Smart agricultural technology最新文献_第7页

Smart energy supply for smart farms: A hybrid ground source heat pump in seasonal greenhouses for growing cherries 智能农场的智能能源供应：用于种植樱桃的季节性温室中的混合地源热泵

IF 5.7 Q1 AGRICULTURAL ENGINEERING

Smart agricultural technology

Pub Date : 2025-12-11 DOI: 10.1016/j.atech.2025.101709

Biao Zhou, Rumeng Zhang, Zhiming Zhang, Kaidi Li, Jianye Wang, Jiufa Chen, George Papadakis, Bin Guo

Growing off-season cherries in seasonal greenhouses is beneficial for increasing the income of growers. The traditional compressed air conditioners commonly used in seasonal greenhouses have the problem of high operating costs. Therefore, in this paper, the TRNSYS simulation software was used to evaluate the operation of a hybrid ground source heat pump system (HGSHP) composed of ground source heat pumps, photovoltaic thermal panels (PV/T), and cooling towers (CT) in seasonal greenhouses. The system features a cooling tower with a heat dissipation capacity of 70.4 kW and PV/T components covering an area of 40 m². The operational performance and feasibility of this system in the Shandong region were investigated.

The results show that compared to a system using only a ground-source heat pump (GSHP), after five years of operation, the soil temperature increased by 6.3 °C for the GSHP, while the HGSHP system resulted in an increase of 0.2 °C. The addition of the cooling tower reduced the number of vertical ground-source heat exchangers from 30 to 22, resulting in a reduction of 46,959 CNY in energy consumption over five years of operation. Compared to GSHP systems, the HGSHP system has an additional payback period of approximately 6.0 years, yielding equivalent annual cost savings of 5981 CNY and reducing carbon dioxide emissions by 13.1 tons. These findings highlight the HGSHP system's potential for providing efficient heating and cooling in seasonal greenhouses, along with its value in terms of high economic viability and long-term sustainability.

在季节性温室中种植反季樱桃有利于增加种植者的收入。季节性大棚常用的传统压缩空调存在运行成本高的问题。因此，本文采用TRNSYS仿真软件对季节性温室中由地源热泵、光伏热板（PV/T）和冷却塔（CT）组成的混合型地源热泵系统（HGSHP）的运行情况进行了评估。系统采用冷却塔散热能力70.4 kW，光伏/T组件面积40 m²。对该系统在山东地区的运行性能和可行性进行了研究。结果表明，与仅使用地源热泵相比，地源热泵系统运行5年后土壤温度升高了6.3℃，而地源热泵系统的土壤温度升高了0.2℃。冷却塔的增加使垂直地源换热器的数量从30台减少到22台，运行5年减少能耗46,959元。与地源热泵系统相比，地源热泵系统的投资回收期约为6.0年，相当于每年节约成本5981元，减少二氧化碳排放13.1吨。这些发现突出了HGSHP系统在季节性温室中提供高效供暖和制冷的潜力，以及它在高经济可行性和长期可持续性方面的价值。

