Research On Calculation Method Of Soybean Canopy Parameters Based On Depth Information

With the development of agricultural mechanization and modern farmland cultivation management technologies of various farmlands,machine vision technology based on artificial intelligence has been widely used in the agricultural field.The fast,non-destructive and high-throughput acquisition of crop plant type parameters in the natural environment is the key to breeding excellent varieties with high yield and strong resistance.The shortcomings of traditional phenotype measurement methods are destructive,time-consuming and laborious,a small measurement range,and non-continuous measurement.Therefore,in order to overcome the inefficiency of traditional measurement methods,this study carried out research on the calculation method of plant type parameters of soybean canopy based on depth information,aiming to make breakthrough progress in plant type parameter acquisition technology.This study took cold soybeans as the research object,and carried out soybean planting and canopy information collection in the Heilongjiang Bayi Agricultural University of Agricultural Innovation and Entrepreneurship Training.The calculation method of soybean canopy plant type parameters was studied by combining theoretical research and experimental analysis,so as to analyze the change rule of soybean canopy plant type parameters based on time series.At the same time,the calculation method of canopy plant type parameters was optimized by measuring the canopy plant type parameters of soybean reproductive growth period.It mainly includes four aspects of work:(1)A synchronous acquisition system of soybean canopy image based on Kinect 2.0 was designed to quickly,accurately and non-destructively acquire color and depth images of soybean canopy during the whole growth period.The three-dimensional structure characteristics of the canopy were analyzed,the design image registration parameters and fusion vectors were optimized,and the three-dimensional structure morphology of soybean canopy in different growth periods was accurately reconstructed.(2)The position relationship between the color camera and the depth camera is calibrated,and color distortion correction is performed on the color image to effectively avoid influence of exposure and distortion on color expression.The three-dimensional structure of soybean canopy with accurate color information was reconstructed.In the Lab color space,K-means clustering algorithm was used to segment the image,and then the effective soybean canopy area was extracted using color and distance information.(3)Based on the reconstruction of the three-dimensional structure of the canopy,a method of extracting soybean canopy plant type parameters was developed.The phenotypic parameters of various organs in soybean canopy space at different growth stages were analyzed with time.The calculation methods of plant type parameters such as plant height,crown width,and leaf area index in three-dimensional space of the canopy were established respectively.(4)The canopy plant type parameters of single-pot and group soybeans were actually measured during the growth process,and used to verify the validity of the calculation method of soybean canopy plant type parameters.The determination coefficient R~2 of soybean canopy height,canopy width,leaf area index are all around 0.9.The experimental results show:The study on the calculation method of soybean canopy plant type parameters based on depth information proposed in this study can quickly and accurately calculate canopy plant height,width,and leaf area index.It not only provides a theoretical basis for breeding excellent soybean germplasm resources,but also provides a practical and feasible technical means for the field high-throughput calculation of other crop phenotype parameters.