectral image, and the algorithms for extracting image features in the overlapping area between the target image to be stitched and the reference unidimensional spectral image include Scale-Invariant Feature Transform (SIFT), Speeded Up Robust Features (SURF) and Small Univalue Segment Assimilating Nucleus (SUSAN). It should be noted that the image registration method based on feature extraction has better robustness in scale, direction and projection reduction.

[0047] When the SURF algorithm is used for image feature extraction, such as Figure 5 The specific steps are as follows.

[0048] 11) Using the Hessian matrix to detect feature points in the target image to be stitched or the reference single-dimensional spectral image. Specifically, performing a differential operation on the image.

[0049] 12) Constructing the scale space by obtaining the response image of the Hessian matrix determinant. The scale space is established through the image pyramid, and the indirect method of continuously increasing the size s of the Gaussian filter template is adopted to obtain the response image of the Hessian matrix determinant through filter templates of different sizes and integral images.

[0050] 13) Locate feature points. This is achieved by using 3D non-maximum suppression on the response image. In a 3×3×3 neighborhood, the point with the largest eigenvalue is selected as the feature point of the area, and the rest are discarded.

[0051] 14) Determine the main direction. Calculate the maximum Haar response cumulative value and extract the main direction feature points.

[0052] 15) Generate feature descriptors and calculate the Hessian matrix response of the selected area. When the Gaussian filter template size is s, the 20s×20s image is divided into 4×4 sub-blocks along the main direction with the feature point as the center. Each sub-block uses the Harr template of size 2s to calculate the response value, and then the response values ​​∑dx, ∑|dx|, ∑dy, ∑|dy| form a feature vector, where x refers to the horizontal coordinate and y refers to the vertical coordinate; the origin of the coordinate system is the center of the first sample single-dimensional spectral image.

[0053] (2) Based on the image features in the overlapped area, determining a feature transformation relationship between the two images; the feature transformation relationship between the two images includes a rotation matrix R and a displacement vector T.

[0054] (3) Based on the transformation relationship of the two image features, the target image to be stitched and the reference single-dimensional spectral image are registered and stitched together to obtain a stitched image. After determining the transformation relationship of the feature points, the two sets of feature points are overlapped, and the pixels of the overlapped part are processed to obtain a stitched image.

[0055] Specifically, the following formula is used for image registration.

[0056] .

[0057] Where F represents the minimum error between the pixel set in the reference single-dimensional spectral image and the pixel set in the target image to be stitched. Express request The values ​​of R and T at the minimum, n represents the number of pixels in the benchmark single-dimensional spectral image, represents the i-th pixel in the benchmark single-dimensional spectral image, Represents the i-th pixel in the target image to be stitched.

[0058] The image after the SURF feature points are repositioned has no obvious misalignment, but there are gaps at the splicing. Based on this, after the target image to be spliced ​​and the reference single-dimensional spectral image are spliced ​​and fused, the method also includes: using curve fitting to process the splicing and fusion area to achieve a smooth transition at the splicing.

[0059] In addition, the algorithm for splicing and fusing the target image to be spliced ​​and the reference single-dimensional spectral image includes pixel weighted fusion and frequency domain conversion method.

[0060] Step 400, taking the first image in the sequence of images to be stitched as a reference single-dimensional spectral image, and then returning to the step of obtaining the corresponding sequence of images to be stitched, until all the single-dimensional spectral images of the sample are searched and stitched, and obtaining a full-slide sample spectral image.

[0061] In a specific example application, such as Figure 6 As shown, configure the device according to the content in Example 1 and check whether the device is connected; confirm that the device is successfully connected, load the configuration information, enter the image preview mode, and the client sets the exposure, focus and other parameters by observing the received preview image. Then enter the image acquisition mode, set the device parameters and the synchronization trigger source, and start the synchronous acquisition. The computer simultaneously acquires and stores the XYZ mobile stage data and the hyperspectral image data, and uses the SURF algorithm to extract the feature points in the overlapping area of ​​the read image, completes the image registration according to the valid feature points, and splices and fuses the registered images to finally obtain a seamless full-slide sample spectral image. After the client acquires the end instruction, the client sends a shutdown instruction, and it ends here.

[0062] In summary, the method of the present invention applies a large-scale dynamic real-time whole-slide stitching algorithm: first read the acquired image, then extract the features in the overlapping area of ​​the read image, perform image registration and image fusion, and compare and analyze the stitched image with the preview image. The stitched image well restores the global image of the tissue section, achieving seamless stitching with good effect.

[0063] Compared with the prior art, the present invention has the following advantages.

[0064] (1) While ensuring image quality, the present invention significantly reduces image registration time and achieves efficient image stitching.

[0065] (2) The microscopic pathology hyperspectral whole-slide scanning imaging proposed in the present invention can output color images and hyperspectral images of a single field of view or the entire slide, and stitch independent single-field images together into a high-definition macroscopic image, making it easier for researchers to observe the microscopic details and macroscopic features of the sample. It has important research value and can provide a richer data source for the analysis of biological tissue samples.