{"title":"Smart energy supply for smart farms: A hybrid ground source heat pump in seasonal greenhouses for growing cherries","authors":"Biao Zhou, Rumeng Zhang, Zhiming Zhang, Kaidi Li, Jianye Wang, Jiufa Chen, George Papadakis, Bin Guo","doi":"10.1016/j.atech.2025.101709","DOIUrl":"10.1016/j.atech.2025.101709","url":null,"abstract":"<div><div>Growing off-season cherries in seasonal greenhouses is beneficial for increasing the income of growers. The traditional compressed air conditioners commonly used in seasonal greenhouses have the problem of high operating costs. Therefore, in this paper, the TRNSYS simulation software was used to evaluate the operation of a hybrid ground source heat pump system (HGSHP) composed of ground source heat pumps, photovoltaic thermal panels (PV/T), and cooling towers (CT) in seasonal greenhouses. The system features a cooling tower with a heat dissipation capacity of 70.4 kW and PV/T components covering an area of 40 m². The operational performance and feasibility of this system in the Shandong region were investigated.</div><div>The results show that compared to a system using only a ground-source heat pump (GSHP), after five years of operation, the soil temperature increased by 6.3 °C for the GSHP, while the HGSHP system resulted in an increase of 0.2 °C. The addition of the cooling tower reduced the number of vertical ground-source heat exchangers from 30 to 22, resulting in a reduction of 46,959 CNY in energy consumption over five years of operation. Compared to GSHP systems, the HGSHP system has an additional payback period of approximately 6.0 years, yielding equivalent annual cost savings of 5981 CNY and reducing carbon dioxide emissions by 13.1 tons. These findings highlight the HGSHP system's potential for providing efficient heating and cooling in seasonal greenhouses, along with its value in terms of high economic viability and long-term sustainability.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101709"},"PeriodicalIF":5.7,"publicationDate":"2025-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145790646","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Smart greenhouse climate control with real-time fault detection and energy-aware automation 智能温室气候控制，实时故障检测和能源感知自动化

IF 5.7 Q1 AGRICULTURAL ENGINEERING

Smart agricultural technology

Pub Date : 2025-12-09 DOI: 10.1016/j.atech.2025.101707

Li Zeping , Noramalina Abdullah , Mohamad Khairi Ishak

In response to increasing global climate variability and the environmental sensitivity of crop production, greenhouse cultivation has become an essential agricultural strategy. This study proposes a low-cost, modular intelligent temperature control system designed specifically for greenhouses. The system integrates ZigBee-based environmental sensing, ESP32-based edge computing, and the Home Assistant platform. Leveraging DHT11 temperature sensors, Tuya smart plugs, and low-code configuration via ESPHome, the architecture enables real-time climate monitoring and automated environmental regulation. A prototype greenhouse was constructed to experimentally evaluate system performance across five key dimensions: sensor placement, response time, energy efficiency, fault tolerance, and ZigBee communication range. Results show that sensors positioned at crop canopy height provided the most representative environmental data. The system maintained a total control latency under 90 s, balancing responsiveness with optimized energy use. A comparative analysis of two temperature control strategies, with ranges of 32 to 35 °C and 33 to 34 °C, respectively, revealed that the stricter range led to 2.2 times greater energy consumption, underscoring the inherent balance between temperature regulation precision and energy efficiency. ZigBee communication achieved over 140 m of line-of-sight range and demonstrated rapid self-healing capability under network disruption. The proposed approach supports the development of intelligent, data-driven environmental control systems for future smart farming applications and precision agriculture.

为了应对日益增加的全球气候变率和作物生产的环境敏感性，温室栽培已成为必不可少的农业战略。本研究提出了一种专门为温室设计的低成本、模块化智能温度控制系统。该系统集成了基于zigbee的环境感知、基于esp32的边缘计算和家庭助理平台。利用DHT11温度传感器、Tuya智能插头和ESPHome的低代码配置，该架构可以实现实时气候监测和自动化环境调节。构建了一个原型温室，从五个关键维度实验性地评估系统性能：传感器放置、响应时间、能效、容错性和ZigBee通信范围。结果表明，位于作物冠层高度的传感器提供了最具代表性的环境数据。系统将总控制延迟保持在90秒以下，平衡了响应性和优化的能源使用。通过对两种温度控制策略（分别为32 - 35°C和33 - 34°C）的对比分析，发现更严格的温度控制范围导致能耗增加2.2倍，强调了温度调节精度和能源效率之间的内在平衡。ZigBee通信实现了超过140米的视距范围，并展示了网络中断下的快速自愈能力。提出的方法支持智能，数据驱动的环境控制系统的发展，用于未来的智能农业应用和精准农业。

