基于计算机视觉的织物疵点自动检测方法研究

为了快速、准确地检测织物上存在的疵点,提出了一种基于计算机视觉的织物疵点自动检测方法。用具有高分辨率的CCD摄像头,在可移动云台及可调标准光源的照射下,捕获织物图像,存储至计算机;采用能量、方差、熵、极差、对比度作为特征值,进行织物图像的特征值提取与归一化处理;基于小波分析方法,提出自动选择子图像的小波重构方法,将织物图像中的疵点部分分离出来进行定位、识别。经试验证明,该方法具有检测速度快、定位准确等优点。

Study on Fabric Defects Automatic Detection Method Based on Computer Vision

To detect defects on fabric quickly and precisely,automatic detection method was proposed based on a computer vision.Image of fabric which were got from removable platform under adjustable standard light by high resolution CCD camera were captured and saved in computer.Eigenvalues were adopted including energy,variance,entropy,range and contrast,fabric image eigenvalue were extracted and normalized.Based on wavelet analyses methods,wavelet reconstruction of automatically choosing image was proposed,defects in fabric image were separate then located and recognized.The practice shows that the method has advantages of exacter locating and higher inspection speed.