摘要
论文针对图像放大问题,建立了矢量化填充放大模型。并针对矢量化填充放大模型不能处理复杂图像的缺点,使用并比较了最邻近插值法、双线性插值法、双三次插值法三种不同的算法。并综合以上四个模型,得到以下结论:矢量化填充放大模型适用于较简单的可矢量化的图像,可无限放大,且不影响其成像。基于插值算法的三个模型中,最邻近法模型算法简单、计算量小、处理速度快,但处理后的图像会出现块效应,造成图像模糊。双线性插值法模型克服了最邻近法处理图像出现块效应的问题,其计算相比双三次法简单,但边缘细微结构会受到影响,产生轮廓模糊的现象。双三次插值法模型解决了以上两种算法的缺点,计算精度高,带有边缘增强的效果,能够较好保持图像的细微结构,但计算量较大。
In view of the problem of image magnification,this paper establishes the vector filling amplification model.And fill the zoom model can't deal with complex vector quantization for image shortcomings,using the adjacent interpolation method,bilinear interpolation method,double three interpolation method of three different algorithms are compared.And integrated the above four models,the following conclusion is gotten:fill amplification vector quantization model suitable for simple images,vector quantization can be unlimited amplifier,and does not affecting the imaging.Three models based on interpolation algorithm,the neighboring method model algorithm is simple,small amount of calculation,fast processing speed,but after processing the image will appear block effect,cause blurred images.Bilinear interpolation model overcame the neighboring method processing image block effect problem,compared the calculation method of double three simple,but the edge structure will be affected,produce the outline of the fuzzy phenomenon.Double three interpolation model to solve the disadvantages of the above two kinds of algorithm,the calculation accuracy is high,with the effect of edge enhancement,can better keep the image structure,but the large amount of calculation.
出处
《计算机与数字工程》
2016年第6期1146-1150,1166,共6页
Computer & Digital Engineering
关键词
矢量填充
最邻近插值法
双线性插值法
双三次插值法
vector filling
adjacent interpolation method
bilinear interpolation method
double three interpolation method
