基于骨架变换的显微图像拼接方法TheMethodof-中国显微图像网.PDF

基于骨架变换的显微图像拼接方法 童锐，简淼夫，周月君 （南京工业大学材料科学与工程学院，江苏 南京 210009 ） 摘 要：针对高放大倍数下显微镜图像视野较小的问题，提出了一种基于骨架变换的显微图像拼接方法。该方 法先调整显微镜的 Z 轴位置，获取样本的序列显微图像；再对相邻序列显微图像进行拉普拉斯算子对比，得到 显微图像清晰区域的二值图像；经过数学形态学处理后，提取出清晰区域的骨架；对相邻骨架插值，得到新骨 架，提取出序列显微图像中新骨架与相邻骨架所包围区域，最终拼接出显微图像的清晰区域。试验证明，该技 术能得到较好的拼接结果。 关键字：骨架变换；显微图像；图像拼接；数学形态学；拉普拉斯算子 中图分类号：TH391.41 文献标识码：A The Method of Image Mosaics from Microscopic Images Based on Skeleton Transform TONG Rui, JIAN Miao-fu, ZHOU Yue-jun （College of Materials Science and Engineering，Nanjing University of Technology ，Nanjing 210009 ，China ） Abstract ：In order to solve the problem that only very limited field of view is available under high magnification, this paper proposed a useful method of microscopic image mosaics based on skeleton transform. In this method, first the Z axis of the microscope was adjusted, so as to obtain the quence microscopic images of the specimen. Afterwards, the binary image of the focused area was obtained by comparing the Laplacian operator of the adjacent microscopic images. And then, the skeleton of the focused area was extracted after the treatment of mathematical morphological. Lastly, a new skeleton was obtained by interpolating of the adjacent skeletons, and the enclosed area between the new skeleton and the adjacent skeleton was extracted, and then the focused area of microscopic image was constructed. It was confirmed that a fairly well image mosaics can be obtained by this method. Key word: Skeleton transform; microscopic images; image mosaics; mathematical morphology; Laplacian operator 1.引言 图像拼接技术在数字视频、宇航探索、虚拟现实、环境智能、医学图像分析等领域都具有 广泛的应用前景[1] 。所谓图像拼接( Image Mosaics) 是指将存在重合区域的多幅图像进行无缝拼 接,重建成一幅较大视野的图像。应用图像拼接技术,对序列显微图像进行拼接,可以得到较大视 野范围的显微图像,从而能够很好地解决上述问题。 图像拼接技术中常用的拼接方法有：基于特征的拼接方法、基于灰度相关的拼接方法和基 于相位相关的拼接方法等。针对显微图像的特点，本文提出了一种通过对序列显微图像进行骨 架