{"title":"Smart greenhouse climate control with real-time fault detection and energy-aware automation","authors":"Li Zeping , Noramalina Abdullah , Mohamad Khairi Ishak","doi":"10.1016/j.atech.2025.101707","DOIUrl":"10.1016/j.atech.2025.101707","url":null,"abstract":"<div><div>In response to increasing global climate variability and the environmental sensitivity of crop production, greenhouse cultivation has become an essential agricultural strategy. This study proposes a low-cost, modular intelligent temperature control system designed specifically for greenhouses. The system integrates ZigBee-based environmental sensing, ESP32-based edge computing, and the Home Assistant platform. Leveraging DHT11 temperature sensors, Tuya smart plugs, and low-code configuration via ESPHome, the architecture enables real-time climate monitoring and automated environmental regulation. A prototype greenhouse was constructed to experimentally evaluate system performance across five key dimensions: sensor placement, response time, energy efficiency, fault tolerance, and ZigBee communication range. Results show that sensors positioned at crop canopy height provided the most representative environmental data. The system maintained a total control latency under 90 s, balancing responsiveness with optimized energy use. A comparative analysis of two temperature control strategies, with ranges of 32 to 35 °C and 33 to 34 °C, respectively, revealed that the stricter range led to 2.2 times greater energy consumption, underscoring the inherent balance between temperature regulation precision and energy efficiency. ZigBee communication achieved over 140 m of line-of-sight range and demonstrated rapid self-healing capability under network disruption. The proposed approach supports the development of intelligent, data-driven environmental control systems for future smart farming applications and precision agriculture.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101707"},"PeriodicalIF":5.7,"publicationDate":"2025-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145790773","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Path planning for indoor potted plant maintenance robots based on IACO and IAPF algorithms 基于IACO和IAPF算法的室内盆栽养护机器人路径规划

IF 5.7 Q1 AGRICULTURAL ENGINEERING

Smart agricultural technology

Pub Date : 2025-12-09 DOI: 10.1016/j.atech.2025.101701

Siyu Chen , Xiaoyang Zhang , Xi Wang , Chunshan Liu

In the complex indoor potted plant environment, the intelligent path planning algorithm directly affects the guarantee of the operational efficiency and stability of the robot. The disordered layout of the indoor potted plant environment contains many unknown obstacles, which impose high requirements on accuracy of positioning. An intelligent path navigation algorithm needs to plan reliable and fast navigation routes in a dynamic and chaotic environment. This paper proposes a collaborative path navigation algorithm that integrates the improved ant colony algorithm and the improved artificial potential field algorithm. Firstly, in the local guidance function for path selection in the ant colony algorithm, factors such as obstacle density, gravitational influence, dynamic monitoring of pheromone evaporation coefficient, and secondary optimization of the path are added to generate the global optimal path. Secondly, the obstacle potential field functions in the potential field guidance algorithm and the method of setting virtual target location are improved to solve the problems of failing to achieve the goal and escaping from local traps, thereby improving the safety distance and operational stability. Ultimately, the efficiency of the algorithm was confirmed through simulation test and actual indoor potting experiments. The simulation test show that the ant colony algorithm based on improved strategies reduces the turning points by 65 % and the path length by 11.69 %. The results of the actual indoor potting environment experiments indicate that under two different paths in scenarios, the collaborative path navigation algorithm reduces the average navigation positioning error by 7.60 % and the alignment deviation by 30.72 %, and the safety distance is increased by 20 %.

在复杂的室内盆栽环境中，智能路径规划算法直接影响到机器人运行效率和稳定性的保证。室内盆栽植物环境的无序布局包含许多未知障碍物，这对定位精度提出了很高的要求。智能路径导航算法需要在动态、混沌的环境中规划出可靠、快速的导航路径。本文提出了一种结合改进蚁群算法和改进人工势场算法的协同路径导航算法。首先，在蚁群算法路径选择的局部引导函数中，加入障碍物密度、重力影响、信息素蒸发系数动态监测、路径二次优化等因素，生成全局最优路径；其次，改进了势场制导算法中的障碍物势场函数和虚拟目标位置设置方法，解决了未达到目标和逃离局部陷阱的问题，提高了安全距离和运行稳定性；最终通过模拟试验和室内盆栽试验验证了算法的有效性。仿真实验表明，基于改进策略的蚁群算法将路径拐点减少65%，路径长度减少11.69%。实际室内盆栽环境实验结果表明，在两种不同路径场景下，协同路径导航算法将平均导航定位误差降低了7.60%，对准偏差降低了30.72%，安全距离提高了20%。

