Information Index Technology For Digital Museum

As an important carrier of cultural preservation and communication, digital museum has an important role to preserve and the spread of traditional culture in the new era. Unlike static and boring display in entity museum, digital museum of high-technology and high experience has become a new direction. Information index technology for digital museum is a new technology based on the existing digital museum, using the idea of augmented reality technology in order to facilitate the human-computer interaction between the user and the entity exhibits. The main goal of this paper is the recognition and information index technology of the information index system.In the field of image recognition, local invariant features detection has become an important step because of good uniqueness and robustness. Effective feature selections directly determine the recognition accuracy, and also is the primary goal to complete the information related system. So choose a high robustness feature extraction algorithm has become the first problem to be solved of this paper. From the robustness of target recognition algorithm, this paper summarizes the reasons by analyzes the scale invariance, affine invariance and robustness to the change of illumination of existing local feature detection. Mainly from the real-time aspects of object recognition, we propose a fast recognition algorithm which can maintain invariance. For the change of scale, we cluster the extracted local feature points using adaptive kernel clustering method, and then match the feature points using multi-scale method. In order to improve the robustness of affine transformation, we draw on the idea of existing full affine invariant feature detection algorithm, and propose an affine invariant feature detection algorithm which is superior to the classical point feature extraction algorithm. The object recognition combining two improved algorithms keep robustness to a variety of affecting factors. Although it is not the best method, improved algorithm can rapidly and accurately complete the identification of museum exhibits target, and we reached the real-time requirements. For the information index after object recognition, The author proposes two different ways to show different information formats. The accuracy of object recognition is the base to calculate the homography mapping matrix. Using the homography mapping matrix, we can easily estimate the posture of the three-dimensional model, thus completing the three-dimensional model superimposed accurately. Through the establishment of target identification and information related subsystems, the author completed a simple information related system for digital museum and provides a direction for the development of new digital museum.