Research On Measurement Method Of Plant Water Use Efficiency Based On Near Infrared And Visible Images

As a country with a large population,with the decrease of cultivated land and the acceleration of industrial construction and urbanization,the most important thing is to ensure the yield and quality of crops.The measurement of plant water use efficiency has important reference significance for the improvement and breeding of crop varieties with low water consumption,strong drought resistance and high yield.The traditional methods for plant water use efficiency detection mainly include manual measurement and sensor contact measurement,but they have some disadvantages,such as strong subjectivity,low repeatability,damage to plants and so on.The crop monitoring based on image processing technology not only has the advantages of non-destructive and tracking observation,but also can provide the basis for the long-term comparative measurement of water use efficiency.Aiming at the research purpose of non-destructive measurement and long-term monitoring of plant water use efficiency,this paper obtains a variety of plant information based on image processing technology,and uses principal component analysis and multiple linear regression to explore a non-destructive measurement method of plant water use efficiency.Based on the comparative analysis of the existing measurement methods of plant water use efficiency,a fast and non-destructive measurement scheme of plant water use efficiency was determined by combining near-infrared and visible images at the individual level.Firstly,the definition and measurement method of plant water use efficiency(WUE)at individual level were introduced as the basis for subsequent measurement.According to the principle of normalized vegetation index(NDVI)to measure leaf water content,filter and camera were selected,and the relationship between NDVI and water content was obtained by image processing.The color and texture information of plants is obtained from visible light images.According to the measurement accuracy and plant growth characteristics,the binocular stereo vision acquisition system is built,and the camera is calibrated.The ad census algorithm is used for stereo matching to obtain parallax information.The three-dimensional reconstruction of spatial coordinates is used to restore the length,width,area,perimeter and other information of plant leaves,which is used to evaluate the growth status of plants.The length,width,area,perimeter and other information of plant leaves are recovered by using the three-dimensional reconstruction of spatial coordinates to evaluate the growth status of plants.At the same time,in order to achieve the rapid acquisition and processing of plant growth environment,the data collected by the sensor is uploaded to the host computer for display and storage.Based on the collection of all parameters and the correlation analysis with water use efficiency,the principal component analysis method was used to reduce the dimension of all parameters,and then the multiple linear regression method was used to establish the measurement model of plant water use efficiency.In order to verify the feasibility of the measurement scheme and evaluate the measurement accuracy of the method,dandelions with a certain economic benefit are the experimental subject,and the gradient water fertilizer is cultivated to dandelion,so that the growth state of the plant is significantly distinct,Effects of different water fertilizer management on plant yield and chlorophyll content.are discussed.The nondestructive measurement of the blade moisture content was achieved by the normalized vegetation index and the measurement accuracy of the measurement method was analyzed.The correlation between the shade of the blade,the length,the width,the area,the perimeter,and the color information and the water use efficiency of the plants were analyzed.After the main component analysis of the reserved parameters,multivariate linear regression predicts plant water utilization.The method of early prediction of plant water use efficiency according to the growth trend of plant geometric parameters was explored,which played a guiding role for the regulation of subsequent planting measures.This system has laid a good foundation for the subsequent construction of plant water use efficiency prediction model.The multiple key information acquisitions have realized the function of fast,nondestructive and digital storage,which provides a certain reference value for other types of plant water use efficiency measurement.