Disclosure of Invention
Technical problem to be solved
Aiming at the defects of the prior art, the utility model provides a super-resolution imaging digital slide scanner, which solves the problems that 1) the traditional microscope cannot be digitalized and has low diagnosis efficiency; 2) the digital scanner has the problems of high cost and inflexible magnification.
(II) technical scheme
In order to achieve the above purpose, the utility model discloses a following technical scheme realizes: a super-resolution imaging digital slide scanner comprises a light source system, wherein the upper end face of the light source system is connected with a slide platform in a sliding manner, the side end face of the slide platform is fixedly connected with a Y-axis motion module, the side end face of the Y-axis motion module is provided with a Y-axis drive motor, the Y-axis drive motor drives the Y-axis motion module to move, the lower end face of the Y-axis motion module is connected with an X-axis motion module in a sliding manner, one side of the side end face of the X-axis motion module is fixedly connected with an X-axis drive motor, the X-axis drive motor drives the X-axis motion module to move, one side of the side end face of the light source system, which is far away from the X-axis drive motor, is provided with a Z-axis focusing module, the upper end face of the Z-axis focusing module is rotatably connected with a Z-axis drive motor, the Z-axis focusing module is driven by the, the lower end face of the optical path imaging system is fixedly provided with an objective lens, the upper end face of the optical path imaging system is provided with a scanning camera, and one side of the side end face of the optical path imaging system, which deviates from the Z-axis focusing module, is fixedly connected with a preview camera.
Preferably, the Y-axis motion module, the X-axis motion module and the Z-axis focusing module are matched with the ball screw through a Y-axis drive motor, an X-axis drive motor and a Z-axis drive motor to drive the Y-axis motion module, the X-axis motion module and the Z-axis focusing module to move in corresponding directions.
Preferably, the slide platform is adjusted in position in the horizontal direction, in the transverse direction and in the longitudinal direction, by the Y-axis movement module and the X-axis movement module.
Preferably, the magnification factor of the objective lens is 10 times, the focusing distance between the objective lens and the slide platform is controlled through the Z-axis focusing module, and the depth of field of the objective lens is larger than 10 um.
Preferably, the upper surface of the slide platform holds a scanning slide.
Preferably, the movement stroke of the X-axis movement module is greater than the length of the slide, and the movement stroke of the Y-axis movement module is greater than the width of the slide.
Preferably, the optical path imaging system is integrated with a super-resolution digital imaging system.
The method also comprises an image processing principle of the super-resolution imaging digital slide scanner, during imaging, a preview camera is used for firstly preliminarily previewing the slide in the slide platform, the position and the size of an area needing to be scanned are judged, then a Z-axis driving motor is used for driving a Z-axis focusing module to drive an objective lens to adjust the focal length, the scanning camera is used for preliminary scanning and splicing the slide, after the scanning is finished, a user can mark the position of the image needing to be amplified according to the preliminarily scanned image, the marking head plans the movement path of the slide platform through software and the preview camera, meanwhile, the Y-axis motion module and the X-axis motion module drive the slide platform to perform super-resolution scanning in a motion shooting mode, after the scanned pictures are spliced and analyzed, the image is embedded into the preliminarily scanned image, and simultaneously the image is converted into a digital format and uploaded to a data cloud.
Preferably, the motion capture includes: line-wise shot, column-wise shot, and s-track shot.
Preferably, the super-resolution scanning obtains a plurality of groups of images through motion shooting, and the high-resolution images are obtained through iterative combination operation based on a resolution reconstruction algorithm of a Deep Convolutional Neural Network (DCNN).
