用AI生成图像增强数据，提升农田杂草检测精度与效率
Exploring Model Quantization in GenAI-based Image Inpainting and Detection of Arable Plants
- 用Stable Diffusion逐步生成新图像，数据量最多增200%
- 在YOLO11和RT-DETR上实现更高mAP50，且支持低精度推理
- 适配Jetson Orin Nano，适合边缘设备部署的智能农业系统
基于深度学习的杂草控制系统的训练数据多样性有限且计算资源受限，影响实际表现。为此，我们提出一种框架，利用Stable Diffusion进行图像修复以逐步增强训练数据，每轮增加10%，最高可额外提升200%的数据量与多样性。该方法在两个先进目标检测模型YOLO11(l)和RT-DETR(l)上评估，采用mAP50指标衡量检测性能。同时探索生成模型与检测模型的量化策略（FP16与INT8），平衡推理速度与精度。在Jetson Orin Nano上的部署验证了该框架在资源受限环境下的可行性，显著提升了智能杂草管理系统的检测准确率与计算效率。
原文摘要 · Abstract (English)
Deep learning-based weed control systems often suffer from limited training data diversity and constrained on-board computation, impacting their real-world performance. To overcome these challenges, we propose a framework that leverages Stable Diffusion-based inpainting to augment training data progressively in 10% increments -- up to an additional 200%, thus enhancing both the volume and diversity of samples. Our approach is evaluated on two state-of-the-art object detection models, YOLO11(l) and RT-DETR(l), using the mAP50 metric to assess detection performance. We explore quantization strategies (FP16 and INT8) for both the generative inpainting and detection models to strike a balance between inference speed and accuracy. Deployment of the downstream models on the Jetson Orin Nano demonstrates the practical viability of our framework in resource-constrained environments, ultimately improving detection accuracy and computational efficiency in intelligent weed management systems.
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。