Background: Traditional dual-energy X-ray inspection technology serves to classify different materials based on the proper value related with effective atomic number. Purpose: This paper aims to solve the problem of the relevance between the proper value and material’s thickness (thickness effect)

which reduces the reliability of dual-energy inspection technology. Methods: A numerical method which combines Simpson formula and sequential quadratic programming (SQP) method is proposed to eliminate the thickness effect of dual-energy values. The process of X-ray inspection is simulated by Monte Carlo method based on MCNP. Probability-classification method is used to evaluate the effectiveness of diminishing error probability of material-classification

and classification boundaries of different materials. A practical X-ray inspection machine is used to verify this method. Results: The experimental results indicate that not only the thickness effect is diminished obviously by this method

but also the classification boundaries are effectively achieved. Conclusion: The dual-energy value of X-ray transmission image can be extracted by our proposed method with effectiveness and practical applicability. Abstract