Template Matching

Image matching, which involves finding the corresponding positions of the same or similar objects in different images, is a fundamental problem in computer vision. Based on image matching, a series of computer vision tasks such as image retrieval, image stitching, and image recognition can be accomplished. The most basic and simplest form of image matching is template matching: given a template image, find the target location in another image that has the highest similarity to the template. For example, as shown in Fig. 4.1 , given a cloud as the template image, it is necessary to determine whether the image contains a cloud, and if so, the location of the cloud. Depending on the number of targets, template matching can be divided into single-target template matching and multi-target template matching. In this chapter, we will introduce the steps of template matching and the methods used to measure similarity for matching.