Surface Defect Target Identification on Copper Strip Based on Adaptive Genetic Algorithm and Feature Saliency

Table 6

Identification accuracy of RBF neural network.

Defect type	Total	Correct	Accuracy
Pits	126	112	88.9%
Loophole	342	323	94.4%
Burr	215	203	94.4%
Perforation	106	96	90.6%