图像特征特点及常用的特征提取与匹配方法（Methods of feature extraction and image matching features and common features）.docVIP

图像特征特点及常用的特征提取与匹配方法（Methods of feature extraction and image matching features and common features） Methods of feature extraction and image matching features and common features The image features are commonly used color features, texture features, shape features and spatial relationship feature. A color feature (a) features: color feature is a global feature, describes the surface properties of the corresponding image or image regions in the scene. The general color feature is based on pixel, then all belong to the image or image pixel regions have their own contribution. The color of the image or the image is sensitive to the change of regional direction and size, so the color characteristics of local features is not well capture the object in image. In addition, only using the color features of the query, if the database is large, often many will not need the image is retrieved. Color histogram is the most commonly used method to express the color features, its advantages are not affected by the image rotation and translation changes, further use of the normalized is not affected by change the image size, the disadvantage is that there is no expression of the base color spatial distribution information. (two) methods commonly used feature extraction and matching (1) color histogram The utility model has the advantages of simple: it can describe the distribution of global image color, different color for the whole image proportion, especially suitable for the automatic segmentation of the image to describe those images and do not need to consider the spatial position of an object. The disadvantage is that: it can not describe the local distribution of image color and the spatial location of each color of the object that can describe a specific object in the image or. The most commonly used color space: RGB color space and HSV color space. Matching method of color histogram feature: histogram intersection method and distance method, center distance method, the refer