Design And Test Of Plant Water Potential Measuring Device Based On Machine Vision

Plant water potential is an important indicator of plant water physiology.At present,the commonly used pressure chamber method requires the operator to observe the seepage of the measured sample section to determine the end point of the water potential measurement.Due to the different criteria of the observer for the occurrence of seepage,it is easy to cause a large measurement error.In order to solve this problem,a method based on machine vision to judge the end point of plant water potential measurement is proposed,and the corresponding measuring device is developed.The specific research work is as follows.(1)Research on determination of end point of water potential measurement based on machine vision.The lighting system of machine vision is designed,and the choice of light source,light source shape,light source layout,light source driver and auxiliary device is made.The end point judgment method of water potential measurement based on machine vision is studied: By selecting different light source shapes,light source layouts,light intensity,light incidence Angle and other factors,24 experimental groups were formed.The USB camera was used to collect video from the cross-section of the sample under test,and the images before and after the cross-section of the sample under test were extracted from video.Matlab was used for image grayscale,binarization,background removal and grayscale average calculation,and the change of grayscale average was obtained.The experimental group with the largest variation was selected as the best lighting scheme.The test results show that the elected with ring light,diffuse lighting,light incident Angle 90 °,3000 lux light intensity detection effect is best;When the change of gray value first appears between 2 and 20,it can accurately indicate the end point of water potential measurement.(2)Structural design of the measuring device.The principle of measuring plant water potential by pressure chamber method is described.The machine vision recognition technology is introduced on the basis of water potential measurement by traditional pressure chamber method.The pressure chamber structure of the measuring device is designed as follows: The pressure chamber is composed of housing,top cover and rotary wrench,etc.304 stainless steel is selected as the material.The pressure chamber is composed of housing,top cover and rotary wrench,etc.304 stainless steel is selected as the material.The top cover and the shell are connected by a rodent quickopening structure;The shell is welded by the cylinder body,the bottom plate,the end flange,the pipe and the fixing lug.The top cover consists of a cover plate and a top cover flange connected by bolts.According to GB150-2011,the dimensions of the cylinder,the bottom plate,the holes and the reinforcement were calculated,and the strength was checked by the process equipment strength calculation software SW6-2011.(3)Control system design.According to the functional requirements of the system,the overall structure of the control system is designed,including the main control unit,data acquisition unit,human-computer interaction unit and execution unit.The main controller,USB camera,pressure sensor,touch screen,solenoid valve,proportional valve and other control system hardware selection;The control system software is developed.This software adopts the embedded Linux technology to write program code on the Windows system and cross-compile through the pc-linux virtual machine system to generate the executable file running on the arm-linux image processor.The control system software carries out man-machine interaction through the touch screen,realizing functions such as sample selection,real-time display of information,data collection,data storage and control decision-making.(4)Performance test of water potential measurement prototype based on machine vision.Taking tomato and sweet pepper as samples to be tested and comparing with traditional pressure chamber method,the measuring speed and precision of the prototype were tested to verify the performance of the prototype.The test results showed that for tomato samples,the time taken by the prototype and the traditional pressure chamber method to measure a sample to be tested was 5.2min and 6.2min respectively,and the time saved by the prototype was 16.1%.The average error was 0.01 MPa and the average relative error was 1.11%.For pepper samples,the time taken by the prototype and the traditional pressure chamber method to measure a sample to be tested was 5.2min and 6.3min,respectively.The time saved by the prototype was 17.5%.The average error was 0.01 mpa and the average relative error was 1.20%.The water potential measurement method based on machine vision improves the measuring speed and precision.