摘要
明中都石刻现存六百余件,是中国建筑雕刻艺术的巅峰,具有极高的历史价值、文化价值、艺术价值.明中都石刻多散落在凤阳县内,且多已消失、残损,严重影响了石刻的保护和修复工作.面对如此数量庞大且类型繁多的石刻,传统的手工整理分类方法显然不能满足当前的需要.利用人工智能技术对石刻图像进行预处理、提取特征、归类分析、建立特征库等步骤实现图像的自动识别和分类,为研究提供了更加丰富、精准的数据资料.
Abstract
There are more than 600 Zhongdu stone carvings in the Ming dynasty,which are viewed as the peak of Chinese architectural carvings with high historical,cultural and artistic value.At present,most of Zhongdu stone carvings in Ming Dynasty are scattered in Fengyang County,and most of them have disappeared or been damaged,which seriously affects the protection and restoration of stone carvings.Faced with various types of stone carvings,traditional manual sorting methods cannot meet the current needs.Automatic recognition and classification of images can be realizes by using artificial intelligence technology,such as preprocessing stone images,extract features,classifying and analyzing them,building feature database,which will provide more abundant and accurate data for the research.
基金项目
安徽省科学研究重点项目(2022AH051604)
安徽省高校自然科学研究项目(KJ2021A0863)
安徽省霍邱县省级农业科技园区"建设实施方案及验收项目"(881046)
安徽省凤阳县规划局"耕地恢复实施方案项目"(881366)
NETL
NSTL
万方数据