Abstract: Traditional technology for detecting strip structure is time-consuming with low accuracy and large strip damage. Therefore image processing technology was combined with fractal theory by using the counting box dimension method to measure the fractal dimension of strips, and the fractal characteristics of strip area and Feret's diameter. The results showed that:1) The T-test (p=0.227>0.05) results for the measurements by IPP image processing software and square standard area indicated that there was no significant difference between the measured values of strip parameters and the real values. 2) Using the box counting dimension method to measure the fractal strip dimension, the determination coefficients of the fitting regression equation were all above 0.81 and the P values of the ANOVA results were all less than 0.05. 3) Both strip area and Fetet's diameter extremely positively correlated to the fractal dimension of the strips. The R² of fitted equation between the fractal dimension D and lnS and lnφ were 0.994 and 0.960, respectively. In other words, the influencing degrees of strip area and Feret's diameter on the fractal dimension were different. 4) The RMSE and SSE of the established fractal dimension model for the strips were 0.016 and 0.025 with the determination coefficient R² of 0.996, the AE and RE of the model were 0.014 and 1.130% respectively, and the model featured good fitting degree and high accuracy. 5) The fractal dimensions were 0.664 < D ≤ 0.829 for small strips, 0.829 < D ≤ 1.000 for medium and short strips, 1.000 < D ≤ 1.201 for medium and long strips, and 1.201 < D ≤ 1.615 for large strips.