Abstract:The appearance of warm ridge in low attitude implies that an unstable weather factor is forming.Therefore, recognizing the warm ridge correctly based on temperature field data is important for weather forecasting. In view of the exposed problems when extracting the warm ridge lines using the ridge line algorithm, this paper proposes a section extreme value correction (SEVC) algorithm, by which candidate warm ridge feature points are extracted, and establishes rules based on weather characteristics to exclude non warm ridge feature points. According to the idea of least squares method used in fitting curve, this paper puts forward a kind of smooth medial axis fitting (SMAF) algorithm to get higher quality of warm ridge lines, which coincide with that weather forecasters manually draw. A large number of experiments show that the recognition rate is as high as 97% and mistaken recognition rate is zero.