针对建筑物图像线条鲜明、轮廓清晰的特点,提出一种自动化的抽象画生成算法.首先通过双边滤波对原始图像的细节进行简化;然后利用快速的、带有误判控制的图像线段检测算法来获得建筑物的外轮廓和门窗等线条特征,通过该线条特征增强基于流的高斯差滤波器生成线条框架;最后将双边滤波处理结果与线条框架进行融合形成最终的抽象画效果.该算法既去掉了建筑物图像中多余的细节信息,使整个图像更趋抽象,同时又通过边缘信息保持了建筑物自身独有的风格,充分展现出富有艺术家手绘风格的建筑物效果.文中算法简单、易于实现,并通过实验充分证明了其有效性.
In this paper we present an automatic abstraction algorithm for building images. Firstly, the bilateral filter technique is used to remove the redundant details of the building image~ and then the fast line segment detector with a false detection control is used to detect the outlines of the door, window or other edge features. Based on the line segments detected, a line-drawing image is generated by Flow-Based Difference-of-Gaussians filter to keep building's sketchy and clear outline features; at last the image from bilateral filtering and the line-drawing image are integrated to form the final abstraction result. With the presented algorithm, the redundant information is removed to abstract the detail content of the image, while a novel line enhanced technique is used to strengthen the special style of the building, which is similar to artist's hand-painting. Our algorithm is simple and easy to implement. And the experimental results verified the effectiveness of the algorithm.