The identification of the boundary geometry is an important branch of the inverse heat conduction problem and is the theoretical basis for the development of thermographic non-destructive inspection technique from the qualitative to quantitative status. For the inverse geometry conduction problem

the boundary of the computational domain has to be updated during the iterative process

which will obviously increase the complexity of the algorithm and the computational time needed. This paper presents a new boundary identification algorithm based on the estimation of the effective thermal conductivity. The method

which converts the boundary estimation in an irregular-shaped computational domain to the identification of the effective thermal conductivity in a regular-shaped domain

can obviously reduce both the complexity and computational time needed for the inverse boundary identification problem. Numerical experiments have certified the effectiveness of the method. The initial boundary guess and the temperature measurement error have no obvious effect on the boundary identification result of the plate.