Method of quickly building target model based on meaning of regular expression

The invention discloses a method of quickly building a target model based on the meaning of a regular expression, and belongs to the field of machine vision and mode identification. First, an image where an identified object is located is preprocessed in order to improve the quality of feature extraction. Then, the local features of an image target are extracted through a Harris feature extraction algorithm. Finally, a target matching model of the identified object is described through the meaning of an extended regular expression and a defined image material library. By extending the 'regularization theory' in the traditional computer theory, the method can be applied to the field of image target identification. The characteristic of high retrieval efficiency is retained. The method has a very good effect under the condition that the contour, color and other physical properties of an identified object are constant. 本发明公开了种基于正则表达式语义的目标模型快速构建方法，属于机器视觉与模式识别领域。首先对被识别物体所在图像进行预处理，以提高特征提取的质量，其后通过Harris特征检测算法提取出图像目标