Radon transform has important value in mathmatics theory, which can be used in algorithms for detecting linear features in an image. However, it has difficulty in detecting line segments in image, because Radon transform can not provide information about the positions of endpoints and the length. These problems become more serious when the transform is applied to an image with a high level of noise. The algorithm's suitability is discussed for linear feature detection in low SNR image, and this algorithm's advantages and drawbacks are analyzed. Then an accessional strategy based on information fusion is put forward to remedy the drawbacks. Aiming at low SNR image of airdrome's runway, the new algorithm reveals the preferable performance of speed and location.