基于CNN的番茄叶片病虫害识别技术

Identification Technology of Tomato Leaf Pests and Diseases Based on CNN

摘要: 近年来，卷积神经网络(CNN)在植物病害检测中得到了迅速的发展和广泛的应用。番茄叶病是植物病害中的一种重要病害，所以设计一种能准确识别番茄叶病的模型是很有必要的。DCNet模型主要利用空洞卷积技术来训练网络模型，并使用批量归一化技术来加速模型收敛，采用随机失活技术避免过拟合问题。同时也利用批量归一化技术和随机失活技术减少了模型的训练次数，提高了植物叶片病害的分类效率。实验表明，与不同的CNN模型相比，在解决番茄病害分类问题上，该模型无论是参数量还是分类精度都达到了最好的效果。

Abstract: In recent years, convolutional neural networks (CNN) have been rapidly developed and widely used in plant disease detection. Tomato leaf disease is an important plant disease, so it is necessary to design a model that can accurately identify tomato leaf disease. The DCNet model mainly uses di-lated convolution technology to train the network model, and uses batch normalization technology to accelerate model convergence, and uses dropout technology to avoid overfitting. At the same time, batch normalization and dropout technology were used to reduce the training times of the model and improve the classification efficiency of plant leaf diseases. The experiment showed that, compared with different CNN models, the model achieved the best effect in solving the problem of tomato disease classification, no matter the number of parameters or classification accuracy.