机织物图像自动纠偏方法的研究

图像纠偏在织物组织结构参数自动识别过程中直接影响识别结果.本文通过对织物图像去噪、直方图均衡化等预处理操作后,对图像进行水平和垂直边缘检测;同时为提高倾斜角度识别的准确度和算法运行效率加入细化处理,然后对细化后的图像进行Hough变换,检测倾斜角度,最后对图像进行旋转实现旋转纠偏.通过对不同机织物图像进行纠偏处理实验,证明了该方法的有效性和可靠性.

Research on the automatic skew rectification of woven fabric image

The skew of woven fabric is very important to the identification of fabric weave pattern,especially effecting the change of result.The level and vertical edge of fabric images are detected by the preprocessing operation,such as image denoising and histogram equalization.The accuracy and the efficiency of the algorithm are added to the refinement process,then the Hough transform is used to detect the skew angle.Finally,the image is rotated.The validity and reliability of the method are proved by experiments of different fabric images.