Cigar tobacco leaf harvesting maturity identification method and system based on integrated learning

A cigar leaf harvesting maturity identification system based on integrated learning, comprising:

• an acquisition module configured to acquire an image of a cigar tobacco leaf during a harvesting process in a field of cigar tobacco leaves;
• a preprocessing module configured to preprocess the image of the cigar tobacco leaf;
• a vectorization dimensionality reduction module configured to carry out vectorization dimensionality reduction on the preprocessed image of the cigar tobacco leaf and extract RGB and HSV eigenvalues to obtain a feature set;
• a feature dimensionality reduction module configured to carry out a nondimensionalizing operation on data in the feature set and carry out feature dimensionality reduction on the data in the feature set by using a Wrapper algorithm to obtain an initial data set of the image of the cigar tobacco leaf to be detected;
• and
• a model identification module configured to input the initial data set into a trained random forest model and output a maturity identification result of the image of the cigar tobacco leaf;
• wherein the feature dimensionality reduction module is configured to
• extract color eigenvalues of tens of thousands of pixels in the image of the cigar tobacco leaf to obtain a matrix having data as [image length, image width, color eigenvalue channel number], and vectorize the data into a matrix [image length*image width, color eigenvalue channel number] to obtain image data before clustering;
• import the image data before clustering into a K-Means algorithm, select the number of clusters to be R, and carry out clustering to obtain R clusters of color eigenvalues in the image data;
• and
• replace color eigenvalues of all pixels under the current cluster with color eigenvalues of pixels of R clusters at centroids.