用热成像与边缘计算实现实时病叶识别，速度快且精度高。
An Edge Computing-Based Solution for Real-Time Leaf Disease Classification using Thermal Imaging
- 在树莓派上部署剪枝量化后的模型，实现轻量推理。
- 模型在边缘设备上比高端显卡快2.13倍，准确率仍达顶尖水平。
- 适合农业物联网和田间实时监测场景使用。
深度学习技术可提升农作物健康监测与管理，从而保障食品安全。本文探索边缘计算在热成像叶片疾病实时分类中的应用潜力。我们构建了一个用于植物病害分类的热图像数据集，并在资源受限设备如Raspberry Pi 4B上评估了InceptionV3、MobileNetV1、MobileNetV2和VGG-16等深度学习模型。通过剪枝与量化感知训练，这些模型在Edge TPU Max上对VGG16的推理速度最快提升1.48倍，在Intel NCS2上对MobileNetV1使用精度降低后速度最快提升2.13倍，相比高端GPU如RTX 3090，同时保持最先进的分类准确率。
原文摘要 · Abstract (English)
Deep learning (DL) technologies can transform agriculture by improving crop health monitoring and management, thus improving food safety. In this paper, we explore the potential of edge computing for real-time classification of leaf diseases using thermal imaging. We present a thermal image dataset for plant disease classification and evaluate deep learning models, including InceptionV3, MobileNetV1, MobileNetV2, and VGG-16, on resource-constrained devices like the Raspberry Pi 4B. Using pruning and quantization-aware training, these models achieve inference times up to 1.48x faster on Edge TPU Max for VGG16, and up to 2.13x faster with precision reduction on Intel NCS2 for MobileNetV1, compared to high-end GPUs like the RTX 3090, while maintaining state-of-the-art accuracy.
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。