基于压缩感知的织物疵点分类研究

Fabric Defects Classification Based on Compressed Sensing

针对织物疵点图像特征,提出了利用压缩感知提取织物疵点特征数据的方法.为了提取到疵点特征,需利用感知波形对图像的数据矩阵进行感知得到特征数据,同时对分类器进行并联设计,继而通过仿真和实验,检测出疵点类型信息.实验表明,该方法能有效地分类出织物疵点的数据信息.

Based on charactersitics of fabric defects image,this paper proposes the use of compressed sensing to extract charactersitic data of fabric defects. In order to implement this extraction,sensing waveform is needed for handling data matrix to obtain characteristic data,while the classifiers must be in parallel design,followed by simulation and experiment,to obtain defect type information. Experiments show that this method can effectively classify the fabric defect data.