Research On Surface Flow Measurement Of River Model

The main drawbacks of the conventional surface flow measurement of river model methods include three aspects,including complex procedures in setting the video capture devices,low measurement accuracy,and high cost of the tracers.To handle these problems,we developed a fast,easy and accurate particle image velocimetry(PIV)method for measuring the surface flow.This method has no constraint on the pose of video capture devices,and has the capability to measure large scale river models.In addition,there is no restraints on the types of the tracers and condition of illumination.In this method,firstly,the distortions of field-of-view are corrected and aerial views are reconstructed after capturing videos.Secondly,a subpixel feature point detection algorithm is used to identify the tracers.Thirdly,the approximate flow velocity vectors are obtained by statistics accumulation during a certain time period,through a pyramidal implementation of the classical Lucas-Kanade algorithm.Finally,the fluid vectors are visualized and analyzed.Our method has been evaluated by Chongqing academy of metrology and quality inspection.The results show that the relative error is less than 10% when the velocity is between 0.01m/s and 0.05m/s,and 5% when the velocity is between 0.05 m/s and 1.5m/s.And the angle deviation is less than 0.5?.Moreover,the method is fast,easy and accurate.The main contribution of the research is explained as the following three aspects.(1)Research and analysis.By investigating and analyzing the restoration methods for the surface flow measurement of river model,it is found that the measurement method based on PIV overcomes the limitations of the single-point contact measuring equipment.Therefore,this paper focuses on the research of PIV,and found that several inherent drawbacks in such methods as follows.(1)The flexibility to deal with the effect of external light is insufficient as the video capture devices must be mounted perpendicularly to the surface of flow fields.(2)The time cost of the multi-channel data analysis is huge but still with low measurement accuracy.(3)The customized tracers in most FSV measurement systems cost too much.(2)Research on the solution.A free perspective method is designed for video capture devices so that the system is easier to operate,and has better robustness for complex terrain.And the flow velocity vectors are obtained by a pyramidal implementation of the classical Lucas-Kanade method,instead of the traditional correlation method.In addition,in order to reduce cost,increase convenience and applications of the system,this paper breaks down the limitation of existing measurement methods for tracers,and improves algorithm accuracy to reduce the dependency of the measurement method on tracers.We take the scraps of papers which crushed by a conventional shredder as tracers,and the natural light as the illumination in our experiments.It is almost zero cost and easy to operate.(3)System development and performance evaluation.A large-scale surface flow field measurement system was developed for this paper.This system passed the identification with Chongqing academy of metrology and quality inspection successfully.And the results show that this system has several advantages,such as high measurement accuracy,simple and convenient operation,powerful capability in large scale river models measurement and real-time measurement.