Numerical Pattern Identification-Application to Inductive Testing Method With Automatic Classifiers

作者：Gizewski Tomasz^*; Goleman Ryszard; Stryczewska Henryka Danuta; Wac Wlodarczyk Andrzej; Nafalski Andrew

来源：IEEE Transactions on Magnetics, 2013, 49(5): 1789-1792.

DOI：10.1109/TMAG.2013.2244200

• Automatic classifiers
• hysteresis
• nondestructive testing

Automatic algorithms which include classifiers require effective systems of data acquisition, data modeling or other data source in order to create probability groups. Their role is to process the information to the basic structure of the model with a significant number of details. Owing to the differences between the probability groups, the classifier allocates the images to a selected class. At the same time the assessment of details%26apos; quality is created. The main topic of this article concerns the numerical modeling of a closed loop, generated from the study of the unbalanced voltage of the Maxwell bridge. It forms an image of material defects, determined by the numerical study of the sample and pattern that were analyzed in the work by changing the shape of a closed loop due to changes in size of the defect. The changes in the shape of the closed loops, as results of changes in the defect size were analyzed.

• 出版日期2013-5
• 单位中国科学院电工研究所