An image stitching method applicable to array micro-CT imaging

Abstract

The present invention relates to image stitching technology, specifically an image stitching method applicable to array micro-CT imaging, and the method is implemented by the following steps: Step 1: Collect images of a sample to be measured from multiple perspectives, thereby obtaining multiple images; Step 2: For each image, define all pixel gray values of the image as a data set; then, for each data set, divide the data set into five clusters; Step 3: Define all gray calibration values of the images as a data set, and divide the data set into five clusters; Step 4: Perform gray value redistribution on each image; Step 5: Identify the overlapping regions of each image, calculate the positional relationship of the overlapping regions, and then stitch each image to obtain a stitched image. The present invention effectively solves the problems that the existing image stitching methods are prone to poor stitching quality and distorted stitched images when applied to array micro-CT imaging, and is applicable to array micro-CT imaging.

Description

Technical Field

[0001] The present invention relates to image stitching technology, and specifically to an image stitching method applicable to array micro-CT imaging. Background Art

[0002] Micro-CT has a higher spatial resolution compared to ordinary CT, but the imaging field of view of micro-CT is often small. Therefore, for a relatively large sample to be measured, an array micro-CT imaging system needs to be built, and images of the sample to be measured are collected from multiple perspectives using the array micro-CT imaging system, and then multiple images are stitched together to achieve array micro-CT imaging. Currently, image stitching methods are mainly divided into two types: The first is the image stitching method based on image grayscale. This method is applicable to the situation where the image grayscale distribution is relatively scattered, but when applied to array micro-CT imaging, due to the relatively concentrated image grayscale distribution, it is easy to result in poor stitching quality. The second is the image stitching method based on image features. This method is applicable to the situation where the overlap rate between images is relatively low, but when applied to array micro-CT imaging, due to the relatively high overlap rate between images, it is easy to cause distortion of the stitched image. Based on this, it is necessary to invent an image stitching method applicable to array micro-CT imaging to solve the problems that the existing image stitching methods are prone to poor stitching quality and distorted stitched images when applied to array micro-CT imaging. Summary of the Invention

[0003] In order to solve the problems that the existing image stitching methods are prone to poor stitching quality and distorted stitched images when applied to array micro-CT imaging, the present invention provides an image stitching method applicable to array micro-CT imaging.

[0004] The present invention is implemented by adopting the following technical solutions:

[0005] An image stitching method applicable to array micro-CT imaging, which is implemented by the following steps:

[0006] Step 1: Use the array micro-CT imaging system to collect images of the sample to be measured from multiple perspectives, thereby obtaining multiple images;

[0007] Step 2: For each image, define all pixel grayscale values of the image as a data set; then, for each data set, use the fuzzy C-means clustering algorithm to divide the data set into five clusters, and define the five cluster centers as the grayscale calibration values of the image corresponding to the data set;

[0008] Step 3: Define the grayscale calibration values of all images as a data set, use the fuzzy C-means clustering algorithm to divide the data set into five clusters, and then define the five cluster centers as the common grayscale calibration values of each image;

[0009] Step Four: Re - distribute the gray levels of each image according to the common gray - level calibration value of each image.

[0010] Step Five: Use the edge - detection algorithm to identify the overlapping regions of each image, use the eight - neighborhood method to calculate the positional relationship of the overlapping regions, and then splice each image according to the positional relationship of the overlapping regions to obtain the spliced image.

[0011] Compared with the existing image - splicing methods, the image - splicing method for array micro - CT imaging described in the present invention uses the fuzzy C - means clustering algorithm to achieve image splicing, and thus has the following advantages: First, compared with the image - splicing method based on image gray levels, the present invention is applicable to the situation where the image gray - level distribution is relatively concentrated. Therefore, when applied to array micro - CT imaging, the present invention can effectively improve the splicing quality. Second, compared with the image - splicing method based on image features, the present invention is applicable to the situation where the overlapping rate between images is relatively high. Therefore, when applied to array micro - CT imaging, the present invention can effectively avoid the distortion of the spliced image.

[0012] The present invention effectively solves the problems that the existing image - splicing methods are prone to poor splicing quality and distortion of the spliced image when applied to array micro - CT imaging, and is applicable to array micro - CT imaging. Detailed Embodiment

[0013] An image - splicing method for array micro - CT imaging is implemented by the following steps:

[0014] Step One: Use the array micro - CT imaging system to collect images of the sample to be measured from multiple perspectives, thereby obtaining multiple images.

[0015] Step Two: For each image, define all the pixel gray - level values of the image as a data set; then, for each data set, use the fuzzy C - means clustering algorithm to divide the data set into five clusters, and define the five cluster centers as the gray - level calibration value of the image corresponding to the data set.

[0016] Step Three: Define all the gray - level calibration values of the images as a data set, use the fuzzy C - means clustering algorithm to divide the data set into five clusters, and then define the five cluster centers as the common gray - level calibration value of each image.

[0017] Step Four: Re - distribute the gray levels of each image according to the common gray - level calibration value of each image.

[0018] Step Five: Use the edge - detection algorithm to identify the overlapping regions of each image, use the eight - neighborhood method to calculate the positional relationship of the overlapping regions, and then splice each image according to the positional relationship of the overlapping regions to obtain the spliced image.

[0019] Although the specific embodiments of the present invention have been described above, those skilled in the art should understand that these are only illustrative examples, and the protection scope of the present invention is defined by the appended claims. Without departing from the principle and essence of the present invention, those skilled in the art can make various changes or modifications to these embodiments, but such changes and modifications all fall within the protection scope of the present invention.

Claims

1. An image stitching method applicable to array micro-CT imaging, characterized in that: The method is implemented by the following steps: Step 1: Use an array micro-CT imaging system to collect images of the sample to be measured from multiple perspectives, thereby obtaining multiple images; Step 2: For each image, define all pixel gray values of the image as a data set; then, for each data set, use the fuzzy C-means clustering algorithm to divide the data set into five clusters, and define the five cluster centers as the gray calibration values of the image corresponding to the data set; Step 3: Define the gray calibration values of all images as a data set, use the fuzzy C-means clustering algorithm to divide the data set into five clusters, and then define the five cluster centers as the common gray calibration values of each image; Step 4: Reassign the gray levels of each image according to the common gray calibration values of each image; Step 5: Use an edge detection algorithm to identify the overlapping regions of each image, use the eight-neighborhood method to calculate the positional relationship of the overlapping regions, and then stitch each image according to the positional relationship of the overlapping regions to obtain a stitched image.