基于三通道注意力网络的番茄叶部病害识别

对番茄病害进行识别,近年来一直是植物病害预防的研究热点。由于受到复杂背景干扰,番茄叶部病害识别准确率不高,针对这一问题,提出一种基于三通道注意力机制网络的番茄叶部病害识别方法。该网络基于ResNeXt50残差网络,将注意力模块嵌入至残差网络的ResNeXt模块中可以并行提取目标的通道特征和空间特征,获取有效的语义信息。训练阶段通过设计双损失函数和数据增强进一步提升分类准确度,并通过迁移学习网络预训练参数的方式提高网络训练效率。实验结果表明,使用双损失函数和数据增强后,基于三通道注意力网络的番茄病害识别算法在测试集上的平均识别准确率达98.4%,相比于传统机器学习方法和其他神经网络方法的准确率更高,检测速度满足实时性,Kappa系数为0.96,满足叶部病害识别的高精度要求。该方法能够有效地对10种番茄叶部病害进行识别,为植物病害识别提供了一种新的思路。

Tomato Leaf Disease Recognition Based on Three-channel Attention Network

The identification of tomato diseases has been a research hotspot in plant disease prevention in recent years. Due to the interference of complex background, the accuracy of tomato leaf disease recognition is not high. To solve this problem, this paper proposed a tomato leaf disease recognition method based on a three-channel attention mechanism. First, image enhancement and expansion on the collected data and the Plant Village public data set were performed to improve the generalization ability of the model; then, a three-channel attention mechanism module was designed, and embed the attention module into the ResNeXt module to correct the residual network structure Optimize; a double loss function was designed to solve the problem of network training over-fitting; finally, the efficiency of network training improved through migration learning network pre-training parameters. The experimental results show that the average recognition accuracy of the tomato disease recognition algorithm based on three-channel attention network reached 98.4% on the test set after using dual loss function and data augmentation, which is more accurate than traditional machine learning methods and other neural network methods, and the detection speed satisfies real-time with a kappa coefficient of 0.96, which meets the high accuracy requirements for leaf disease recognition. The method can effectively identify 10 tomato leaf diseases, which provides a new idea for plant disease identification.