A new fuzzy based fast and high definition template matching algorithm

The accuracy of automatic sample pages pasting seriously affects the efficiency of sample books´ production. To achieve a more rapid and low cost production of sample books, according to the structure characteristics of the sample pages, a new high speed, high precision sample pages detection algorithm is proposed. The algorithm combines correlation coefficient based template matching and image space transformation, designed a novel real time sample pages image detection system. The algorithm has improved the standard template matching method with image histogram similarity structure characteristics of template matching to meet the high accuracy of real-time image processing requirement. The use of image spatial transformation for target plate location has dramatically improved the accuracy of the feature point´s location, and reduced more than 80% calculating burden approximately. The experimental results show that the time cost running by two 5 megapixel camera synchronously is within 100ms. This algorithm not only can accurately match and locating the template, and significantly reduce the overall computational complexity, but also can effectively improve recognizable efficiency and robustness, gets a better performance than previous system.