复杂场景下害虫目标检测算法:YOLOv8-Extend  

Crop Pest Target Detection Algorithm in Complex Scenes:YOLOv8-Extend

作　　者：张荣华 白雪 樊江川[2,3] ZHANG Ronghua;BAI Xue;FAN Jiangchuan(Jinghang Chuangzhi(Beijing)Technology Co.Ltd.,Beijing 102404,China;National Engineering Research Center for Information Technology in Agriculture,Beijing 100097,China;Beijing Key Laboratory of Digital Plants,Beijing 100097,China)

机构地区：[1]京航创智(北京)科技有限公司,北京102404 [2]国家农业信息化工程技术研究中心,北京100097 [3]数字植物北京市重点实验室,北京100097

出　　处：《智慧农业（中英文）》2024年第2期49-61,共13页Smart Agriculture

基　　金：北京市科技新星计划(Z211100002121065,Z20220484202);“十四五”国家重点研发计划项目(2022YFD2002302-02)。

摘　　要：[目的/意义]实现复杂的自然环境下农作物害虫的识别检测,改变当前农业生产过程中依赖于专家人工感官识别判定的现状,提升害虫检测效率和准确率具有重要意义。针对农作物害虫目标检测具有目标小、与农作物拟态、检测准确率低、算法推理速度慢等问题,本研究提出一种基于改进YOLOv8的复杂场景下农作物害虫目标检测算法。[方法]首先通过引入GSConv提高模型的感受野,部分Conv更换为轻量化的幻影卷积(Ghost Convo⁃lution),采用HorBlock捕捉更长期的特征依赖关系,Concat更换为BiFPN(Bi-directional Feature Pyramid Network)更加丰富的特征融合,使用VoVGSCSP模块提升微小目标检测,同时引入CBAM(Convolutional Block Attention Module)注意力机制来强化田间虫害目标特征。然后使用Wise-IoU损失函数更多地关注普通质量样本,提高网络模型的泛化能力和整体性能。之后,对改进后的YOLOv8-Extend模型与YOLOv8原模型、YOLOv5、YOLOv8-GSCONV、YOLOv8-BiFPN、YOLOv8-CBAM进行对比,验证模型检测准确度和精度。最后将模型移植到边缘设备进行推理验证,在实际应用场景中验证模型的有效性。[结果和讨论]YOLOv8-Extend在对比实验中均取得良好的表现,其中与原模型对比实验中,精确率、召回率、mAP@0.5和mAP@0.5∶0.95评价指标分别提升2.6%、3.6%、2.4%和7.2%,表现突出,具有更好的检测效果。改进前后的模型分别运行在边缘计算设备JETSON ORIN NX 16 GB上并通过TensorRT加速相比,mAP@0.5提升4.6%,达到57.6 FPS,满足实时性检测要求。在复杂农业场景中YOLOv8-Extend模型具有更好的适应性,在实际采集数据中微小害虫与生长环境相似的害虫检测方面有明显优势,在困难数据检测方面准确率提高了11.9%。[结论]本研究提出的YOLOv8改进模型有效提高了检测精度和识别率同时保持了较高的运行效率,能够部署在边缘终端计算设备上实现农作物害虫[Objective]It is of great significance to improve the efficiency and accuracy of crop pest detection in complex natural environments,and to change the current reliance on expert manual identification in the agricultural production process.Targeting the problems of small target size,mimicry with crops,low detection accuracy,and slow algorithm reasoning speed in crop pest detection,a complex scene crop pest target detection algorithm named YOLOv8-Entend was proposed in this research.[Methods]Firstly,the GSConv was introduecd to enhance the model's receptive field,allowing for global feature aggregation.This mechanism enables feature aggregation at both node and global levels simultaneously,obtaining local features from neighboring nodes through neighbor sampling and aggregation operations,enhancing the model's receptive field and semantic understanding ability.Additionally,some Convs were replaced with lightweight Ghost Convolutions and HorBlock was utilized to capture longer-term feature dependencies.The recursive gate convolution employed gating mechanisms to remember and transmit previous information,capturing long-term correlations.Furthermore,Concat was replaced with BiFPN for richer feature fusion.The bidirectional fusion of depth features from top to bottom and from bottom to top enhances the transmission of feature information acrossed different network layers.Utilizing the VoVGSCSP module,feature maps of different scales were connected to create longer feature map vectors,increasing model diversity and enhancing small object detection.The convolutional block attention module(CBAM)attention mechanism was introduced to strengthen features of field pests and reduce background weights caused by complexity.Next,the Wise IoU dynamic non-monotonic focusing mechanism was implemented to evaluate the quality of anchor boxes using"outlier"instead of IoU.This mechanism also included a gradient gain allocation strategy,which reduced the competitiveness of high-quality anchor frames and minimizes harmful gradients from

关 键 词：YOLOv8 害虫检测 注意力机制 边缘计算 CBAM BiFPN VoVGSCSP GSConv 

分 类 号：S433[农业科学—农业昆虫与害虫防治] TP391.41[农业科学—植物保护]