Research On The Flow Measurement Method Of Irrigation Channels Based On Machine Vision

Water measurement in irrigation areas is a powerful measure to save agricultural water,improve irrigation quality and irrigation efficiency,and is an important means to achieve accurate water diversion,transmission and distribution.The non-contact flow measurement method is simple,fast and has the feature of not destroying the flow field,especially the development of image processing technology and target tracking algorithm has brought new opportunities for machine vision flow measurement.In this paper,a machine vision-based channel flow measurement system is established to obtain the surface flow field distribution using machine vision-based flow measurement method,and obtain the flow rate through the function relationship between the surface basin flow rate and the cross-sectional flow rate in each small part,which realizes the channel water flow monitoring based on machine vision technology.The main research work and conclusions of this paper are as follows:(1)The study of machine vision-based water flow measurement technology in the channel system.Firstly,based on the characteristics of monocular machine vision technology and target detection technology,the optical flow velocity of the water flow surface ripple in each frame of the video is calculated using the pyramid L-K optical flow method,and the velocity conversion model is constructed to convert the optical flow velocity to the actual water flow surface velocity by the camera calibration method.A RANSAC algorithm based on multi-threshold Canny edge detection is proposed in the paper to accurately identify the water surface width,and the water level height is obtained through a pinhole imaging model.Finally,the intelligent measurement of channel water flow is realized by obtaining the standard channel cross-sectional flow based on the principle of flow velocity-area method.(2)Calibration of machine vision flow measurement system based on laboratory conditions.In this paper,we developed a channel water flow measurement system based on machine vision technology,and built a machine vision channel flow measurement equipment calibration device under laboratory conditions to realize the calibration process of the flow measurement system.Among them,the water flow measurement is performed when the water flows relatively smoothly in the channel,and the flow calibration can be performed using the standard mass method when the water flows out from the inlet.Through the analysis of the test results,the maximum relative error between the measured water level and the monitored water level can be obtained as 3.67%,and its accuracy is millimeter level;by comparing with the flow measured by the standard mass method,it is found that the relative error of the flow is within 10%,and both the water level and the flow meet the requirements of the flow measurement specification of the water conservancy industry,indicating that the method can be used for real-time measurement of water flow in irrigation areas.(3)Research on the correction of flow coefficient based on FLOW3 D numerical simulation.The current non-contact flow measurement method based on machine vision can measure the surface flow rate of the channel,but to get the actual cross-sectional flow needs to be multiplied by a certain correction factor of the water surface,so the value of the surface flow rate coefficient is essential for the acquisition of the channel water flow.In this paper,by using FLOW3 D software to simulate the real flow field environment,the surface velocity coefficient was corrected,and finally the relative error between the flow measurement result and the constant flow result of electromagnetic flow meter was within 5% by machine vision flow measurement system,which improved the accuracy of flow measurement and met the measurement requirement of water conservancy industry,and provided a new idea and method for the improvement of the accuracy of machine vision flow measurement;meanwhile,the machine vision flow measurement system was used to At the same time,the machine vision flow measurement system is used to measure the flow rate and flow rate at different water levels,and the relative error between the measured results and the actual flow rate is small,with the maximum error not exceeding 5%,which further indicates that the stability of the machine vision based flow measurement system is good and can be applied to the actual measurement of water flow in irrigation channels.This paper combines machine vision technology with motion target detection technology,and researches and discusses a set of machine vision-based channel surface flow measurement method,analyzes the relationship between the average flow velocity and surface flow velocity of a typical channel cross-section,and obtains a preliminary machine vision-based channel flow measurement method.