An Adaptive On-Line Inspection Method Based on Singular Value Decomposition

Singular value decomposition(SVD) is an effective method of algebraic feature extraction. It has the stability, rotation invariability, brightness invariability and other important features. In this thesis, through autonomous learning in small sample space and extracting the SVD feature, the similarity calculation method of singular value feature is given, the similarity is used to recognition. This method significantly reduce the requirements for the training image, and it can be applied to wider fields. Finally, the method is testified by a experiment of button battery case.