<p>High-speed railway is a comprehensive carrier of high-technology. As an important infrastructure of high - speed rail tracks, rail breakage and derailment accidents caused by rail rolling contact fatigue crack damage are quite common. Considering that a single eddy current pulsed thermography cannot directly and very accurately make a quantitative estimate of the geometric parameters of Rolling Contact Fatigue (RCF), this paper proposes a rail contact fatigue crack damage identification system and method based on the fusion of 2D optical images and eddy current thermography. A mathematical - physical model of Poisson reconstruction degree for the fusion of 2D optical images and eddy current thermography is proposed to quantitatively characterize the geometric parameters of rail cracks, such as length and depth. A rail contact fatigue crack damage identification robot system based on the fusion of 2D optical images and eddy current thermography is designed and built. Experiments are carried out to compare the detection accuracy of crack geometric parameter identification in static and dynamic (different motion states) modes, as well as to study the impact on the quantification of crack geometric parameters. The experimental results show that the proposed Poisson reconstruction degree based on the fusion of 2D optical images and eddy current thermography has good robustness in quantifying crack geometric parameters in the slow - speed mode, verifying the effectiveness of the system and method.</p>