基于排列组合熵的铝合金腐蚀灰度特征提取方法

Aluminum Alloy Corrosion Gray Feature ExtractionMethod Based on the Permutation Entropy

摘要:

针对铝合金试件腐蚀图像包含的大量腐蚀信息,提出了利用排列组合熵提取其灰度图特征参数的方法。该方法具有计算快速简单,抗噪能力较强,只需较短的序列长度就能估计出较稳定的系统特征值的特点。计算结果表明,基于排列组合熵的特征提取方法可以有效地提取各时间段内试件的灰度图在2个方向上的特征参数,熵值随着腐蚀时间的增加而规律性地增加,并可以作为评价腐蚀程度的一个重要特征参数。计算结果也验证了排列组合熵具有较好的抗噪性能。

Abstract:

For the aluminum alloy specimen corrosion images containing a wealth of information, the use of permutation entropy to extract grayscale characteristic parameters was proposed. The method has the following characteristics, counting quickly simple, the resisted chirp ability strongly, and estimating more steady system characteristic value by only needing shorter alignment length. The result showed that the feature extraction method based on permutation entropy could effectively extract the characteristic parameters in all directions of the specimen grayscale in different time. Entropy increased regularly with increasing corrosion time, could be used as an important parameter for the evaluation of corrosion degree. PE was verified to have a better anti-noise ability.