Corrosion Character Extract Algorithm Based on Correlation Dimension of Fractal

For the defects testing of civil aviation aircraft maintenance, an extraction algorithm of corrosion feature based on fractional dimension of fractal is proposed. Firstly, the ration denotation method of corrosion phenomena in fractal dimension is studied, and then distance vectors between the boundary and the centroid are ascertained by fractal feature of corrosion region´s contour and the vectors is fitted. Secondly, the feature value of corrosion contour which results from fitting is modified depending on self-similar of step variation when the corrosion region is contracted. At last, the fitted value is regarded as a correlation dimension and the corrosion defects of aircraft skin is detected and recognized. The experimental results demonstrate that the proposed algorithm is easy to extract feature of corrosion of aircraft skin, and it has high recognition ratios and robustness for corrosion detection, especially, it can meet the real-time request.