Dynamic Deoxynivalenol Prediction for Wheat Harvest
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for predicting deoxynivalenol content in wheat at harvest are inaccurate due to reliance on static predictions and large differences in climatic data between years, leading to errors and excessive computational overhead when handling multi-dimensional large-batch data.
Innovation Solution
A dynamic prediction method using historical data from wheat samples, including geographic and climatic factors, to establish prediction sub-models based on particle swarm optimization, allowing for real-time prediction of deoxynivalenol content by normalizing climatic factors and applying them to specific growth stages, thereby reducing errors and improving computational efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If static prediction method is used at a single time point, then prediction process is simple, but prediction accuracy is low
Solution Approach 1:
The patent transforms the static prediction method into a dynamic multi-time point prediction system. Instead of performing prediction at a single fixed time point, the system establishes multiple prediction sub-models corresponding to different growth stage time periods (jointing stage, heading stage, flowering stage, grain filling stage, maturity stage) and performs predictions at multiple time points throughout the wheat growth cycle, thereby improving prediction accuracy while managing complexity through structured segmentation.
Solution Approach 2:
The patent segments the wheat growth period into five distinct growth stage time periods and establishes separate prediction sub-models for each stage. This segmentation allows the system to capture temporal variations in deoxynivalenol accumulation at different growth stages, improving overall prediction accuracy while organizing the complex prediction task into manageable modular components.
2Measurement precision
If multi-dimensional large-batch data is processed using traditional algorithms, then comprehensive analysis is achieved, but computational overhead time is excessive
Solution Approach 1:
The patent extracts and selects only the most relevant climatic factors from the multi-dimensional data using a particle swarm optimization algorithm. Instead of processing all available climatic data, the system identifies and extracts key climatic factors that have the most significant impact on deoxynivalenol accumulation, thereby maintaining prediction accuracy while significantly reducing computational overhead time.
Solution Approach 2:
The patent applies normalization processing to climatic factor data to transform it into a standardized format suitable for prediction model input. This parameter transformation optimizes the data for computational efficiency while preserving the essential information needed for accurate prediction, reducing the computational burden of processing multi-dimensional large-batch data.
3Adaptability or versatility
If prediction is based on previous years' growth period data, then historical trends are utilized, but prediction error increases when climatic data differs significantly
Solution Approach 1:
The patent incorporates real-time climatic data from the current year into the prediction model through a feedback mechanism. The system continuously updates the prediction by comparing actual current year climatic conditions with historical data and adjusting the prediction accordingly, allowing the model to adapt to significant climatic variations while maintaining prediction accuracy.
Solution Approach 2:
The patent creates a universal prediction framework that can handle both historical trend analysis and current year specific conditions. The prediction model is designed to work with diverse climatic scenarios by integrating historical growth period data with real-time climatic data, making it versatile enough to adapt to different years' conditions while maintaining accuracy.
Data Source
AI summary
The present application provides a method and system for dynamically predicting a deoxynivalenol content of wheat at harvest, including: on the basis of historical data, screening out by particle swarm optimization algorithm combined factors suitable for establishing a prediction model, and establishing the prediction model by using the combined factors; on the basis of data of a current year, predicting a second flowering date and a second harvest date of wheat in the current year by an agricultural model; then obtaining a weather forecast on the basis of the second flowering date and the second harvest date, and combining the weather forecast and geographic data into correlated factors; and finally predicting the deoxynivalenol content of wheat at harvest by means of the prediction model and the correlated factors. Compared with the prior art, statistical items in the prediction model are more comprehensive, and growth period data of the current year can be dynamically predicted on the basis of growth period indexes model, thus continuously adjusting and establishing the prediction model. In addition, an overhead time for screening multi-dimensional large-batch data by the particle swarm optimization algorithm has more advantages, and the prediction model established by a multiple linear regression algorithm has higher precision.