Research on Support Vector Machines Method Modeling for Rice Potassium Nutrition Diagnosis

摘要

A machine learning method was used to establish a diagnostic model of potassium nutrition for rice obtained from image processing techniques. In this study, super hybrid rice "Liangyoupeijiu" was used as experimental object to set up four kinds of rice cultivation experiments at different potassium fertilization levels, the image data of a total of 1920 groups of the 1st leaves and the 2nd leaves, the 3rd leaves, and their corresponding sheaths were obtained by scanning with a scanner. Nineteen rice characteristic indexes were obtained. Support vector machine was used to establish the diagnostic model of potassium nutrition in nineteen rice characteristic indexes, and to diagnose and identify the potassium nutrition in rice. The experimental results show that the identification method based on image processing and SVM is suitable for the diagnosis of potassium nutrition of the 3rd leaves in rice young panicle differentiation stage with an accuracy of 89%, which provides a reliable and universal method for studying the recognition of potassium nutrition in rice and can meet the needs of agronomic research.