景物三维重建的关键之一是从景物图像中获得景物目标到摄像机的距离。研究了一种基于散焦图像的计算景物距离的算法。该方法基于不均匀散焦模型,并且只需两幅改变光圈指数得到的散焦程度不同的图像,因此可以避免图像的大小匹配问题。根据图像的点扩散函数的形式,可以通过优化的方法求得深度。模拟和真实实验表明了算法的有效性。
The key of 3-D reconstructing is to determine the depth information of a scene from images.A new method for depth from defocus images based on uneven defocusing model is proposed.This method only needs two images taken with different aperture diameters and it is valid without process of scale normalization or images matching.Experimental results on synthetic and real images are presented to demonstrate the effectiveness of the proposed algorithm.