(III) advantageous effects
The utility model provides a super-resolution imaging digital slide scanner possesses following beneficial effect: the device is provided with a preview camera to position and guide the slide and the picture position and preliminarily scan so as to provide a basis for the automatic movement and super-resolution imaging of a slide platform, the slide platform horizontally and longitudinally moves through a Y-axis movement module and an X-axis movement module, an objective lens vertically moves through a Z-axis focusing module, the Y-axis movement module, the X-axis movement module and the Z-axis focusing module are respectively driven by a Y-axis driving motor, an X-axis driving motor and a Z-axis driving motor, and the Y-axis driving motor, the X-axis driving motor and the Z-axis driving motor can be programmed by software and are combined with the preview camera to position, so that the automatic control steps such as automatic focusing and automatic scanning are realized, the operation difficulty in use is greatly reduced, the precision of the scanned picture is improved, the objective lens is a 10X objective lens, and the view field diameter of the whole light path, the imaging field area of a single picture is 4 times of that of a 20X objective lens by matching with an imaging camera, so that the number of pictures to be shot is 1/4 of the 20X objective lens, the number of the shot pictures is greatly reduced, the scanning efficiency is greatly improved, as with other scanners, the focal plane detection is required to be carried out on the whole scanning area before continuous scanning so as to judge the focusing motion track of the objective lens in the scanning process, as the 10-time objective lens is adopted, the depth of field range is larger, the selection of focal points can be less than that of the 20X objective lens, the scanning time can be saved, the scanning efficiency of the equipment is further improved, meanwhile, a super-resolution imaging system is embedded in an optical path imaging system, a plurality of low-resolution images are utilized to obtain relevant information, a high-resolution image is formed through the reconstruction process, the super-resolution imaging method adopted by the system is a resolution reconstruction algorithm based on a Depth Convolution Neural Network (DCNN), the DCNN network includes two alternating convolutional layers, a maximize pool layer, two fully-connected layers, and a last classification layer. The convolution layer and the maximization pool layer respectively generate a convolution sum maximization pool characteristic map through continuous convolution and maximization pool operation. These feature maps then support the extraction and combination of an appropriate set of image features from the training set. The algorithm carries out iterative training on a large number of image sets acquired by the scanner, optimizes and updates the characteristic parameters of the model, and obtains a final network model. Carry out the final high resolution image that arrives of iterative composition to the low resolution image through neural network model, realized shooting high definition image with low power objective, the image resolution who forms under 10 times objective promotes to and is unanimous with 20 times or 40 times objective imaging resolution and detail simultaneously not lost, make the cost of reduction equipment, carry out the form conversion through software to high resolution image and preliminary scanning image after embedding integration, send the data high in the clouds, the person of facilitating the use is to the storage of data, the operation of transmission and sharing, the scanner uses more conveniently.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the accompanying drawings are exemplary only for the purpose of explaining the present invention, and should not be construed as limiting the present invention.
In the description of the present invention, it is to be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, and are only for convenience of description and to simplify the description, but do not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore should not be construed as limiting the present invention. Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first", "second", may explicitly or implicitly include one or more of the described features. In the description of the present invention, "a plurality" means two or more unless specifically limited otherwise.
In the description of the present invention, it is to be noted that, unless otherwise explicitly specified or limited, the terms "mounted," "connected," and "connected" are to be construed broadly, and may be, for example, fixedly connected, detachably connected, or integrally connected; may be mechanically connected, may be electrically connected or may be in communication with each other; either directly or indirectly through intervening media, either internally or in any other relationship. The specific meaning of the above terms in the present invention can be understood according to specific situations by those skilled in the art.
In the present disclosure, unless expressly stated or limited otherwise, the first feature "on" or "under" the second feature may comprise direct contact between the first and second features, or may comprise contact between the first and second features not directly. Also, the first feature being "on," "above" and "over" the second feature includes the first feature being directly on and obliquely above the second feature, or merely indicating that the first feature is at a higher level than the second feature. A first feature being "under," "below," and "beneath" a second feature includes the first feature being directly under and obliquely below the second feature, or simply meaning that the first feature is at a lesser elevation than the second feature.
