为了解决传统图像合成方法在处理具有不同纹理特征的图像时产生明显边界的问题,提出新的图像合成框架.把图像分解为基本层和细节层,基本层解决边界处颜色的无缝融合;细节层实现边界处纹理细节的光滑过渡.算法流程为进行基本层的融合,合成新的细节层,新细节层添加到融合的基本层得到最终结果.通过计算一张权重图来指导新细节层的生成,权重图度量纹理特征的强弱和像素到边界的距离.实验结果显示:算法可以有效保证颜色的无缝融合,保持纹理的光滑过渡.相对于传统图像合成方法,该算法可以生成更真实的合成图像.
In order to resolve the apparent boundary issue that the conventional image compositing method encounters when applied to images of different texture details,a new image compositing framework was proposed,which decomposed an image to a base layer and a detail layer,the base layer implementing the seamless fusion of color at the boundary,while the detail layer implementing the smooth transition of texture details at the boundary.The processing was as follows: fusing the base layer,synthesizing a new detail layer and then adding it to the fused base layer to get the final result.A weight map was computed by the measure of texture details and distances of pixels to the boundary and used to guide the generation of new detail layer.Experimental results on many images of different texture features showed that the proposed algorithm can effectively realize the seamless fusion of color and smooth transition of texture details.The proposed method produced more realistic composites than the traditional method.