abstract |
The invention provides a method for extracting and classifying surface defects of foamed nickel. The steps of the present invention are as follows: firstly, preprocess the nickel foam image; then, extract the defects from the preprocessed image, and finally use the geometric feature to train the BP neural network to classify the defects. The method proposed by the invention effectively extracts seven typical defects of foam nickel edge cracks, scratches, scribes, voids, missing plating, gaps and through holes, and can realize efficient and accurate classification of defects. |