Bas-relief modeling from RGB monocular images with regional division characteristics

Traditional Bas-relief modeling methods are often limited to inefficient and difficult to be altered after the product is formed. This paper presents a novel method for bas-relief modeling from RGB monocular images with regional division characteristics. The problem discussed in this paper involves edge detection, region division, height value recovery and three-dimensional reconstruction of image. In our framework, we can automatically obtain the pixel height of each area in the image and can adjust the concave-convex relationship of each image area to obtain a bas-relief modeling which can be printed directly in 3D. The edge detection algorithm used Gaussian difference pyramid to combine the luminance information and chrominance information of digital image; the regions of the RGB monocular image are divided by the improved connected-component labeling algorithm;and the 3D pixel point cloud of each region is calculated by the shape-from-shading algorithm. Different from previous work, our method can fully obtain the image height field data and completely restored the depth information,which makes it possible to use any RGB monocular image for bas-relief modeling. Experiments with groups of images show that our method can effectively generate bas-relief modeling.

Automated summary

This paper presents a novel method for bas-relief modeling from RGB monocular images, which adjusts the concave-convex relationship through regional division and three-dimensional reconstruction, and enables direct printing of 3D bas-relief models.