The following disclosure provides many different embodiments or examples for implementing different features of the invention. In order to simplify the disclosure of the present invention, the components and arrangements of specific examples are described below. Of course, they are merely examples and are not intended to limit the present invention. Furthermore, the present invention may repeat reference numerals and/or reference letters in the various examples, which have been repeated for purposes of simplicity and clarity and do not in themselves dictate a relationship between the various embodiments and/or arrangements discussed. In addition, the present disclosure provides examples of various specific processes and materials, but one of ordinary skill in the art may recognize applications of other processes and/or use of other materials.
The embodiment of the utility model provides a super-resolution imaging digital slide scanner, including light source system 9, the up end of light source system 9 is connected with slide platform 4 in a sliding manner, the side end of slide platform 4 is fixedly connected with Y-axis motion module 5, the side end of Y-axis motion module 5 is equipped with Y-axis drive motor 6, and Y-axis drive motor 6 drives Y-axis motion module 5 to move, the lower end of Y-axis motion module 5 is connected with X-axis motion module 7 in a sliding manner, X-axis drive motor 8 is fixedly connected with one side of the side end of X-axis motion module 7, and X-axis drive motor 8 drives X-axis motion module 7 to move, Z-axis focusing module 11 is equipped with Z-axis focusing module 11 on the side of the side end of light source system 9 departing from X-axis drive motor 8, the up end of Z-axis focusing module 11 is rotationally connected with Z-axis drive motor 12, Z-axis focusing module 11 is driven by, the side end face of the Z-axis focusing module 11 is fixedly connected with an optical path imaging system 3, the lower end face of the optical path imaging system 3 is fixedly provided with an objective lens 10, the upper end face of the optical path imaging system 3 is provided with a scanning camera 1, and one side of the optical path imaging system 3, which deviates from the Z-axis focusing module 11, is fixedly connected with a preview camera 2.
Y-axis motion module 5, X-axis motion module 7 and Z-axis focus module 11 through Y-axis driving motor 6, X-axis driving motor 8 and Z-axis driving motor 12 and ball screw adaptation drive Y-axis motion module 5, X-axis motion module 7 and Z-axis focus module 11 remove in corresponding direction, slide platform 4 is horizontal and fore-and-aft position control through Y-axis motion module 5 and X-axis motion module 7 on the horizontal direction, objective 10 magnification factor specifically is 10 times, and focus the distance of focusing of control objective 10 and slide platform 4 through Z-axis focus module 11, objective depth of field of objective 10 is greater than 10um, the upper surface of slide platform 4 is held the scanning slide, the motion stroke of X-axis motion module 7 is greater than slide length, the motion stroke of Y-axis motion module 5 is greater than the width of slide.
Before use, optimizing and verifying the established network model, and the first step is as follows: 5000 groups of low-resolution scanning images are collected as a data set, and high-resolution reconstruction is carried out on each group of low-resolution images through a neighbor difference method, a bilinear difference method or a bicubic difference algorithm to obtain an initial high-resolution image. From all data sets, 70% were randomly selected as training set, 15% as test set, and 15% as validation set.
The second step is that: inputting the images of the test set into a deep convolution network model ResNet, performing end-to-end learning by using the high resolution images reconstructed in the step 1 as network output, obtaining a minimized loss function by iteration by using a mean square error as a loss function, and optimizing model parameters. Compared with the traditional method for reconstructing the single super-resolution image, the network integrates multiple reconstruction methods to obtain an overall optimal algorithm.
The third step: and testing and evaluating the trained model by using the test set. And feeding the images of the test set to a network model to obtain corresponding high-resolution images, comparing the high-resolution images with the high-resolution images obtained before training, calculating the average peak signal-to-noise ratio and the structural similarity of the high-resolution images, and finally proving that the image reconstruction effect of the network model is better.
The fourth step: and embedding the optimized and verified network model program on a scanner system to realize super-resolution imaging of the real-time acquired image of the objective lens.
