Visible-Light Crop Yield Estimation With Weighted Grade Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional crop yield estimation methods, including field sampling surveys and UAV-based multispectral imaging, are labor-intensive, time-consuming, and yield predictions have high errors, necessitating improved accuracy and efficiency.
Innovation Solution
A method utilizing a deep convolutional neural network to classify yield grades from visible light images, applying a yield grade classification model to obtain confidence scores, and calculating estimated yields through weighted decision-making, reducing the complexity of image processing and enhancing prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If field sampling survey method is used for crop yield estimation, then the method is simple to implement, but the error is large and it is time-consuming and labor-intensive
Solution Approach 1:
The patent replaces the mechanical field sampling survey method with a deep learning-based image processing system. A convolutional neural network automatically extracts features from crop images and predicts yield, eliminating manual sampling and calculation while significantly improving accuracy to within 5% error margin.
Solution Approach 2:
The patent introduces an intermediate processing layer between image acquisition and yield prediction. The deep learning model acts as an intermediary that automatically extracts meaningful features from images and transforms them into accurate yield estimates, avoiding both manual sampling and the complexity of traditional multispectral analysis.
2Measurement precision
If UAV multispectral image processing method is used for crop yield estimation, then the prediction accuracy is improved, but the processing complexity is increased
Solution Approach 1:
The patent extracts only the essential visible light spectrum information needed for yield prediction, discarding the complexity of multispectral and hyperspectral analysis. The deep learning model focuses on extracting relevant features from standard visible light images, achieving high accuracy without processing multiple spectral bands.
Solution Approach 2:
The patent uses standard visible light cameras instead of expensive multispectral or hyperspectral sensors. The approach relies on readily available imaging technology combined with powerful algorithms, making the system more accessible and easier to implement while maintaining high prediction accuracy.
3Ease of operation
If conventional crop yield estimation methods are used, then the implementation is straightforward, but the time consumption and labor intensity are high
Solution Approach 1:
The patent implements a self-service system where the deep learning model automatically performs all yield estimation tasks without human intervention. The system autonomously processes images, extracts features, and generates predictions, eliminating the need for manual sampling, data entry, and calculation while dramatically improving processing efficiency.
Solution Approach 2:
The patent enables continuous automated processing of crop images for yield estimation. Unlike discrete manual sampling methods, the system can continuously analyze images and generate predictions in real-time, significantly increasing productivity and reducing the time required for yield assessment.
Data Source
AI summary
The present disclosure provides a crop yield estimation method and system based on level identification and weighted decision fusion. The method includes: inputting an obtained visible light image of a target crop field at maturity into a yield level classification model, to obtain yield level output values corresponding to all yield levels of the visible light image of the target crop field at maturity, where the yield levels are determined according to a yield level rule; calculating a confidence score corresponding to each of the yield level output values, and sorting the confidence scores in descending order; selecting the first m confidence scores, and performing normalization on the first m confidence scores to obtain m yield level weights; and respectively multiplying the m yield level weights and corresponding yield levels, and adding up all products, to obtain an estimated yield of the target crop field at maturity.