Search

CN-121981897-A - Cell image stitching method and device and computer equipment

CN121981897ACN 121981897 ACN121981897 ACN 121981897ACN-121981897-A

Abstract

The embodiment of the application is suitable for the fields of biomedical technology and image processing technology, and provides a cell image stitching method, a cell image stitching device and computer equipment, wherein the method comprises the steps of shooting a plurality of cell images and performing image enhancement processing; the method comprises the steps of respectively determining relative displacement vectors between any two adjacent images after image enhancement processing, respectively determining global coordinates of target points of a first image in a plurality of cell images by taking the target points of the first image as global coordinate origins based on the relative displacement vectors, cutting overlapping areas and non-overlapping areas in each cell image, fusing the overlapping areas cut out from any two adjacent images to obtain a plurality of non-overlapping area images and a plurality of overlapping area fused images, and splicing the plurality of non-overlapping area images and the plurality of overlapping area fused images based on the global coordinates of the target points of the plurality of cell images. By adopting the method, the cell images can be spliced rapidly and accurately.

Inventors

  • CHEN YUCHAO
  • ZHANG FENG

Assignees

  • 利德健康科技(广州)有限公司

Dates

Publication Date
20260505
Application Date
20251225

Claims (10)

  1. 1. A method of stitching a cell image, comprising: shooting a plurality of cell images, and carrying out image enhancement processing on each cell image, wherein an overlapping area exists between adjacent images in the plurality of cell images; respectively determining relative displacement vectors between any two adjacent images after image enhancement processing; Taking a target point of a first image in the plurality of cell images as a global coordinate origin, and respectively determining global coordinates of the target point of each cell image based on the relative displacement vectors; cutting an overlapping region and a non-overlapping region in each cell image, and fusing the overlapping regions cut from any two adjacent images to obtain a plurality of non-overlapping region images and a plurality of overlapping region fused images; And based on global coordinates of target points of the plurality of cell images, stitching the plurality of non-overlapping region images and the plurality of overlapping region fusion images.
  2. 2. The method according to claim 1, wherein determining the relative displacement vector between any two adjacent images after the image enhancement processing includes: performing coarse registration on any two adjacent images after the image enhancement processing to obtain a coarse registration displacement; Carrying out fine registration on the overlapped area in any two adjacent images after the image enhancement processing to obtain a fine registration displacement; And determining a relative displacement vector between any two adjacent images based on the coarse registration displacement amount and the fine registration displacement amount.
  3. 3. The method according to claim 1 or 2, wherein the target point is a same position point in each of the cell images, the target point of the first image of the plurality of cell images is taken as a global origin of coordinates, and global coordinates of the target point of each of the cell images are respectively determined based on the relative displacement vectors, comprising: Taking a target point in a first image in a plurality of cell images as a global coordinate origin, and determining global coordinates of the target point in a second image according to a relative displacement vector between the first image and an adjacent second image; And determining the global coordinates of the target point in the next cell image adjacent to the cell image according to the relative displacement vector between the cell image and the next adjacent image by taking the global coordinates of the target point in any cell image as a reference, so as to obtain the global coordinates of the target point of each cell image.
  4. 4. A method according to claim 3, wherein the target point in each of the cell images is an upper left corner point in each of the cell images, the method further comprising: determining the minimum value of the global coordinates of the upper left corner position points in the plurality of cell images in the horizontal direction and the vertical direction; if the minimum value of the global coordinate of any position point in the horizontal direction or the vertical direction is smaller than zero, the global coordinate in the corresponding direction is adjusted so that the global coordinate of the upper left corner position point in each cell image is larger than or equal to zero.
  5. 5. The method of any one of claims 1,2 or 4, wherein fusing the overlapping regions cropped from any two adjacent images comprises: for any two adjacent images, respectively determining the weight value of the overlapping area in the two adjacent images; and carrying out weighted fusion on the overlapped region according to the weight value to obtain an overlapped region fusion image.
  6. 6. The method of claim 5, wherein said image enhancement processing of each of said cell images comprises: In the process of shooting each cell image, guiding and filtering the shot cell images in parallel to obtain a background smooth image; and performing linear transformation on the image obtained by subtracting the corresponding background smooth image from the cell image to obtain an enhanced image after image enhancement processing, wherein the enhanced image is subjected to normal distribution, and the mean value and the variance of the normal distribution are calculated based on the pixel value in the first image.
  7. 7. The method according to any one of claims 1 or 2 or 4 or 6, wherein stitching the plurality of non-overlapping region images and the plurality of overlapping region fusion images based on global coordinates of target points of the plurality of cell images comprises: constructing an initial image; And embedding each non-overlapping region image and each overlapping region fusion image into the corresponding position of the initial image according to the global coordinates of the target points of the cell images to obtain a spliced complete cell image.
  8. 8. The method of claim 7, wherein the length of the initial image is equal to the sum of the global coordinate maximum of the target point of the plurality of cell images in the horizontal direction and the length of any one of the cell images, and the width of the initial image is equal to the sum of the global coordinate maximum of the target point of the plurality of cell images in the vertical direction and the width of any one of the cell images.
  9. 9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the computer device is caused to implement the method of any one of claims 1 to 8 when the processor executes the computer program.
  10. 10. A computer program product comprising a computer program which, when run, causes the method of any one of claims 1 to 8 to be performed.

