基于数字图像处理的非织造布均匀性检测

为测量薄型非织造布的均匀性,采用数字图像处理技术,通过目标区域提取、中值滤波及Otsu二值化等操作,对图像进行预处理,再对图像进行分块后分别计算各块的纤维覆盖率,最后计算出变异系数,并与取样称重法的结果进行比较。结果发现:图像分析法的结果与取样称重法的结果具有高度的趋势一致性;通过阈值校准之后,图像分析法结果与取样称重法结果非常接近,这验证了图像分析法的有效性;此外,相较于取样称重法,图像分析法效率提高了50倍,高效、准确。

Uniform measurement of nonwovens based on digital image processing

In order to measure the uniform of light-weight nonwovens, the digital image processing was adopted to preprocess the image with target areas'extraction, median filtering and Otsu binarization, then the image was partitioned to calculate the fiber coverage of each partition, and finally the coefficient of variation was calculated, which was compared with the result of weighing method.The results showed that, the result of image analysis method was consistent with the trend of the weighing method`s result.With the threshold calibration, the result of the image analysis method was close to that of weighing method, which proved the effectiveness of image analysis method.Furthermore, the image analysis method was efficient and accurate, which efficiency was 50 times higher than that of the weighting method.