Study On Rapid Detection Of Phenotypic Character Parameters Of Plants

With the rapid development of Digital Agriculture and agricultural iot technology,the research and development of software and hardware platforms such as the rapid non-destructive testing technology and sensing instrument of plant life information has become the hot spot of modern agriculture research.We can use a lot of expensive sophisticated equipment when we're doing agricultural science.But agricultural production should consider the return on investment and cannot afford the expensive equipment.Therefore,in the process of plant growth information detection,this paper focuses on low cost and fast.It can be seen that the study and development of low-cost rapid detection device for plant morphology and nutrient content can provide technical support for the refinement and large-scale management of crop planting and the improvement of crop yield and quality.The development of Digital Agriculture is of great significance to the advancement of the construction of Intelligent AgricultureFocusing on the key issues of rapid detection technology for plant phenotypic traits,this paper studies the rapid detection technology for plant phenotypic traits from two different perspectives.From the first perspective,the plant growth process as a clue,from the study of grain phenotype information to the acquisition of leaf phenotype information,from the plant leaf monomer to multiple leaves,and then to the measurement of canopy leaves;The second point of view,starting from the interior in vitro testing research,extended to outdoor living testing,indoor in vitro detection can be in a particular environment to achieve higher measurement accuracy,provide high precision detection scheme for follow-up studies,outdoor living detection is on the premise of meet the accuracy,rapid,accurate,convenient access to plant phenotypic traits parameter,improve the work efficiency of the detection of plant phenotypic information to achieve high throughput,more suitable for the vast number of agricultural science and technology workers and agricultural producers for actual testing workIn this paper,a lot of research work has been carried out on the phenotypic information acquisition of grains and leaves,and the methods and means of rapidly obtaining the phenotypic parameter information of grains and leaves have been studied.From the rapid and accurate counting of grains,the measurement of shape parameters,to the measurement of leaf shape,area and nutrient content.The study was carried out under the premise of reducing the use cost.A variety of rapid detection technologies for plant phenotypic traits were designed,and low-cost rapid detection devices for plant grain and leaf information were studied and developed.In terms of research methods,this paper proposes the research idea of image eigenvalue extraction using HSV color model,and applies this research idea to the research of plant phenotypic character parameter extraction.Many current studies convert RGB color image to gray image,and then according to gray image processing,such as research methods have lost a lot of useful color information,there is a lot of adverse effects,the result of the decision,especially in agriculture,application fields,plant leaves,flowers,fruits and so on,often have very bright color,the color information reflects many plant phenotypic information of research value.The research contents of this paper mainly include the following six aspects:(1)The rapid method for the determination of phenotypic parameters of seeds in laboratory was studied,proposed the CM-Watershed algorithm for image segmentation,the accuracy of CM-Watershed algorithm in the identification of line stick contour can reach more than 99%.(2)The rapid method for the determination of phenotypic parameters of leaves with indoor high precision was studied.In this paper,a Minlabel algorithm for image contour recognition is proposed,The high precision measuring instrument for plant leaf area designed by Minlabel algorithm has high measurement accuracy.The maximum standard deviation is 0.05 from experimental data and the relative error is 0.7%.(3)A rapid method for the determination of phenotypic parameters of leaves was studied:using HSV color model is used to identify the reference and background separation,and then identify the object to be tested.The area of the leaf is calculated by the ratio of the number of pixels between the blade and the reference.It is easy to carry,quick to measure,and can support multiple blades to measure at the same time.(4)A rapid method for detecting the phenotypic parameters of leaves with high outdoor flux was studied,carry out research from the perspective of ease of use,and use binocular vision technology to achieve the fast detection of non-contact and close distance.(5)Studied the rapid detection method of nitrogen content in plant leaves,and used image processing technology to analyze the color of the sampled leaves.HSV CAM leaf nitrogen content detection method is proposed,and the SPAD meter test results are compared,the experimental results show that HSV CAM test method of the nitrogen content of the measured values of N and highly correlated with SPAD meter measuring the nitrogen content of the values,R2=0.901.It has the characteristics of rapid detection and convenient use and can meet the needs of agricultural production.(6)In terms of the implementation of the rapid detection device for plant phenotypic character parameters.Two embedded system development modes are proposed.One is using the combination of MCU and Android smart camera,suitable for the external control operation and external information acquisition type of image processing.Another is to use embedded Linux+QT development mode,this mode is suitable for application types with high requirements for image acquisition and processing.These two development modes are all realized,forming two development templates,which play a good reference role for the development of other agricultural instruments in the later project.