结合图像频域和空间域的纬编针织物密度检测方法

为解决当前基于图像处理测量纬编针织物密度准确度不高、稳定性较差等问题，提出一种结合频域和空间域来测定针织物密度的方法。首先通过离散小波变换重构针织物线圈图像获得线圈结构清晰的图像，再分别提取线圈灰度及二值图的灰度曲线。利用提出的基于概率密度统计的波谷坐标校验算法，结合线圈坐标校验算法获得实际线圈所在列坐标；再利用八邻域广度优先搜索算法求出纬编针织物横、纵方向的线圈个数，得到针织物横密和纵密。结果表明，密度检测方法与人工测量数据相比误差小于1.7%,该方法适用性好，运算量小，准确率高，可实现纬编针织物密度的自动化测量。

Density detection method of weft knitted fabrics making use of combined image frequency domain and spatial domain

Aiming at the problems of low accuracy and poor applicability in measuring the density of weft knitted fabrics based on image processing, this paper presents a method for locating the loop position based on the gray curve of weft knitted fabric image to measure the fabric density, which highlights the loop image based on discrete wavelet transform. By extracting the gray curves of weft knitted fabrics and binary image respectively, the coordinates of the actual loop column were obtained by using the trough coordinate verification algorithm based on probability density statistics and the loop coordinate verification algorithm proposed in this study. The eight-neighborhood search algorithm was used to calculate the number of loops in the transverse and longitudinal directions of the weft knitted fabric, and the course and wale densities of the weft knitted fabric were obtained. Compared with the existing methods based on image processing, the method proposed in this paper involves less computation, and the error is less than 1.7％ compared with the standard data, which shows that this method has fast measurement and high accuracy, and it is beneficial for automatic measurement of weft knitted fabric density.