Enhancement of polychrome pottery hyperspectral images based on multilevel feature extraction method
摘要
Polychrome pottery, as an ancient art form, embodies profound historical and cultural significance. This study focuses on a polychrome pottery lid unearthed from Zhaojiazui, Yunyang, using hyperspectral imaging for feature enhancement. Hyperspectral data from the visible to near-infrared range enables non-destructive analysis of fine surface details. A multilevel feature extraction method is proposed by combining Principal Component Analysis (PCA) and Sequential Maximum Angle Convex Cone (SMACC). PCA first identifies the principal component with the highest contribution, which is then inversely transformed to reconstruct a true-color image. SMACC is applied to extract endmembers, and HSV conversion is performed to enhance features. The final fusion of enhanced images results in improved visual clarity and detail, especially in blurred areas.