A Recognition Method of Water-stained Tobacco Leaf After Conditioning Based on Feature Extraction of Industrial Camera

Abstract

The invention discloses a method for identifying water-stained tobacco leaves after moistening based on feature extraction of industrial cameras. The method comprises the following steps: 1) collecting all original tobacco leaf images of online tobacco leaf samples, and manually classifying the tobacco leaf samples into normal tobacco leaves, water-stained tobacco leaves Tobacco leaves, near-normal tobacco leaves and near-water-soaked tobacco leaves, and obtain the information of all original tobacco leaf images; 2) Process the information of the original tobacco leaf images to obtain the feature vectors of all original tobacco leaf images; 3) Perform SNV on the feature vectors Preprocessing to eliminate abnormal numerical information; 4) Supervised two-dimensional projection processing on the preprocessed features to obtain two-dimensional projection vectors. Quickly solve the rapid identification and proportion calculation of water-stained tobacco leaves after one-run, lay a solid foundation for reducing the proportion of water-stained tobacco leaves through process control, and accumulate original identification data of water-stained tobacco leaves.

Description

technical field [0001] The invention relates to the field of image feature information extraction and identification, in particular to a method for identifying water-stained tobacco leaves after moistening based on industrial camera feature extraction. Background technique [0002] Affected by high temperature and high humidity in the moistening process, the tobacco leaves are prone to water-soaked tobacco and steamed sheets. It is difficult to adjust the water-soaked tobacco and steamed sheets to the appropriate moisture in the re-baking process, resulting in excessive moisture in the water-stained sheets. During the packing process of tobacco sheets, the temperature and moisture of water-stained sheets are relatively high, resulting in excessive moisture and temperature of local tobacco leaves, which are prone to mildew and carbonization during storage, resulting in losses. The identification of water-soaked tobacco leaves is generally based on the fact that the leaf surfa...

AI Technical Summary

Problems solved by technology

[0004] As for how to accurately and simply extract the characteristic information of water-stained tobacco leaves from industrial cameras to identify whether the tobacco leaves are water-stained tobacco leaves, combined with image stitching, and according to the area ratio of the group of feature information to identify water-stained tobacco leaves The actual proportion of tobacco leaves within a period of time is measured and calculated, and there is no record of such a technical solution in the prior art

Images