基于注意力机制的松材线虫病树识别应用研究

Application of Bursaphelenchus xylophilus Disease Identification Based on Attention Mechanism

松材线虫病是一种松树的传染病,对森林具有极强的破坏性,近几年,深度学习技术广泛应用于森林病虫害的监测领域。利用深度学习技术对病树进行了监测,通过使用基于注意力机制的目标检测算法对无人机遥感影像下的松材线虫病树进行了检测,以实现对林区松材线虫病树的高效、智能化的识别。实验展示了基于空间注意力机制、通道注意力机制和混合注意力机制的目标检测算法对松材线虫病树的识别结果,并分析总结了基于不同注意力机制的目标检测算法在松材线虫病树识别应用上的差异。实验结果表明:基于混合注意力机制的目标检测算法在松材线虫病树影像上的检测效果最好,使用基于注意力机制的目标检测算法在大老岭自然保护区的无人机影像中,识别出了561棵松材线虫病树。

Bursaphelenchus xylophilus disease is an infectious disease of Pinus trees,which is very destructive to forests.In recent years,deep learning technology has been widely used in the detection field of forest diseases and insect pests.In this paper,deep learning technology is used to monitor the diseased trees,and the target detection algorithm based on attention mechanism is used to detect the Bursaphelenchus xylophilus diseased trees in the remote sensing image of UAV and to realize the efficient and intelligent identification application of the Bursaphelenchus xylophilus diseased trees in the forest area.The experiment shows the identification results of Bursaphelenchus xylophilus diseased trees based on spatial attention,channel attention,and mixed attention mechanisms,and summarizes the differences in the application of Bursaphelenchus xylophilus tree identification based on different attention mechanisms.The experimental results show that the target detection algorithm based on the mixed attention mechanism has the best detection effect on the Bursaphelenchus xylophilus diseased tree image.1435 Bursaphelenchus xylophilus diseased trees were identified using the target detection algorithm based on the attention mechanism under the UAV image in the Dalaoling Nature Reserve.