Abstract: Ice accretion detection is an important means to ensure flight safety and an important issue in the field of aircraft anti-icing. In this paper, the method of identifying the boundary between the ice surface and the interior is discussed by using the infrared thermal wave detection technology. With a flash infrared thermal wave detection system established, regular ice accretion samples and ice accretion samples with internal boundary were made, the ice accretion detection experiments were carried out, and the data of the infrared thermal wave sequence were collected. In addition, the traditional algorithm based on the first-order differential operator and the second-order differential operator was exploited for processing the ice edge. A new boundary recognition method was proposed as well, which combined the gauss-Pierre-Simon Laplace pyramid algorithm and the area filtering algorithm. Then, the feasibility of the proposed algorithm to identify the boundary of the ice accretion surface was discussed and compared. The experiments and the image data processing methods show that the traditional algorithm can successfully recognize the outer boundary of ice accretion, but can not accurately recognize the internal boundary of ice accretion. The new fusion algorithm can effectively recognize the ice edge and the internal boundary, but the image noise is higher than that of the traditional algorithm. It can be concluded that the new fusion algorithm has some advantages in the detection of the irregular icing surface, and it is expected to provide a new research idea for icing detection in the field of aircraft anti-icing.