Method for distinguishing male and female cotton bollworm adults based on computer vision technology

Abstract

The invention discloses a method for distinguishing female and male adults of cotton bollworm based on computer vision technology, after pre-processing the collected images of cotton bollworm of different sexes, the images with feet and antennae removed are obtained, and the color, shape and texture of the images are extracted respectively. The simulated annealing algorithm is used to optimize thedimensionality reduction process. The obtained feature data is trained and tested by the support vector machine classifier, and finally the automatic classification and recognition are realized. Theinvention has the advantages of simple operation, strong robustness and high identification accuracy, and has ideal time performance, and can remarkably improve the identification efficiency of insectsex classification.

Description

technical field [0001] The invention relates to the technical field of agricultural and forestry insect pest control, in particular to a method for discriminating male and female adults of cotton bollworm based on computer vision technology. Background technique [0002] Cotton bollworm is found in all climatic zones of the world. It hosts more than 200 species of more than 30 families. It likes to eat cotton, corn and other crops. It is one of the world's major pests. my country's total cotton output ranks first in the world, and the loss caused by cotton bollworms reaches 15%-20% every year, and some areas with serious losses are as high as more than 50%. There was only a major pest outbreak in the cotton area of ​​North China in 1992. According to estimates, its direct economic The loss was as high as 10 billion yuan. Although transgenic Bt cotton has gone through more than 20 years of research on insect resistance and large-scale planting applications, it has saved huge ...

AI Technical Summary

Problems solved by technology

[0003] At present, the methods used to distinguish the sex of insects at home and abroad mainly include artificial identification method, biological information discrimination method, near-infrared spectral analysis method, image recognition method, etc. The requirements are too high, human errors are prone to occur, and inaccurate or even wrong information is provided for control measures. The bioinformatics discrimination method can accurately classify pests, but it has high time and manpower requirements and cannot be widely used in various pests. The shortcomings of pest sex judgment, near-infrared spectroscopy analysis method has the disadvantages of difficult modeling and low recognition accuracy, and the above methods are not suitable for the control and research needs of cotton bollworm pests

Images