This paper proposes a method for the classification and identification of glass artefacts based on a K-mean clustering model. Using weathered and unweathered glass with high potassium and lead-barium as the main constituent as the main samples, a cluster analysis of their chemical composition was carried out to derive subclass subdivision and extract characteristic chemical components, based on which a sub-classification model of unknown glass based on the K-mean clustering model was established to provide a new method for the identification of glass artefact types. The results were analysed for reasonableness and sensitivity, and the model was proved to be of generalisation and application value.

No abstract

Reducing the number of comparisons is the most common way to improve the effectiveness of data cleaning. We investigate the problem by using inconsistency. We split redundant data into three categories. For each category, we give an algorithm and analyze its complexity, and combine them together finally. In particular, we address the chasing problem for the method under functional dependency. At the last, we experimentally verify that these algorithms effective and scale well, and that the method helps us more efficiently detecting duplications.

Applying the bilateral filter, Canny method to detect the edge of the image, the Morphological operation to modify the result, we proposed an improved SURF algorithm. The improved SURF algorithm can effetely handle the image match problems where the image undergoes noise and texture blurry. We utilize the improved SURF algorithm to deal with the infrared image recognition problem and obtain a good result.