Multi-source data efficient processing method and system for agricultural greenhouse

By receiving, classifying, preprocessing, extracting features and fusing multi-source data in agricultural greenhouses, combined with adaptive threshold filtering and multi-scale feature fusion, the problem of integrating multi-source heterogeneous data is solved, deep mining and precise management of data are achieved, and the efficiency of greenhouse management and crop yields are improved.

CN120653967APending Publication Date: 2025-09-16CHUNHE FUTURE FARM (GUANGDONG) CO LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202510808538.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-06-17
Publication Date
2025-09-16

AI Technical Summary

Technical Problem

Existing agricultural greenhouse data processing technologies are unable to effectively integrate and deeply mine heterogeneous data from multiple sources, resulting in an inability to provide sufficient support for precise management.

Method used

It uses data reception and classification, preprocessing, feature extraction, fusion and in-depth analysis methods, combined with adaptive threshold filtering, multi-scale feature fusion and machine learning algorithms to process multi-source data from sensors, cameras, drones and WMS systems.

Benefits of technology

It achieves unified management and in-depth mining of multi-source data, improves the availability and value of data, provides accurate decision-making support for the intelligent management of agricultural greenhouses, timely detects crop growth problems and reduces production costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120653967A_ABST
    Figure CN120653967A_ABST
Patent Text Reader

Abstract

The invention relates to a multi-source data processing method and system for an agricultural greenhouse. The method comprises the following steps: receiving multi-source data from a sensor, a camera, an unmanned aerial vehicle, an inspection robot and a WMS system, dividing the multi-source data into numeric data, image video data and text data, and then preprocessing the data, for example, removing abnormal values from the numeric data and normalizing the data; performing feature extraction on the preprocessed data, fusing different types of data into a comprehensive feature vector by adopting a multi-scale feature fusion algorithm, and finally analyzing and generating decision suggestions by utilizing a machine learning algorithm. The system comprises corresponding function modules for executing the operation. The agricultural greenhouse multi-source heterogeneous data processing method can effectively process agricultural greenhouse multi-source heterogeneous data, realizes deep mining and fusion of the data, provides accurate decision support for greenhouse intelligent management, improves the crop yield and quality, and reduces the production cost.
Need to check novelty before this filing date? Find Prior Art

Claims

1. A multi-source data processing method for agricultural greenhouses, characterized in that: include: Receive multi-source data from sensors, cameras, drones, inspection robots, and WMS systems; Multi-source data is classified into numerical, image and video, and text data and then preprocessed. Numerical data is filtered by adaptive threshold to remove outliers and normalize, image and video data is format converted, compressed, and enhanced, and text data is segmented and stop words are removed. Perform feature extraction on preprocessed data. Statistical analysis is used to extract features from numerical data, deep learning models are used to extract features from image and video data, and TF-IDF algorithm is used to extract features from text data. A multi-scale feature fusion algorithm is used to fuse different types of data into a comprehensive feature vector; Use machine learning algorithms to analyze comprehensive feature vectors and generate decision recommendations.

2. The multi-source data processing method for an agricultural greenhouse according to claim 1, characterized in that: The adaptive threshold filtering algorithm dynamically adjusts the threshold according to the historical distribution and real-time fluctuation of the data to eliminate outliers.

3. The multi-source data processing method for an agricultural greenhouse according to claim 1, characterized in that: The numerical data is normalized to the interval [0, 1] using minimum-maximum normalization.

4. The multi-source data processing method for an agricultural greenhouse according to claim 1, characterized in that: The image and video data formats are converted into JPEG or MP4 formats and compressed using H.264 coding.

5. The multi-source data processing method for an agricultural greenhouse according to claim 1, characterized in that: The image and video data enhancement processing includes histogram equalization and sharpening.

6. The multi-source data processing method for an agricultural greenhouse according to claim 1, characterized in that: The deep learning model is a convolutional neural network (CNN).

7. The multi-source data processing method for an agricultural greenhouse according to claim 1, characterized in that: The multi-scale feature fusion algorithm assigns weights according to the importance and relevance of different types of data to perform weighted fusion.

8. The multi-source data processing method for an agricultural greenhouse according to claim 1, characterized in that: The machine learning algorithm includes a support vector machine (SVM) or a decision tree.

9. A multi-source data processing system for agricultural greenhouses, characterized in that: It includes a data receiving module, a data preprocessing module, a feature extraction module, a data fusion module and a data analysis and decision module, and each module performs the operation of the corresponding step in claim 1 respectively.

10. The multi-source data processing system for agricultural greenhouses according to claim 9, characterized in that: The adaptive threshold filtering algorithm in the data preprocessing module adjusts the threshold to remove outliers based on the historical distribution and real-time fluctuations of the data.

Citation Information

Cited By

  • Vision, language and behavior linkage grassland robot inspection and driving integrated method

    CN121209519A