{"title":"Path planning for indoor potted plant maintenance robots based on IACO and IAPF algorithms","authors":"Siyu Chen , Xiaoyang Zhang , Xi Wang , Chunshan Liu","doi":"10.1016/j.atech.2025.101701","DOIUrl":"10.1016/j.atech.2025.101701","url":null,"abstract":"<div><div>In the complex indoor potted plant environment, the intelligent path planning algorithm directly affects the guarantee of the operational efficiency and stability of the robot. The disordered layout of the indoor potted plant environment contains many unknown obstacles, which impose high requirements on accuracy of positioning. An intelligent path navigation algorithm needs to plan reliable and fast navigation routes in a dynamic and chaotic environment. This paper proposes a collaborative path navigation algorithm that integrates the improved ant colony algorithm and the improved artificial potential field algorithm. Firstly, in the local guidance function for path selection in the ant colony algorithm, factors such as obstacle density, gravitational influence, dynamic monitoring of pheromone evaporation coefficient, and secondary optimization of the path are added to generate the global optimal path. Secondly, the obstacle potential field functions in the potential field guidance algorithm and the method of setting virtual target location are improved to solve the problems of failing to achieve the goal and escaping from local traps, thereby improving the safety distance and operational stability. Ultimately, the efficiency of the algorithm was confirmed through simulation test and actual indoor potting experiments. The simulation test show that the ant colony algorithm based on improved strategies reduces the turning points by 65 % and the path length by 11.69 %. The results of the actual indoor potting environment experiments indicate that under two different paths in scenarios, the collaborative path navigation algorithm reduces the average navigation positioning error by 7.60 % and the alignment deviation by 30.72 %, and the safety distance is increased by 20 %.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101701"},"PeriodicalIF":5.7,"publicationDate":"2025-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145790768","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

FIEA: An android tool for sustainable furrow irrigation FIEA：可持续沟灌的机器人工具

IF 5.7 Q1 AGRICULTURAL ENGINEERING

Smart agricultural technology

Pub Date : 2025-12-08 DOI: 10.1016/j.atech.2025.101704

Lenard Kumwenda , Grivin Chipula , Patsani Gregory Kumambala , Thomas Nyanda Reuben , Lameck Fiwa , Stanley Phiri

Furrow irrigation performance is often constrained by limited field measurements and the reliance on desktop hydraulic models unsuitable for data-scarce environments. This study presents the Furrow Irrigation Evaluation App (FIEA), an Android-based tool that computes application efficiency (AE), runoff depth (RO), and deep percolation (DP) using simplified SCS hydraulic relationships. FIEA was validated against WinSRFR and SURDEV using published datasets and Monte-Carlo–generated scenarios. Results show high agreement for AE (R² = 0.92–0.99; NRMSE = 1.8–9.6 %) and RO (R² = 0.91–0.99), while DP showed lower correspondence (R² = 0.02–0.31) due to geometric simplifications. Sensitivity analysis identified inflow rate and furrow length as dominant drivers (|r| ≤ 0.62). By providing an offline, field-ready alternative to desktop models, FIEA enables rapid diagnosis of water losses and supports sustainable, climate-resilient irrigation management. Future improvements will refine DP estimation through enhanced geometric representation and sensor-integrated modelling.

