一种基于样本传播的图像风化方法

对图像中的物体进行风化处理是一种提高模拟场景真实感的有效方法。利用一种图像处理方法,实现了图像中物体表面的精细风化效果。首先,利用径向基函数(RBF)方法计算风化程度。即使物体表面的颜色变化很大,该算法也可以较精确地计算风化程度。然后,使用图像分割算法来提取风化程度最深的区域作为"风化样本"。最后,利用patchmatch图像修补算法对"风化样本"进行补全,以此在图像上的待风化区域进行"风化样本"的无缝传播。结果表明,此方法可以产生各种类型的风化效应。利用样本传播的方法可以更好地模拟风化图像。

An Image Weathering Method Based on Exemplar Propagation

In order to improve the realism of the simulated scene, weathering the objects of the image is an effective method. Using an image processing method to achieve the fine weathering effect of the objects surface in the image. First of all, use the radial basis function (RBF) method to calculate the weathering degree image. Even if the color of the surface of the object changed greatly, the algorithm could calculate the degree of weathering more accurately. Then, use the image segmentation algorithm to extract the most weathered region as the "weathering exemplar". Finally, the patchmatch image repair algorithm was used to complete the "weathering exemplar", so that the "weathering exemplar" could be propagated seamlessly in the weathering area of the image. The results show that the proposed method can generate various types of weathering effects. Using the method of exemplar propagation can better simulate the weathering image.