Font Size: a A A

The Interactive Deterioration Identification System Of Ancient Murals

Posted on:2017-12-23Degree:MasterType:Thesis
Country:ChinaCandidate:N ZhangFull Text:PDF
GTID:2335330512977433Subject:Pattern Recognition and Intelligent Systems
Abstract/Summary:PDF Full Text Request
Mural was one of the most precious cultural heritages of our country with great history and artistic values.However,due to the destruction caused by natural disasters such as shabby preserve environment,fragile materials,coupled with floods and dust in addition with man-made destruction,the existing ancient murals present different kinds of deterioration.The preventive monitoring of murals and restore of the deterioration requires different measures depending on the type of deterioration.One of the most important tasks is precisely determining the type of deterioration area.Dunhuang Mogao Grottoes and other sites have preserved huge amounts of murals.At present,the commonly used tool to label the deterioration of the mural is general utility tools AutoCAD which has poor professionalism,long learning cycle,complicated operation etc.Thus it can't satisfy the repair jobs.Through the survey of the deterioration labelling process,we propose a fully functional ancient mural labelling system.It has strong profession,short learned circles and simple operation merits.It proved by the reality used in Dunhuang Mogao Grottoes that our system reduces the heavy deterioration labelling works and extremely improves the labelling efficiency.The paper introduces the design and implementation of deterioration identification system of ancient murals.The main contribution lies in two aspects: adding the data statistics module and outputing of deterioration image to meeting the need for deterioration monitoring,and optimizing user operations and functions for the efficiency of the identification work,,which greatly enhance the support for large images and multi images.
Keywords/Search Tags:Ancient Murals, Deterioration, Auto-identification, Sparse Coding
PDF Full Text Request
Related items