In yarn evenness inspection based on image processing,due to serious noise,morphological method can not separate yarn from the blackboard entirely.In order to solve this problem,this paper proposes a denosing method t...In yarn evenness inspection based on image processing,due to serious noise,morphological method can not separate yarn from the blackboard entirely.In order to solve this problem,this paper proposes a denosing method that combines morphology with the image decomposition based on partial differential equation(PDE).Before denosing automatical skew detection is introduced to avoid the calculation error caused by image tilting.The experimental results show that the hairiness can be removed completely,and the characteristics of yarn can be reserved.The inspection result is well coincided with visual assessment.展开更多
文摘In yarn evenness inspection based on image processing,due to serious noise,morphological method can not separate yarn from the blackboard entirely.In order to solve this problem,this paper proposes a denosing method that combines morphology with the image decomposition based on partial differential equation(PDE).Before denosing automatical skew detection is introduced to avoid the calculation error caused by image tilting.The experimental results show that the hairiness can be removed completely,and the characteristics of yarn can be reserved.The inspection result is well coincided with visual assessment.