Carving surface modeling of image based on Gaussian RBF

Carving surface modeling is directly established using the RBF interpolation algorithm according to the pixels information of the image with clear texture and relatively simple structure. This method segments the image, and analyses and computes the pixel values and their position parameters for the sub-images, and those feature parameters are used for construction of the surface patches based on quadratic linear combinations of Gaussian functions. The modeling is implemented by limited smooth connection of the retrieved Gaussian surface patches with good slickness and positive definiteness. The established surface model can represent the basic structure characteristics of the original objects in the image, and satisfy the requirements of digital carving processing. Relatively low algorithm complex and strong practicability are the features of this method.