AUTHOR=Pan Pan , Guo Wenlong , Zheng Xiaoming , Hu Lin , Zhou Guomin , Zhang Jianhua 

TITLE=Xoo-YOLO: a detection method for wild rice bacterial blight in the field from the perspective of unmanned aerial vehicles

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 14 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2023.1256545

DOI=10.3389/fpls.2023.1256545

ISSN=1664-462X

ABSTRACT=Wild rice, as a natural gene pool for rice germplasm innovation and variety improvement, holds immense value in rice breeding due to its disease-resistance genes. However, the conventional approach to large-scale disease resistance identification in wild rice heavily depends on timeconsuming and subjective manual methods. In recent years, a new trend has emerged, combining unmanned aerial vehicles (UAV) and deep learning to achieve intelligent disease resistance identification. To achieve target detection of bacterial blight in wild rice under natural field conditions, this study proposes a model named the Xoo-YOLO based on the YOLOv8 model for detecting bacterial blight in wild rice, which backbone network introduced into the Large Selective Kernel Network (LSKNet) to better achieve the detection of disease targets under the UAV view by dynamically adjusting its large spatial receptive field. Simultaneously, the neck network is enhanced by introducing the hybrid convolution module of the GSConv to reduce the amount of calculation and parameters of the model. To overcome the challenge posed by the elongated and rotational nature of the disease when detected from the UAV's perspective, a rotational angle (theta dimension) was introduced to the output of the head layer. This enhancement enables precise detection of bacterial blight in wild rice from any direction. The experimental results highlight the effectiveness of our proposed Xoo-YOLO model, boasting a remarkable mean average precision (mAP) of 94.95%. This outperforms other models, underscoring its superiority. Our model strikes a harmonious balance between precision and speed in disease detection. It stands as a technical cornerstone, facilitating the intelligent identification of disease resistance in wild rice on a large scale.