中文摘要：

为解决显微镜观测视野狭小的问题，提出了一种基于尺度不变特征变换（SIFT）特征的显微图像自动拼接方法。该方法充分考虑了显微图像采集过程以及显微图像自身的特点，通过一系列适用于显微图像拼接的算法，控制单幅图像拼接时间在1秒内，进而可以保证显微图像拼接随图像采集在线完成，同时，利用SIFF特征匹配准确度高的优点，可达到亚像素级的拼接精度。该方法拓展了显微镜的应用，尤其适用于生物和医学显微图像观测。

英文摘要：

To solve the problem of narrow field of view in microscope, this paper proposes a new method of microscopic image stitching based on scale invariant feature transform （SIFT） features. This approach fully considers the properties of image acquisition and microscopic images and uses a series of algorithms suitable for microscopic image stitching, so it can limit its single inlage stitching time to 1 second. This means the on-line microscopic image stitching. For the high precision of the SIFT feature based matching, this approach can give the sub-pixel stitching accuracy.