AgriJetsonBench：面向Jetson边缘平台的农业视觉模型的外部功率参考TensorRT基准测试
AgriJetsonBench: External-Power-Referenced TensorRT Benchmarking of Agricultural Vision Models on Jetson Edge Platforms
浏览论文内容
中文总结 AI 辅助
该研究提出AgriJetsonBench基准，测试Jetson AGX Orin 64GB和Orin Nano Super上7个农业视觉模型的部署性能，发现不同精度、功率模式下两款板卡表现有差异，部署需综合多维度因素。
中文摘要 AI 辅助
农业视觉模型的选择通常基于验证准确率和报告的帧率（FPS），但在嵌入式农业边缘GPU系统上的部署还取决于时序边界、数值精度、功率模式、板级输入能耗以及持续运行的有效性。我们提出AgriJetsonBench，这是一个可复现的部署基准，用于在NVIDIA Jetson AGX Orin 64GB和Jetson Orin Nano Super上进行作物/杂草检测与分割任务。使用现有的锁定式农业数据集作为固定部署工作负载，通过ONNX和TensorRT导出了7个模型家族。该基准结合了纯TensorRT引擎时序、外部板级输入功率日志、内部Jetson遥测、电池续航测试、运行有效性筛选以及存档的可复现性工件。主要能耗测试使用YOLO11s和SegFormer-B0，批量大小为1，输入尺寸为640×640。在匹配的15 W INT8操作下，Orin Nano Super在两个工作负载上均优于AGX Orin：YOLO11s达到145.82 FPS，每次推理能耗为0.0767 J，而AGX Orin分别为80.90 FPS和0.1814 J；SegFormer-B0达到47.18 FPS，每次推理能耗为0.2808 J，而AGX Orin分别为31.95 FPS和0.5013 J。在原生SegFormer-B0 FP16操作下，AGX Orin实现了比Nano Super更高的吞吐量和更低的p95延迟，分别为103.61 FPS和9.441 ms，而Nano Super为67.47 FPS和15.020 ms，每次推理能耗相近。对76691个内部-外部功率样本的时间对齐分析显示，二者具有高度时间关联性，但合并的内部减外部偏差为-1.988 W。结果表明，农业Jetson部署决策应基于工作负载、精度、功率模式和测量边界，而非仅基于板级型号。
英文摘要
Agricultural vision models are often selected from validation accuracy and reported frames per second, but deployment on embedded agricultural edge-GPU systems also depends on timing boundary, numeric precision, power mode, board-input energy, and sustained-run validity. We present AgriJetsonBench, a reproducible deployment benchmark for crop/weed detection and segmentation on NVIDIA Jetson AGX Orin 64GB and Jetson Orin Nano Super. Existing locked agricultural datasets were used as fixed deployment workloads, and seven model families were exported through ONNX and TensorRT. The benchmark combines pure TensorRT engine timing, external board-input power logging, internal Jetson telemetry, battery-continuation testing, run-validity screening, and archived reproducibility artifacts. Main energy tests used YOLO11s and SegFormer-B0 at batch size 1 and $640 \times 640$. Under matched 15 W INT8 operation, Orin Nano Super outperformed AGX Orin on both workloads: YOLO11s reached 145.82 FPS and 0.0767 J/inference versus 80.90 FPS and 0.1814 J/inference, and SegFormer-B0 reached 47.18 FPS and 0.2808 J/inference versus 31.95 FPS and 0.5013 J/inference. In native SegFormer-B0 FP16 operation, AGX Orin achieved higher throughput and lower p95 latency than Nano Super, 103.61 FPS and 9.441 ms versus 67.47 FPS and 15.020 ms, with similar energy per inference. Time-aligned analysis of 76,691 internal--external power samples showed high temporal association but a pooled internal-minus-external bias of $-1.988$ W. The results show that agricultural Jetson deployment decisions should be based on workload, precision, power mode, and measurement boundary rather than board class alone.
