基于YOLO v5的病虫害检测与优化

《A fast and lightweight detection algorithm for passion fruit pests based on improved YOLOv5》

• a new point-line distance loss function is proposed to reduce redundant computations and shorten detection time
• the attention module is added to the network for adaptive attention, which can focus on the target object in the channel and space dimensions to improve the detection and identification rates.
• the mixup online data augmentation algorithm is added to expand the online training set, which increases the model robustness and prevents over-fitting.
  
  

《Real‑time and effective detection of agricultural pest using an improved YOLOv5 network》

• a lightweight feature extraction network GhostNet is adopted as the backbone and an efficient channel attention(ECA) mechanism is introduced to enhance feature extraction
• introduce BiFPN and add high-resolution feature map C2 and horizontal residual connections to it to enhance the expression of small pest features.
• We propose feature fusion with the attentional multiple receptive fields (FFARF) module, which dynamically assigns weights to each receptive field to highlight their unequal contribution to the global information after obtaining multiple receptive fields.