Year: 2024
Type: article
Abstract: Insects are one of the most diverse biological groups on Earth, playing a crucial role in human agricultural production. However, there is a relative shortage of professionals capable of classifying insects. With the rapid advancement of computer vision technology, image classification techniques have been employed to classify insects. Traditional insect image classification requires manual extraction of image features, a process that is both time-consuming and labor-intensive, and the accuracy of the identification results is relatively low. To address these issues, this study adopts deep learning technology and uses Google's TensorFlow framework to build a convolutional neural network (CNN) model for insect classification. The article further analyzes the impact of different optimizers and learning rates on the model's classification performance. Experimental results show that using the Adam optimizer with a learning rate of 0.009 yields the highest recognition accuracy for the CNN model, reaching up to 92%.
DOI: 10.54097/hrh97f63
Author: Zhiqiang Zhang
Institution: Zhejiang DongFang Vocational and Technical College
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4405900848