Recognition of Fracture Image Based on Gray Level Co-occurrence Matrix

Fracture image recognition is an important part of failure analysis,but the problem of low recognition rate exists at large.A recognition method of fracture image based on gray level co-occurrence matrix is presented in this paper.Gray level co-occurrence matrix method is applied to metal fracture image recognition by extracting features,and separability criterion is used to obtain building factors including the forming step d,the gray-level of image g and forming direction θ which can affect texture feature parameter.And redundant feature parameters which are used to express metal fracture image texture are eliminated by correlation analysis.Finally,BP neural network classifier is used to classify the fracture image.Experiment results show that it is feasible to classify metal fracture image and the recognition rate is satisfied based on this method.