基于图像智能识别的引黄灌区干渠视频测流方法OA

To evaluate the measurement accuracy and adaptability of video-based discharge gauging,this study targets the main canals of the Ningxia Yellow River Diversion Irrigation District.Using video data from two representative cross-sections,it proposed a real-time dynamic discharge measurement method for irrigation canals based on intelligent video-image recognition and validated it against synchronous discharge observations from a vertical acoustic Doppler current profiler(V-ADCP).The results indicate that in wide and shallow cross-sections,the video-based method achieves higher monitoring and recognition accuracy,with relative deviations ranging from-21.5%to 15.2%.Moreover,when the water surface exhibits abundant texture and floating tracers,measurement accuracy improves and adaptability is further enhanced.These findings provide a technical reference for video-based discharge gauging in the main canals of the Ningxia Yellow River Diversion Irrigation District.