Description

Cell image stitching method and device and computer equipment Technical Field The embodiment of the application belongs to the technical field of biomedicine and image processing, and particularly relates to a cell image stitching method, device and computer equipment. Background The cell image stitching refers to a process of automatically or semi-automatically stitching together a plurality of high-resolution partial cell images shot under a microscope and having an overlapping area, and finally generating a complete and seamless large-view high-resolution image. The object of cell image stitching is mainly various microscopic cell images. In the process of splicing cell images, aiming at a plurality of partial fragment images with partial overlapping areas, which are obtained by shooting by an optical microscope area by area, in the prior art, a traditional feature detection algorithm is mainly adopted to extract feature points from the overlapping areas of adjacent partial images and screen candidate matching pairs, then mismatching points are removed by a random sampling consistency algorithm and the like, then a geometric transformation matrix between the adjacent images is calculated, alignment of the images to be spliced on pixel coordinates is realized, accurate overlapping of the overlapping areas is ensured, and finally a complete wide-view cell image is output. The method has the defects of large characteristic matching error, easiness in splicing dislocation, poor image fusion effect, easiness in amplifying grid effect due to uneven illumination, low splicing efficiency, difficulty in adapting to dynamic observation scenes, poor adaptability to sample deviation, insufficient robustness and the like, and seriously influences the efficiency and accuracy of cell image splicing. Disclosure of Invention In view of the above, the embodiments of the present application provide a cell image stitching method, device and computer device, which are used for improving the efficiency and accuracy of cell image stitching. A first aspect of an embodiment of the present application provides a cell image stitching method, including: shooting a plurality of cell images, and carrying out image enhancement processing on each cell image, wherein an overlapping area exists between adjacent images in the plurality of cell images; respectively determining relative displacement vectors between any two adjacent images after image enhancement processing; Taking a target point of a first image in the plurality of cell images as a global coordinate origin, and respectively determining global coordinates of the target point of each cell image based on the relative displacement vectors; cutting an overlapping region and a non-overlapping region in each cell image, and fusing the overlapping regions cut from any two adjacent images to obtain a plurality of non-overlapping region images and a plurality of overlapping region fused images; And based on global coordinates of target points of the plurality of cell images, stitching the plurality of non-overlapping region images and the plurality of overlapping region fusion images. Optionally, the determining the relative displacement vector between any two adjacent images after the image enhancement processing includes: performing coarse registration on any two adjacent images after the image enhancement processing to obtain a coarse registration displacement; Carrying out fine registration on the overlapped area in any two adjacent images after the image enhancement processing to obtain a fine registration displacement; And determining a relative displacement vector between any two adjacent images based on the coarse registration displacement amount and the fine registration displacement amount. Optionally, the target point is the same position point in each of the cell images, the global coordinate of the target point of each of the cell images is determined by using the target point of the first image of the cell images as a global coordinate origin and based on the relative displacement vector, including: Taking a target point in a first image in a plurality of cell images as a global coordinate origin, and determining global coordinates of the target point in a second image according to a relative displacement vector between the first image and an adjacent second image; And determining the global coordinates of the target point in the next cell image adjacent to the cell image according to the relative displacement vector between the cell image and the next adjacent image by taking the global coordinates of the target point in any cell image as a reference, so as to obtain the global coordinates of the target point of each cell image. Optionally, the target point in each of the cell images is an upper left corner point in each of the cell images, and the method further comprises: determining the minimum value of the global coordinates of the upper left corner position points