LAUSR

Crop pests are a major agricultural problem worldwide because the severity and extent of their occurrence threaten crop yield. However, traditional pest image segmentation methods are limited, ineffective and time-consuming, which causes difficulty in their promotion and application. Deep learning methods have become the main methods to address the technical challenges related to pest recognition. We propose an improved deep convolution neural network to better recognize crop pests in a real agricultural environment. The proposed network includes parallel attention mechanism module and residual blocks, and it has significant advantages in terms of accuracy and real-time performance compared with other models. Extensive comparative experiment results show that the proposed model achieves up to 98.17% accuracy for crop pest images. Moreover, the proposed method also achieves a better performance on the other public dataset. This study has the potential to be applied in real-world applications and further motivate research on pest recognition.