Application of improved YOLOv7-based sugarcane stem node recognition algorithm in complex environments.

Area of Science:

• Agricultural Robotics
• Computer Vision
• Machine Learning

Background:

• Sugarcane stem node detection is crucial for intelligent harvesting robots.
• Accuracy is compromised in complex field conditions due to shadows and background clutter.

Purpose of the Study:

• To improve the accuracy and robustness of sugarcane stem node detection in challenging environments.
• To enhance the performance of small intelligent sugarcane harvesting robots.

Main Methods:

• An improved YOLOv7 model incorporating SimAM attention mechanism for feature preservation.
• Integration of Deformable Convolutional Networks to replace traditional convolutional layers.
• Introduction of WIoU Loss function to address sample imbalance and improve convergence.

Main Results:

• The enhanced model achieved a mean Average Precision (mAP) of 94.53% and an F1 score of 92.41%.
• This represents a significant improvement over the standard YOLOv7 model (3.43% mAP, 2.21% F1).
• Outperformed state-of-the-art methods with a 0.43% increase in mAP.

Conclusions:

• The proposed model effectively enhances sugarcane stem node detection accuracy in complex environments.
• Provides a robust technical foundation for developing advanced intelligent agricultural machinery.
• Offers potential applications for crop detection in similar agricultural settings.