Font Size: a A A

Research On The Key Technology Of Salient Regions Extraction And Matching Of Terra-cotta Warrior’s Fragments

Posted on:2015-03-31Degree:MasterType:Thesis
Country:ChinaCandidate:N HaoFull Text:PDF
GTID:2268330428977217Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
Fragments of Terra-cotta Warrior are complex cultural relics’models which have a lot of data noise, complex structures, large quantity and incomplete information. The traditional method that reassembles fragments directly is very time-consuming. If we match fragments with the feature-region-library of body-parts firstly, determine which parts they belong to and reassemble them only inside of their own libraries of body-parts, the whole efficiency will be improved. It will cause second damage on cultural relics that matching them in the help of the archaeologists only, so this paper researches on fragments’matching facing their reassembling:(1) Propose a feature points’extraction method based on integral invariants. At first, give the definition of integral invariants and use volume integral invariants to measure the bending degree of a model’s surface. Then calculate value of integral invariants at each vertex on surface. At last, use the region-expansion method to define local surface patches iteratively and select their central points as feature points.(2) Implement a feature regions’extraction method for Terra-cotta Warrior’s fragments based on the region-growing method. Merge local surface patches which are represented by feature points to maximize the salience grade of a local region using the region-growing method. Take those local regions that have higher salience grade as a model’s salient regions.(3) Propose a matching method for fragments based on prior knowledge and salient features. Use the affine ICP algorithm to measure the similarity between two salient regions. Compare each salient region of a fragment with the feature-region-library to determine its body-parts. At last, combine the results and prior knowledge to obtain the fragments’ body-parts semi-automatically. The experimental results show that taking salient region as unit to match fragments can reduce the complexity greatly and that the accuracy of matching results is relatively high.(4) Design and develop a Terra-cotta Warrior’s fragments-matching system. The system which is easy to operate, has strong interaction and can obtain fragments’ parts information quickly provides constraints and guidance for reassembling and restoration of damaged Terra-cotta Warrior.
Keywords/Search Tags:Integral invariants, Salient regions, Feature regions matching, Affine ICP, Fragments matching
PDF Full Text Request
Related items