基于自适应正交小波的机织物密度自动检测的研究

Automatic Inspection of Weaving Density for Woven Fabrics Using Adaptive Wavelets

为了将计算机视觉应用于织物经纬密自动检测,提出了建立机织物自适应小波分解滤波器并研究了织物密度自动检测的方法。应用织物自适应滤波器对织物灰度图像进行小波分解,就可使分解后的子图像能很好地分别包含织物的纬向和经向纹理信息,从而可将织物子图像二值化后,根据其黑白交替变化规律,计算出织物的经纬纱密度。实验结果证明自动检测和人工测量相比,误差在3%以内。

A method of automatic detection of woven fabric density based on adaptive wavelets filters is proposed for the application of computer vision to fabric inspection. The algorithm of adaptive wavelet and the decomposition of fabric image with it are introduced. The fabric images can be decomposed into vertical and horizontal sub-images. Then the weaving density of woven fabrics can be inspected by thresholding the sub-images. The results show that the relative error of automatic inspection compared with manual inspection is less than 3%.