The image processing principle of the super-resolution imaging digital slide scanner is that during imaging, a preview camera 2 is used for preliminarily previewing a slide in a slide platform 4, the position and the size of an area to be scanned are judged, then a Z-axis driving motor 12 is used for driving a Z-axis focusing module 11 to drive an objective lens 10 to adjust the focal length, the slide is preliminarily scanned and spliced through a scanning camera 1, after scanning is completed, a user can mark the position of the image to be amplified according to the preliminarily scanned image, a marking head is used for planning the movement path of the slide platform 4 through software and in combination with the preview camera 2, meanwhile, a Y-axis movement module 5 and an X-axis movement module 7 are used for driving the slide platform 4 to carry out super-resolution scanning in a movement shooting mode, after the scanned image is spliced and analyzed, the image is embedded into the preliminarily scanned image, and simultaneously the image is converted into a digital format, the method comprises the steps of line shooting, column shooting and s-type track shooting, wherein multiple groups of images are obtained after super-resolution scanning is carried out through motion shooting, and iterative combination operation is carried out on a resolution reconstruction algorithm based on a Deep Convolutional Neural Network (DCNN) to obtain a high-resolution image.
In conclusion, the preview camera 2 is arranged to position and guide the slide and the picture position and perform preliminary scanning to provide a basis for automatic movement and super-resolution imaging of the slide platform 4, the slide platform 4 horizontally and longitudinally moves through the Y-axis movement module 5 and the X-axis movement module 7, the objective lens 10 vertically moves through the Z-axis focusing module 11, the Y-axis movement module 5, the X-axis movement module 7 and the Z-axis focusing module 11 are respectively driven by the Y-axis driving motor 6, the X-axis driving motor 8 and the Z-axis driving motor 12, and meanwhile, the Y-axis driving motor 6, the X-axis driving motor 8 and the Z-axis driving motor 12 can be programmed by software and combined with the preview camera 2 to perform position positioning, so as to realize automatic focusing, automatic scanning and other automatic control steps, greatly reduce the use operation difficulty and improve the picture scanning precision, the objective lens 10 adopts a 10X objective lens, the field diameter of the whole light path is about 2mm, the imaging view area of a single picture is 4 times of that of the 20X objective lens by matching with an imaging camera, so that the number of pictures to be shot is 1/4 of the 20X objective lens, the number of the shot pictures is greatly reduced, the scanning efficiency is greatly improved, as with other scanners, the detection of a focal plane is required to be carried out on the whole scanning area before continuous scanning so as to judge the focusing movement track of the objective lens in the scanning process, because the 10X objective lens is adopted and the depth of field range is larger, the selection of a focusing point can be less than that of the 20X objective lens, the scanning time can be saved, the scanning efficiency of the equipment is further improved, meanwhile, a super-resolution imaging system is embedded in the light path imaging system 3, the related information is obtained by utilizing a plurality of low-resolution images, and a high-resolution image is formed through the, the super-resolution imaging method adopted by the system is a resolution reconstruction algorithm based on a Deep Convolutional Neural Network (DCNN), and the DCNN comprises two alternating convolutional layers, a maximization pool layer, two full-connection layers and a last classification layer. The convolution layer and the maximization pool layer respectively generate a convolution sum maximization pool characteristic map through continuous convolution and maximization pool operation. These feature maps then support the extraction and combination of an appropriate set of image features from the training set. The algorithm carries out iterative training on a large number of image sets acquired by the scanner, optimizes and updates the characteristic parameters of the model, and obtains a final network model. Carry out the final high resolution image that arrives of iterative composition to the low resolution image through neural network model, realized shooting high definition image with low power objective, the image resolution who forms under 10 times objective promotes to and is unanimous with 20 times or 40 times objective imaging resolution and detail simultaneously not lost, make the cost of reduction equipment, carry out the form conversion through software to high resolution image and preliminary scanning image after embedding integration, send the data high in the clouds, the person of facilitating the use is to the storage of data, the operation of transmission and sharing, the scanner uses more conveniently.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.