基于深度学习航拍图像识别的农作物病虫害研究

Research on Aerial Image Recognition Technology of Crop Pests and Diseases Based on Deep Learning

海南省多山地,地貌多为坡地,采用无人机拍摄作物图像可以降低工劳动强度,提高生产效率以及数据准确性。深度学习作为航拍图像识别的重要技术手段,具有识别准确率高、速度快的优势。基于此,笔者概述了国内外深度学习领域农作物病虫害图像识别技术的研究进展,对深度学习技术在图像识别研究中存在的问题进行归纳总结,提出了一种卷积神经网络(CNN)与生成式对抗网络(GAN)的组合模型,经过实验对比,模型的图像识别正确率达95.55%,比卷积神经网络模型提高了5.05%。随着深度学习技术的不断发展,这种优化组合模型的研究将是未来发展的趋势。

Hainan Province has many mountains and landforms are mostly slopes.Using UAV to take crop images can reduce labor intensity,improve production efficiency and data accuracy.Deep learning,as an important technical means for aerial image recognition,has high recognition accuracy,Advantages of fast speed.This paper summarizes the research progress of crop disease and insect image recognition technology in the field of deep learning at home and abroad,and summarizes the problems existing in deep learning technology in image recognition research.A combined model of CNN and GAN is proposed.The experimental comparison shows that the model’s image recognition accuracy rate is 95.55%,which is an improvement of 5.05 percentage points over the CNN model.With the continuous development of deep learning technology,the study of this optimized combination model will be the future development trend.