The Recognition Method Of Ceramic Microcosmic Images Based On SURF And Blockchain Storage

Ceramics in China have a long history and possess significant cultural and artistic value,representing a dazzling gem in traditional Chinese culture.Ceramic artworks embody unique artistic value,reflecting humanity’s pursuit of life and aesthetics,and they hold high collectible worth.With the development of the market and the promotion of art,ceramics have become an extremely appealing investment.However,the increasing advancement of counterfeit ceramic production techniques has led to a proliferation of fake ceramic products in the market.This behavior has seriously hindered the normal development of the ceramic art investment and collection market.Traditional ceramic authentication methods heavily rely on experts’ experience and expertise.While physical and chemical identification methods may cause irreversible damage to ceramic artworks themselves.Therefore,this study proposes an identification method based on the distinctive microstructural features of ceramic surface images,using Accelerated Robust Features(SURF)matching and blockchain storage.To achieve higher identification accuracy,this paper compares several commonly used feature extraction algorithms from multiple angles,ultimately selecting the SURF algorithm that best suits the purpose of this study for analysis and improvement.The proposed method utilizes an improved SURF algorithm to extract image features.Initially,a bidirectional Fast Library for Approximate Nearest Neighbors(FLANN)matching strategy is employed to preliminarily filter out incorrect features.Subsequently,the Random Sample Consensus(RANSAC)algorithm is utilized to filter out unreliable matching point pairs.Finally,a judgment is made based on the matching accuracy of the remaining reliable matching point pairs.By comparing the proposed algorithm with other typical feature extraction algorithms,experimental results demonstrate that the proposed algorithm achieves improvements in both recognition efficiency and accuracy.The average matching similarity on the ceramic microstructural image dataset exceeds 90%,and it also exhibits good robustness under complex conditions such as rotation,blur,brightness variation,and scale transformation.In order to enhance the credibility of ceramic identification and traceability work,it is essential to avoid using traditional centralized data storage methods.The decentralized nature and immutability of blockchain technology greatly ensure data security and reliability.However,blockchain’s handling speed and storage efficiency for large-scale image data are relatively low,making it challenging to meet the increasing data storage demands of ceramic authentication work.To address this issue,this study introduces the Inter Planetary File System(IPFS)in conjunction with blockchain technology to establish a distributed storage network.By utilizing IPFS for data storage and leveraging blockchain technology to guarantee data integrity and security,a decentralized storage solution is provided for ceramic authentication,specifically targeting microstructural identity features.In the actual authentication process,the comparison and verification of the target image with the image features on the blockchain enable the prevention of counterfeiting and the traceability of ceramic artworks.