Edge Intelligence and Vision-Based Robotics for Sustainable Disease and Pest Management in Field Crops

Ensuring global food security relies heavily on effective phytoprotection, particularly for staple field crops like wheat and maize. Conventional practices of broadcasting chemical pesticides not only cause severe environmental pollution but also disrupt ecological balance, contradicting sustainable development goals. Although artificial intelligence and machine vision have advanced crop disease diagnosis and weed detection, deploying these models in real-world agricultural fields faces significant hurdles. Cloud-based computing suffers from network latency and high bandwidth bottlenecks in rural areas, while complex neural networks often exceed the computational capacity of onboard devices on agricultural robots. To achieve sustainable agriculture, there is an urgent need to shift toward "edge intelligence," deploying lightweight, robust AI models directly on spraying robots or UAVs for real-time, precision phytoprotection.

This Research Topic fully shifts its core focus toward Proof-of-Research. Submitted manuscripts are required to provide 100% complete in-field evaluation results as essential supporting evidence. At the current stage of agricultural intelligent robotics research, simulation-only or pure-modelling analysis is insufficient to verify practical performance. Studies relying solely on simulation or modelling will not be accepted, unless they are fully validated by real-world field-based experiments.

This Research Topic aims to achieve the highly efficient engineering deployment of advanced deep learning algorithms on resource-constrained intelligent phytoprotection hardware. We seek cutting-edge research that synergizes machine vision, edge computing, and agricultural robotics to enable real-time identification and targeted treatment of crop diseases, pests, and weeds. By focusing on field crops, our goal is to provide practical, proof-of-concept solutions validated under real-world farm conditions, such as vision-based navigation for spraying robots, lightweight leaf disease diagnosis systems, and UAV-assisted ecological control. We aim to establish benchmarks for balancing model accuracy with the computational efficiency required in complex field environments. Ultimately, this collection will disseminate impactful discoveries that significantly reduce chemical pesticide usage, directly advancing United Nations Sustainable Development Goals SDG 2 (Zero Hunger) and SDG 12 (Responsible Consumption and Production) by promoting novel, sustainable agricultural production modes.

We welcome Original Research, Reviews, and Methods articles. Submissions must focus on algorithmic innovation, edge deployment, or robotic applications aligned with sustainable development goals. All original research manuscripts are required to include rigorous in-field experimental evaluation. Pure simulation-based or modelling-only submissions will not be considered. Themes of interest include, but are not limited to:
(1) Lightweight Machine Vision Architectures: Highly efficient neural networks designed for real-time identification of crop diseases, pests, or weeds at the edge, with field-tested performance validation.
(2) Edge-AI Deployment on Ag-Robotics: Hardware-software co-optimization, model quantization, and acceleration on phytoprotection robots and UAVs to achieve precise, targeted chemical applications verified in field environments.
(3) Vision-Based Navigation Mechanisms: Crop row recognition, visual path planning, and multi-source information fusion navigation for autonomous phytoprotection equipment, validated under real-world field scenarios.
(4) Algorithmic Robustness in Field Environments: Robust visual detection solutions capable of handling harsh field conditions, such as drastic outdoor lighting changes, heavy occlusions, and complex backgrounds.
(5) Intelligent Ecological and Biological Control: Distributed phytoprotection monitoring frameworks driven by IoT, multi-sensor systems, and AI to minimize reliance on chemical pesticides, supported by practical field evaluation data.
---------------------
We welcome contributions from researchers across the field and are happy to support you as you prepare your submission. If you would like initial feedback on whether your planned manuscript aligns with the scope of this Research Topic, please reach out to us at plantscience@frontiersin.org before submitting. Early input can help ensure your work is a strong fit and save you time later in the process.

If you would like to join us, simply click “Participate in this topic” at the top right of this page. This lets us know you plan to contribute, ensures you never miss key deadlines, and gives you access to helpful resources to guide you through the submission process. We very much look forward to your participation and to building a vibrant and cohesive Research Topic together.