“Carbon black” is the major influential factor in categorizine the quality rank of rubber. This paper presents a new approach to simplify the ranking procedure with limited samples by using Grey relational grading method while the ranking accuracy is ensured. The proposed approach is described as follows: First, rubber samples are sliced and snapshot into black & white images. These images are further enhanced by digital image processing methods before applying this new clustering approach. Then, the “carbon black” of a sampled rubber image is classified into ten ranks according to the method B of ISO11345 document. The measured data are processed by using the algorithm of partial Grey relational grading methods. The “carbon black” ranking of rubber is obtained by sorting and determining the maximum value of Grey relational algorithm. In order to verify this new approach, test results are compared to the results taken by standard statistical method as reference. The results confirm that this new approach can use fewer samples and preserve the ranking accuracy.
展开▼
