多光源图像细化和细节增强的协同图像处理算法研究

多光源图像合成处理过程中，为了增强图像的轮廓和表面细节，同时使合成后的图像边缘无伪迹、保真度高，提出了一种多光源图像细化与细节增强的协同处理算法。该算法通过梯度域法构建辅助层，采用二次过滤法提取出细节层，提出了新的阴影检测算法用来去除细节层的伪迹，同时使该细节层包含了输入图像的全部信息；使用一幅输入图像构造一个基础层并进行暗区亮度增强处理；细节层和基础层进行复合处理后，重现了暗影区丢失的细节，同时增强了现有的细节。该算法与其他算法进行图像处理对比实验，结果表明，该算法用于实现多光源图像细化和细节增强是可行的、有效的，采用这种方法合成的图像看起来更自然，相对于输入图像，输出图像保持了较高的保真度。该算法具有交互性，用户可以手动或自动调整合成图像的效果。

Multi-light source image defined and enhanced image processing algorithms

In order to enhance the image contour and surface details, and make the image edge without artifact after synthesis, high fidelity, the authors propose a multi light source image processing algorithm with details enhancement. The algorithm, through the gradient domain method to construct assist, secondary filtering method, is used to extract details layer, and a new shadow detection algorithm is used to remove detail layer artifact, at the same time make the detail layer contain all the information in the input image; Using a picture of the input image constructs a base layer and dark space brightness enhancement processing; After detail layer and base layer composite processing, the shadow area missing details are reconstructed. We enhance the existing details at the same time. The results of algorithm and other algorithms for image processing experiments show that the algorithm is used to realize the source image and detail enhancement is feasible and effective; using the method of synthetic image looks more natural; relative to the input image, the output image remains high fidelity. The algorithm is interactive; the user can adjust the effect of the composite image manually or automatically.