基于纹理抑制平滑滤波的织物图案识别

Fabric pattern recognition based on texture suppression smoothing filter

针对织物图案特征提取效率低的问题,提出以纹理抑制平滑滤波算法与Adaboost算法为基础的织物图案识别法。首先对织物扫描图案的纱线纹理进行滤除;灰度化后再分别对图像的3个特征进行提取,最终构建图像特征库,以此作为训练对象,采用Adaboost算法进行实验。结果显示,以纹理抑制平滑滤波算法为基础的识别法能够提高图案识别率。

In order to solve the problem of inefficient feature extraction of fabric pattern,a fabric pattern recognition method based on texture suppression smoothing filter algorithm and Adaboost algorithm was proposed.Firstly,the yarn texture of the fabric scanning pattern was filtered out.After graying,the three features of the image were extracted respectively,and finally the image feature library was constructed,which was used as the training object and tested by Adaboost algorithm.The results showed that the recognition method based on texture suppression smoothing filtering algorithm could improve the pattern recognition rate.