【作者】 龚奕刚；

【导师】 吴锡生；

【作者基本信息】 江南大学 ， 计算机应用， 2008， 硕士

【摘要】 如今,数字图像已经在工程和数学领域引起广泛关注,而图像插值技术也已被广泛应用于数字图像处理,例如图像缩放、图像变形、图像恢复、图像重建、图像配准等。图像插值就是利用已知邻近像素点的灰度值来产生未知像素点的灰度值,以便由原始图像再生出具有更高分辨率的图像。本文首先通过分析实际图像获取系统将图像数字化后,建立图像的模型,并给出了图像的表示方法,根据数字图像的特点阐述了图像放大的基本原理。其次,本文介绍了传统的线性移不变插值算法。主要介绍了最近邻插值、线性插值、三次卷积插值、多项式插值、样条插值和高斯插值等。通过对传统的线性移不变图像插值算法的分析,归纳了线性移不变图像插值的共同技术缺陷和它们的理论成因。然后,论文从线性空不变图像插值和距离加偏差图像插值两方面介绍了现代自适应图像插值技术,讨论了各算法的优势和局限性。自适应插值技术的插值效果比线性移不变插值技术有了较大的改善,但算法复杂计算量大,在实际图像处理系统中应用困难。本文提出了一种基于图像边缘信息的自适应图像插值算法。根据边缘部分映射点邻域图像的复杂程度,自适应地调节插值权值的图像插值方法。应用该算法插值后的边界清晰、自然,忠实地反映了原始图像的面貌,与传统的插值算法相比,其边界处理效果好且易于实现,实验也验证了该方法的有效性。更多还原

【Abstract】 Today, the digital image in the field of engineering and mathematics has aroused widespread concern, and the image interpolation technology has been widely used in digital image processing, such as image zoom, image deformation, image restoration, image reconstruction.Image Interpolation is known to use the adjacent pixels to generate unknown gray value of the gray pixels value to the original image from renewable to a higher-resolution images.In this paper, through analysis of the actual image acquisition system will be digital image after image of the established model, and gives the image of that method, in accordance with the characteristics of digital images on the image to enlarge the basic principles.Secondly, this paper, the traditional linear interpolation algorithm shift unchanged. Mainly on the nearest neighbor interpolation, linear interpolation, three convolution interpolation, polynomial interpolation, Gaussian spline interpolation, and so on. By shifting the traditional linear interpolation algorithm for the image analysis, summed up the same image interpolation linear shift the common technical defects and their causes of the theory. Then, this paper from the same linear space and image interpolation and deviation from the negative image interpolation even introduced the image of modern adaptive interpolation technology, discussed the advantages and limitations of the algorithm.Adaptive interpolation technology transfer linear interpolation effect than the same interpolation technology have been greatly improved, but the algorithm complex calculation, the actual image processing system in the application of difficulties. This paper presents a message based on the Edge of adaptive image interpolation algorithm. According to the edge of the neighborhood image mapping, the complexity of the adaptive value of the right to adjust the interpolation image interpolation. Application of the algorithm after the interpolation of the border clear, natural, faithfully reflects the look of the original image, with the traditional interpolation algorithm compared to its borders and easy to good effect to achieve ,the experiment also proved the effectiveness of the method.更多还原