Edge-Based Standing-Water Detection via FSM-Guided Tiering and Multi-Model Consensus
For agricultural robotics and edge AI practitioners, this work offers a deployed solution for real-time water detection under resource constraints, though the improvements are incremental over existing methods.
The paper presents an edge architecture for standing-water detection in agricultural fields using Raspberry-Pi-class devices, combining camera input and environmental sensors with an FSM-guided tiering system and multi-model YOLO ensemble. The approach improves flood-detection performance over static local baselines, reduces energy consumption compared to naive offload policies, and maintains bounded tail latency in real agricultural settings.
Standing water in agricultural fields threatens vehicle mobility and crop health. This paper presents a deployed edge architecture for standing-water detection using Raspberry-Pi-class devices with optional Jetson acceleration. Camera input and environmental sensors (humidity, pressure, temperature) are combined in a finite-state machine (FSM) that acts as the architectural decision engine. The FSM-guided control plane selects between local and offloaded inference tiers, trading accuracy, latency, and energy under intermittent connectivity and motion-dependent compute budgets. A multi-model YOLO ensemble provides image scores, while diurnal-baseline sensor fusion adjusts caution using environmental anomalies. All decisions are logged per frame, enabling bit-identical hardware-in-the-loop replays. Across ten configurations and sensor variants on identical field sequences with frame-level ground truth, we show that the combination of adaptive tiering, multi-model consensus, and diurnal sensor fusion improves flood-detection performance over static local baselines, uses less energy than a naive always-heavy offload policy, and maintains bounded tail latency in a real agricultural setting.