Thunderstorm winds are recognized as a highly sudden and destructive type of disastrous weather. The bow echo, a characteristic radar signature associated with thunderstorm winds, is considered critical in early warning systems. To improve the timeliness of thunderstorm wind warnings and minimize associated losses, accurate identification of bow echoes is required. For enhanced bow echo recognition accuracy, a bow echo identification algorithm based on morphological operations and curvature fitting is proposed in this study.First, low-elevation reflectivity data are processed using threshold segmentation (reflectivity threshold ≥45 dBZ) to extract strong echo regions. Next, erosion and dilation operations are applied to eliminate discrete noise, followed by the removal of outliers via the Local Outlier Factor (LOF) algorithm. The Breadth-First Search (BFS) algorithm is then employed to traverse connected regions, enabling extraction of the bow-shaped skeleton. Finally, B-spline fitting is utilized to identify the apex of the bow echo, allowing both the bow echo structure and potential high-wind zones to be detected.To validate the algorithm, 15 bow echo weather events from the U.S. NEXRAD dataset were tested. Results showed a Critical Success Index (CSI) of 0.87, Probability of Detection (POD) of 0.87, Miss Rate (MR) of 0.13, and False Alarm Rate (FAR) of 0, demonstrating reliable performance in bow echo recognition.