Related Experiment Video
Updated: Jan 10, 2026
10:37
A Multimodal Imaging Framework to Advance Phenotyping of Living Label-free Breast Cancer Cells
Published on: August 22, 2025
1.0K
Non-Destructive Detection and Grading of Plum Quality Based on Multimodal Data
Xian Liu1, Weibin Tong2,3, Biao Di2,3
1Institute of Digital Agriculture, Fujian Academy of Agricultural Sciences, Fuzhou 350008, China.
Sensors (Basel, Switzerland)
|November 27, 2025
Summary
Deep learning fuses color and spectral data for accurate plum quality assessment. This multimodal approach achieves 100% classification accuracy, surpassing single methods for non-destructive plum grading.
Area of Science:
- Agricultural Science
- Computer Science
- Spectroscopy
Background:
- Plum quality assessment traditionally relies on external appearance, which can be unreliable.
- Objective and accurate methods are needed for non-destructive plum grading.
Purpose of the Study:
- To develop a multimodal deep learning technique for accurate plum quality assessment.
- To fuse color image and spectral data for enhanced plum grading.
Main Methods:
- Utilized Visual Geometry Group 16-layer network (VGG16) for color image feature extraction.
- Employed a 1D Convolutional Neural Network (1D-CNN) for near-infrared spectral data feature extraction.
- Integrated features from both modalities for classification using fully connected layers.
Main Results:
- The multimodal fusion approach achieved 100% plum quality classification accuracy.
- Significantly outperformed single-modal methods: color imaging (85.71%) and spectroscopy (83.33%).
- Demonstrated robustness and applicability for non-destructive plum quality detection.
Conclusions:
- Multimodal data fusion provides a superior method for plum quality assessment.
- This technique enables simultaneous detection of internal and external quality indicators.
- Offers advanced technical solutions for non-destructive plum grading.
