基于离散高斯滤波器的纺织品图像增强

Enhancement of textile images based on discrete gaussian filter

在纺织品中,常需要对织物纹理等性能进行分析。基于计算机视觉技术的纺织品分析,对提取的织物图像要进行图像预处理,增强图像,为后续分析提供有效数据。采用不同模板的均值滤波器和高斯滤波器来增强图像。由实验得出高斯滤波器去除常见的高斯噪声效果好,但模板并不是越大越好,模板尺寸约为9×9时的滤波效果较好。

The analysis of textiles property, such as texture, is generally acquired in textile research community. Analyzing the textile based on computer vision technology, which need image pre-processing and image enhancement, provides valuable data for the next analysis. Textile images were enhanced by using mean filter and gaussian filter with different templates. From the experiments, it could be shown that gaussian filter has good efficiency in disposing ordinary gaussian noise. However, larger template size does not necessarily mean better efficiency and the Efficiency of the template with about size 9 × 9 is preferable to that of others.