基于Radon变换的机织物纬斜检测方法

利用图像处理技术分析纺织品的参数、确定其品质是纺织工业领域的一个发展趋势。为了准确检测织物的纬斜角度,提出了一种基于Radon变换的机织物纬斜检测方法。通过全1矩阵补偿织物图像的Radon变换,使其方差-倾角特性曲线呈现单一的峰值特性,将纬斜检测问题转化为峰值搜索问题,提高了检测的可靠性。提出了一种变尺寸搜索算法,利用图像尺寸的变化加快了纬斜检测速度,减少了检测耗时。实验结果表明,在合理的参数配置下,该方法用于检测机织物纬斜,能够确保不大于0.15°的检测误差,单次检测耗时不超过30ms。

Detection method for weft skew based on Radon transform

A developing trend is to utilize the image processing technique to analyze the parameters of textile and to determine the quality of the textile in the field of the spinning and weaving industry.The detection method for weft skew based on radon transform is provided to detect accurately the weft skew of fabric.The variance-dip characteristic curve presented the single peak value feature through radon transform of the All 1 matrix compensation fabric image.The problem of fabric weft skew was converted to the problem of the peak value searching.The reliability of detection was raised.One varying size searching algorithm was provided to utilize the change of image size to quicken the speed of weft skew detection and to reduce the time of the detection.The experimental result demonstrated that this method could be used to detect the fabric weft skew,the variance of the detection was limited 0.15 degree under the rational parameter configuration and the time was not exceed 30 ms per detection.