沟灌性能常常受到有限的田间测量和依赖于不适合数据匮乏环境的台式水力模型的限制。本研究介绍了沟灌评估应用程序（FIEA），这是一个基于android的工具，可以使用简化的SCS水力关系计算应用效率（AE）、径流深度（RO）和深层渗透（DP）。FIEA使用已发布的数据集和蒙特卡罗生成的场景对WinSRFR和SURDEV进行了验证。结果表明，AE （R²= 0.92 ~ 0.99;NRMSE = 1.8 ~ 9.6%）和RO （R²= 0.91 ~ 0.99）的一致性较高，DP （R²= 0.02 ~ 0.31）的一致性较低。敏感性分析发现流入速率和沟长是主要驱动因素（|或|≤0.62）。通过提供离线、现场可用的桌面模型替代方案，FIEA可以快速诊断水分流失，并支持可持续的、适应气候变化的灌溉管理。未来的改进将通过增强几何表示和传感器集成建模来改进DP估计。

{"title":"FIEA: An android tool for sustainable furrow irrigation","authors":"Lenard Kumwenda , Grivin Chipula , Patsani Gregory Kumambala , Thomas Nyanda Reuben , Lameck Fiwa , Stanley Phiri","doi":"10.1016/j.atech.2025.101704","DOIUrl":"10.1016/j.atech.2025.101704","url":null,"abstract":"<div><div>Furrow irrigation performance is often constrained by limited field measurements and the reliance on desktop hydraulic models unsuitable for data-scarce environments. This study presents the Furrow Irrigation Evaluation App (FIEA), an Android-based tool that computes application efficiency (AE), runoff depth (RO), and deep percolation (DP) using simplified SCS hydraulic relationships. FIEA was validated against WinSRFR and SURDEV using published datasets and Monte-Carlo–generated scenarios. Results show high agreement for AE (R² = 0.92–0.99; NRMSE = 1.8–9.6 %) and RO (R² = 0.91–0.99), while DP showed lower correspondence (R² = 0.02–0.31) due to geometric simplifications. Sensitivity analysis identified inflow rate and furrow length as dominant drivers (|r| ≤ 0.62). By providing an offline, field-ready alternative to desktop models, FIEA enables rapid diagnosis of water losses and supports sustainable, climate-resilient irrigation management. Future improvements will refine DP estimation through enhanced geometric representation and sensor-integrated modelling.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101704"},"PeriodicalIF":5.7,"publicationDate":"2025-12-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145737827","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

From mechanistic-driven to data-driven: A review of the evolution of crop models 从机械驱动到数据驱动：作物模型的发展综述

IF 5.7 Q1 AGRICULTURAL ENGINEERING

Smart agricultural technology

Pub Date : 2025-12-08 DOI: 10.1016/j.atech.2025.101699

Lebing Zheng , Hong-Yu Zhang , Shanmei Liu , Zaiwen Feng , JunWen He , Hai Liang , Fang Tian , Hui Peng

The convergence of climate change,resource scarcity, and rising global food demand necessitates advanced tools for sustainable agricultural intensification.Traditional farming practices, often based on static guidelines,are increasingly inadequate to manage the nonlinear and interactive effects of multiple stressors. Crop models—originally mechanistic, process-based simulators—have evolved into hybrid, data-integrated systems that support precision and intelligent agriculture. This review traces their evolution from early physiological simulations to contemporary paradigms combining mechanistic interpretability with machine learning adaptability,and examines applications in crop growth simulation, management optimization and strategic decision-making. Persistent challenges, including parameter overfitting,computational demands and limited cross-regional transferability, highlight the need for “mechanism-guided, data-enhanced” approaches that anchor interpretability in physiological knowledge while leveraging data-driven flexibility. This synthesis provides both the conceptual and technical foundation for the development of next-generation crop models, offering theoretical support for more precise and adaptive decision-making in smart agriculture.