[0066] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0067] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A full-slide hyperspectral scanning imaging method, characterized in that the method include: Acquire single-dimensional spectral images of all samples collected by a full-slide hyperspectral scanning imaging device; Determine a reference single-dimensional spectral image, and use a breadth-first traversal method to perform a four-neighborhood search on all the sample single-dimensional spectral images to obtain a sequence of images to be spliced ​​corresponding to the reference single-dimensional spectral image; specifically, the method includes: in the breadth-first traversal algorithm, first access the four-neighborhood images of the reference single-dimensional spectral image for registration: first put the nearest image after the search of the reference single-dimensional spectral image into a queue for registration and splicing with the reference single-dimensional spectral image, and then release the registered image; after completing the above steps, perform a reorganization search to find the image closest to the reference single-dimensional spectral image, and use it as a new reference single-dimensional spectral image, and perform a four-neighborhood image search and subsequent registration and splicing again; Selecting a target image to be stitched from the sequence of images to be stitched, registering and stitching the target image to be stitched with the reference single-dimensional spectral image to obtain a stitched image; specifically comprising: extracting image features in an overlapping area between the target image to be stitched and the reference single-dimensional spectral image; determining a feature transformation relationship between the two images based on the image features in the overlapping area; performing image registration and stitching fusion on the target image to be stitched and the reference single-dimensional spectral image based on the feature transformation relationship between the two images to obtain a stitched image; The first image in the sequence of images to be stitched is used as a reference single-dimensional spectral image, and then the process returns to the step of obtaining a corresponding sequence of images to be stitched, until all the single-dimensional spectral images of the sample are searched and stitched, and a full-slide sample spectral image is obtained; The full-slide hyperspectral scanning imaging device includes a light source, a three-dimensional moving component, a grating spectroscopic component, an imaging component, a hyperspectral microscope and a control component; The three-dimensional moving component is located below the objective lens of the hyperspectral microscope, and the three-dimensional moving component is used to accommodate the sample to be tested; The light source is used to illuminate the sample to be tested on the three-dimensional moving component; the objective lens of the hyperspectral microscope is used to collect the transmitted light carrying the sample information and transmit it to the grating spectroscopic component; the grating spectroscopic component is used to perform spectral dispersion on the received light to form monochromatic light of different wavelengths; the imaging component is used to image the monochromatic light of different wavelengths to obtain a single-dimensional spectral image of the sample, and then send the single-dimensional spectral image of the sample to the control component; The three-dimensional moving component is also used to receive the moving scanning instruction issued by the control component, drive the sample to be tested to move according to the moving scanning instruction, and record the coordinates of each moving position; The control component is used to store all the single-dimensional spectral images of the samples and all the moving position coordinates, and to splice all the single-dimensional spectral images of the samples to obtain a full-slide sample spectral image.

2. The full-slide hyperspectral scanning imaging method according to claim 1, characterized in that: The three-dimensional moving assembly includes an XYZ moving stage and a stage controller; The XYZ moving stage is used to accommodate the sample to be tested; The stage controller is connected to the XYZ mobile stage and the control component respectively, and the stage controller is used to: receive a mobile scanning instruction issued by the control component, control the movement of the XYZ mobile stage according to the mobile scanning instruction, and move the sample to be tested when the XYZ mobile stage moves; record the moving position coordinates of the XYZ mobile stage after movement and send them to the control component for storage.

3. The full-slide hyperspectral scanning imaging method according to claim 1, characterized in that: The sample to be tested is a biological tissue pathological section, food, plant section or soil.

4. The full-slide hyperspectral scanning imaging method according to claim 1, characterized in that: The device also includes data transmission components, condenser and dual-frequency damping vibration isolation platform; The data transmission component is used to transmit the sample single-dimensional spectral image obtained by the imaging component to the control component; The condenser is arranged between the light source and the three-dimensional moving component, and is used to focus the light emitted by the light source on the sample to be tested of the three-dimensional moving component; The hyperspectral microscope is placed on the dual-frequency damping vibration isolation platform; The grating light splitting component comprises a slit, a collimating objective lens, a plane grating and a converging objective lens which are arranged in sequence.

5. The full-slide hyperspectral scanning imaging method according to claim 1, characterized in that: The imaging component is an area array camera detector; the hyperspectral microscope is a trinocular microscope; the area array camera detector is arranged at the vertically arranged eyepiece position of the hyperspectral microscope.

6. The whole-slide hyperspectral scanning imaging method according to claim 1, characterized in that: Based on the feature transformation relationship between the two images, image registration is performed on the target image to be stitched and the reference single-dimensional spectral image, specifically including: The following formula is used for image registration: ; Wherein, F represents the minimum error between the pixel set in the reference single-dimensional spectral image and the pixel set in the target image to be spliced, and the feature transformation relationship between the two images includes the rotation matrix R and the displacement vector T. Express request The values ​​of R and T at the minimum, n represents the number of pixels in the benchmark single-dimensional spectral image, represents the i-th pixel in the benchmark single-dimensional spectral image, Represents the i-th pixel in the target image to be stitched.

7. The full-slide hyperspectral scanning imaging method according to claim 1, characterized in that: The algorithm for extracting the image features in the overlapping area between the target image to be stitched and the reference single-dimensional spectral image includes a scale-invariant feature conversion algorithm, an accelerated robust feature algorithm and a minimum core value similar area feature detection algorithm; The algorithm for splicing and fusing the target image to be spliced ​​and the reference single-dimensional spectral image includes pixel weighted fusion and frequency domain conversion method.

8. The full-slide hyperspectral scanning imaging method according to claim 1, characterized in that: After the target image to be stitched and the reference single-dimensional spectral image are stitched and fused, the method further includes: processing the stitched and fused area by using a curve fitting method.