轻量化模型实现无人机农田杂草实时检测，速度提升超28%。
TinyEcoWeedNet: Edge Efficient Real-Time Aerial Agricultural Weed Detection
- 采用结构化通道剪枝与量化训练压缩模型，适配边缘设备。
- 在棉田杂草数据集上达到85.9% mAP50，推理速度达184 FPS。
- 适合部署在资源受限的农业无人机等边缘设备上使用。
在农业中部署深度学习模型面临边缘设备资源有限的挑战，本文通过结构化通道剪枝、量化感知训练（QAT）及NVIDIA TensorRT加速，在Jetson Orin Nano上实现了EcoWeedNet的压缩。尽管复杂架构包含残差连接、注意力机制、拼接操作和CSP模块，模型规模仍减少68.5%，计算量降低3.2 GFLOPs，FP16下推理速度达184 FPS，比基线快28.7%。在CottonWeedDet12数据集上，39.5%剪枝率的模型优于YOLO11n和YOLO12n（仅20%剪枝），达到83.7%精度、77.5%召回率和85.9% mAP50，验证了其在精准农业中的高效性与有效性。
原文摘要 · Abstract (English)
Deploying deep learning models in agriculture is difficult because edge devices have limited resources, but this work presents a compressed version of EcoWeedNet using structured channel pruning, quantization-aware training (QAT), and acceleration with NVIDIA's TensorRT on the Jetson Orin Nano. Despite the challenges of pruning complex architectures with residual shortcuts, attention mechanisms, concatenations, and CSP blocks, the model size was reduced by up to 68.5% and computations by 3.2 GFLOPs, while inference speed reached 184 FPS at FP16, 28.7% faster than the baseline. On the CottonWeedDet12 dataset, the pruned EcoWeedNet with a 39.5% pruning ratio outperformed YOLO11n and YOLO12n (with only 20% pruning), achieving 83.7% precision, 77.5% recall, and 85.9% mAP50, proving it to be both efficient and effective for precision agriculture.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。