Methods, apparatus, equipment, and media for tobacco leaf grading based on hyperspectral band selection
By combining hyperspectral imaging and predictive classification networks, the problems of low efficiency and inaccuracy in tobacco leaf grading in existing technologies have been solved, achieving efficient and accurate tobacco leaf grade classification.
CN116625955BActive Publication Date: 2026-06-30SHENZHEN UNIV
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHENZHEN UNIV
- Filing Date
- 2023-04-07
- Publication Date
- 2026-06-30
AI Technical Summary
Technical Problem
Existing technologies cannot provide an effective method for grading tobacco leaves, resulting in low grading efficiency and inaccuracy, especially after band selection, making accurate classification difficult.
Method used
By taking hyperspectral images of the tobacco leaves to be graded, selecting the target bands of the target number of bands, segmenting them, and then using a pre-trained prediction classification network to classify the tobacco leaf grades, the classification prediction values are statistically analyzed and the highest prediction value is set as the tobacco leaf grade.
Benefits of technology
This improved the efficiency and accuracy of tobacco leaf grading, ensuring accurate classification after band selection.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN116625955B_ABST
Abstract
This invention relates to the field of deep learning technology and provides a method, apparatus, device, and medium for grading tobacco leaves based on hyperspectral band selection. The method includes: taking a picture of the tobacco leaves to be graded to obtain a first hyperspectral image of the tobacco leaves; selecting bands from the first hyperspectral image according to a preset number of target bands and a pre-acquired set of representative bands to obtain target bands corresponding to the target number of target bands; segmenting a second hyperspectral image composed of all target bands to obtain several average spectral curves; classifying the tobacco leaves according to the average spectral curves using a pre-trained prediction classification network to obtain a classification prediction value corresponding to each average spectral curve; statistically analyzing the classification prediction values; and setting the classification prediction value with the most statistical occurrences as the grade of the tobacco leaves, thereby improving the efficiency and accuracy of tobacco leaf grading.
Need to check novelty before this filing date? Find Prior Art