While working on aircraft navigation on taxiways, the line detection is one of the main challenging problems to be solved. This subject has been widely studied in the literature in the automotive field. In this paper, we propose a comparison of three line detection algorithms based on methods validated in the automotive field but transposed in aeronautics where this subject has not been widely addressed. Some problematics appear: the tarmac environment differs from the usual road model and the camera's position impacts the visibility on the image. The first method presented here uses a particle filter while the second one is based on the Hough transform. In the second method, we perform a color-based detection and introduce a method to compute the reference color, using technical specifications for airport markings. The last method is the LaneNet neural network. Criteria such as the precision or the max range of the detection are computed and exploited to discuss the algorithms relevance. The comparison is performed on both simulated images (from a product of the OKTAL-SE company) and real ones (from Airbus Operations S.A.S.).