IP camera-based LSPIV system for on-line monitoring of river flow

To improve the real-time performance and robustness of the existing Large-Scale Particle Image Velocimetry (LSPIV) system, an IP camera-based LSPIV system has been developed for on-line monitoring of river flow. The camera was refitted to support NIR imaging to enhance the contrast between flow tracers and water surface. The image streams are automatically processed by a software consists of image acquisition, image enhancement, motion estimation, flow field processing and flow field calibration. Experiment results show that it takes at most 30 s for single measurement, which can meet the needs for river flow monitoring and hydrometry. The combination of NIR imaging and high-pass filtering is proved to be robust to quasi-static background noise such as shadow and glare, which improves the Peak Signal-to-noise Ratio (PSNR) in motion estimation. A nonlinear processing strategy is used to reconstruct time-averaged flow field. It utilizes the prior and redundant information in river flows, making it reliable for unseeded conditions with the low density and uneven distribution of nature tracers.