A nondestructive detection method, device and equipment for sugar content of fruits on a tree and a storage medium

CN122113035APending Publication Date: 2026-05-29AGRI INFORMATION INST OF CHINESE ACAD OF AGRI SCI
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
AGRI INFORMATION INST OF CHINESE ACAD OF AGRI SCI
Filing Date
2024-11-25
Publication Date
2026-05-29

AI Technical Summary

Technical Problem

Existing technologies suffer from low model detection accuracy due to the large number of bands and data volume in hyperspectral images. Hyperspectral cameras are expensive and have complex structures, making them difficult to use for detecting the sugar content of fruit on trees in orchard environments.

Method used

By reconstructing the spectrum from RGB images, processing and analyzing the reconstructed spectral data using a deep learning model based on the Transformer architecture, and combining the spectral reconstruction model with a fruit sugar content detection model, end-to-end non-destructive testing is achieved.

Benefits of technology

This reduces detection costs, improves the accuracy and stability of the model, and enables low-cost, high-precision detection of sugar content in tree fruits.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122113035A_ABST
    Figure CN122113035A_ABST
Patent Text Reader

Abstract

The application provides a nondestructive detection method, device and equipment for sugar content of fruits on a tree and a storage medium, and the method comprises the following steps: a data acquisition and preprocessing step, in which a hyperspectral image of fruits in a first area on a tree is acquired by using a first camera and is preprocessed, hyperspectral information of each fruit is extracted, a true value of sugar content of each fruit is measured, and a sugar content dataset is formed; a dataset construction step, in which hyperspectral information is subjected to integral operation by using a response function of a second camera, a first RGB image is rendered, an RGB-HSI dataset is formed, and the dataset is divided into a training set and a test set; a spectral reconstruction model training step, in which a spectral reconstruction model is trained by using the training set, optimal parameters are determined by using the test set, a spectrum is reconstructed, and a first reconstructed HSI sugar content dataset is formed; and a construction and training step of a fruit sugar content detection model, in which a fruit sugar content detection model based on a Transform architecture is designed, the fruit sugar content detection model is trained by using the first reconstructed HSI sugar content dataset, and a first fruit sugar content value is output.
Need to check novelty before this filing date? Find Prior Art