Tree species classification method and system based on deep learning

Abstract

The invention discloses a tree species classification method and system based on deep learning. The method includes: collecting sample images for labeling; utilizing labeled sample images to constructa data set, and carrying out preprocessing; then utilizing a convolutional neural network to extract image features for a preprocessed image data set to acquire a feature vector set; utilizing the acquired feature vector set to train and verify a thirty-classifier; and finally, collecting leaf image data of a to-be-detected tree species, carrying out preprocessing operations, extracting a featurevector, and then utilizing the trained thirty-classifier to judge and classify the extracted leaf image feature vector to realize automatic tree species classification. According to the method, the problem that species leaf contour features cannot be completely reflected by existing tree species classification methods, and thus inaccurate results are caused is solved, and accuracy and reliabilityof tree species classification are guaranteed.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, and more specifically, to a tree species classification method and system based on deep learning. Background technique [0002] The identification and classification of tree species is of great significance for exploring the evolution law of plant ecosystems. After long-term development, the botanical research community has proposed many tree species classification methods. These classification methods mainly select some relatively stable appearance traits of trees, that is, appearance characteristics, collect characteristic data about these traits through observation and measurement, and then perform cluster analysis and principal component analysis on these trait characteristic data to achieve classification. The selection of tree traits is generally based on the parts of the tree, such as some characteristics of leaves, flowers, fruits, stems, and branches, and identify...

AI Technical Summary

Problems solved by technology

These methods are inefficient and difficult to guarantee data objectivity

[0003] At present, researchers have begun to use computer vision technology to classify leaves, specifically by calculating the shape features of leaf contours such as curvature, aspect ratio, rectangularity, eccentricity, etc., but this method has parameters for leaf contours. The processing can not reflect the original appearance of the contour, and there is a certain deviation in the result

Images