气候变化、资源短缺和全球粮食需求上升三者的共同作用，需要先进的工具来实现可持续农业集约化。传统的农业实践通常基于静态准则，越来越不足以管理多种压力源的非线性和相互作用效应。作物模型——最初是机械性的、基于过程的模拟器——已经演变成混合的、数据集成的系统，支持精准和智能农业。本文回顾了它们从早期生理模拟到结合机械可解释性和机器学习适应性的当代范式的演变，并研究了它们在作物生长模拟、管理优化和战略决策中的应用。持续存在的挑战，包括参数过拟合、计算需求和有限的跨区域可转移性，突出了对“机制指导、数据增强”方法的需求，这些方法可以在利用数据驱动的灵活性的同时，锚定生理知识的可解释性。这种综合为下一代作物模型的开发提供了概念和技术基础，为智能农业中更精确和适应性的决策提供了理论支持。

{"title":"From mechanistic-driven to data-driven: A review of the evolution of crop models","authors":"Lebing Zheng , Hong-Yu Zhang , Shanmei Liu , Zaiwen Feng , JunWen He , Hai Liang , Fang Tian , Hui Peng","doi":"10.1016/j.atech.2025.101699","DOIUrl":"10.1016/j.atech.2025.101699","url":null,"abstract":"<div><div>The convergence of climate change,resource scarcity, and rising global food demand necessitates advanced tools for sustainable agricultural intensification.Traditional farming practices, often based on static guidelines,are increasingly inadequate to manage the nonlinear and interactive effects of multiple stressors. Crop models—originally mechanistic, process-based simulators—have evolved into hybrid, data-integrated systems that support precision and intelligent agriculture. This review traces their evolution from early physiological simulations to contemporary paradigms combining mechanistic interpretability with machine learning adaptability,and examines applications in crop growth simulation, management optimization and strategic decision-making. Persistent challenges, including parameter overfitting,computational demands and limited cross-regional transferability, highlight the need for “mechanism-guided, data-enhanced” approaches that anchor interpretability in physiological knowledge while leveraging data-driven flexibility. This synthesis provides both the conceptual and technical foundation for the development of next-generation crop models, offering theoretical support for more precise and adaptive decision-making in smart agriculture.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101699"},"PeriodicalIF":5.7,"publicationDate":"2025-12-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145791350","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A UAV-based tobacco plant detection model integrating NDVI and multi-scale feature fusion for precision agriculture 融合NDVI和多尺度特征融合的精准农业无人机烟草植物检测模型

IF 5.7 Q1 AGRICULTURAL ENGINEERING

Smart agricultural technology

Pub Date : 2025-12-07 DOI: 10.1016/j.atech.2025.101703

Xinbao Chen, Junqi Lei, Yaohui Zhang, Xianzhao Liu, Xiangyue Chen

Context

Accurate and early detection of tobacco plants is essential for optimizing field management and ensuring stable yield in precision agriculture. Yet, achieving reliable detection at the transplanting stage using Unmanned Aerial Vehicles (UAVs) is particularly challenging due to complex soil backgrounds and the prevalence of small, obscure targets.

Objective

This study aims to develop a robust UAV-based detection framework that integrates vegetation indices with deep learning to enhance discrimination between crops and non-crops during the critical transplanting stage.

Methods

We propose YOLONTD, an innovative detection framework that incorporates NDVI (Normalized Difference Vegetation Index) spectral information into a deep learning pipeline. The architecture integrates three dedicated modules: (i) Small Object Enhanced Pyramid (SOEP) for capturing fine-grained features, (ii) Feature Complementary Mapping (FCM) for enriching multi-scale contextual information, and (iii) Fusion and Pyramid Spatial Channel (FPSC) for optimized feature fusion. Additionally, the Normalized Wasserstein Distance (NWD) metric is introduced to reduce localization sensitivity in small-object detection.

Results and conclusions

Experimental results show that YOLONTD achieves state-of-the-art performance, reaching 69.9 % mAP@50–95 and 54.6% APtiny, significantly surpassing the baseline model while maintaining low computational overhead. These findings confirm the efficacy of combining vegetation indices with deep learning for enhanced small-object detection.

Significance

This study provides a reliable and efficient solution for UAV-based crop monitoring, demonstrating that the integration of spectral indices with advanced detection models can substantially improve precision agriculture practices, particularly in early-stage crop management.

在精准农业中，准确和早期检测烟草植株对优化田间管理和确保稳定产量至关重要。然而，由于复杂的土壤背景和小而模糊的目标普遍存在，在移植阶段使用无人机（uav）实现可靠的检测尤其具有挑战性。本研究旨在开发一个强大的基于无人机的检测框架，该框架将植被指数与深度学习相结合，以增强移栽关键阶段作物和非作物的区分。方法提出了一种将归一化植被指数（NDVI）光谱信息整合到深度学习管道中的创新检测框架YOLONTD。该架构集成了三个专用模块：(i)用于捕获细粒度特征的小对象增强金字塔（SOEP），（ii）用于丰富多尺度上下文信息的特征互补映射（FCM），以及（iii）用于优化特征融合的融合和金字塔空间通道（FPSC）。此外，引入归一化Wasserstein距离（NWD）度量来降低小目标检测中的定位灵敏度。结果与结论实验结果表明，YOLONTD达到了最先进的性能，达到了69.9% % mAP@50 -95和54.6% APtiny，在保持较低的计算开销的同时显著超过了基线模型。这些发现证实了将植被指数与深度学习结合起来增强小目标检测的有效性。本研究为基于无人机的作物监测提供了一种可靠、高效的解决方案，表明将光谱指数与先进的检测模型相结合可以极大地提高精准农业实践水平，特别是在作物早期管理方面。

{"title":"A UAV-based tobacco plant detection model integrating NDVI and multi-scale feature fusion for precision agriculture","authors":"Xinbao Chen, Junqi Lei, Yaohui Zhang, Xianzhao Liu, Xiangyue Chen","doi":"10.1016/j.atech.2025.101703","DOIUrl":"10.1016/j.atech.2025.101703","url":null,"abstract":"<div><h3>Context</h3><div>Accurate and early detection of tobacco plants is essential for optimizing field management and ensuring stable yield in precision agriculture. Yet, achieving reliable detection at the transplanting stage using Unmanned Aerial Vehicles (UAVs) is particularly challenging due to complex soil backgrounds and the prevalence of small, obscure targets.</div></div><div><h3>Objective</h3><div>This study aims to develop a robust UAV-based detection framework that integrates vegetation indices with deep learning to enhance discrimination between crops and non-crops during the critical transplanting stage.</div></div><div><h3>Methods</h3><div>We propose YOLO<img>NTD, an innovative detection framework that incorporates NDVI (Normalized Difference Vegetation Index) spectral information into a deep learning pipeline. The architecture integrates three dedicated modules: (i) Small Object Enhanced Pyramid (SOEP) for capturing fine-grained features, (ii) Feature Complementary Mapping (FCM) for enriching multi-scale contextual information, and (iii) Fusion and Pyramid Spatial Channel (FPSC) for optimized feature fusion. Additionally, the Normalized Wasserstein Distance (NWD) metric is introduced to reduce localization sensitivity in small-object detection.</div></div><div><h3>Results and conclusions</h3><div>Experimental results show that YOLO<img>NTD achieves state-of-the-art performance, reaching 69.9 % mAP@50–95 and 54.6% APtiny, significantly surpassing the baseline model while maintaining low computational overhead. These findings confirm the efficacy of combining vegetation indices with deep learning for enhanced small-object detection.</div></div><div><h3>Significance</h3><div>This study provides a reliable and efficient solution for UAV-based crop monitoring, demonstrating that the integration of spectral indices with advanced detection models can substantially improve precision agriculture practices, particularly in early-stage crop management.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101703"},"PeriodicalIF":5.7,"publicationDate":"2025-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145737824","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Classification of soybean genotypes based on physiological clustering (PCA + k-means) integrated with VIS-NIR hyperspectral data and machine learning models 结合VIS-NIR高光谱数据和机器学习模型的生理聚类（PCA + k-means）大豆基因型分类

IF 5.7 Q1 AGRICULTURAL ENGINEERING

Smart agricultural technology

Pub Date : 2025-12-07 DOI: 10.1016/j.atech.2025.101702

Regimar Garcia dos Santos , Dthenifer Cordeiro Santana , Larissa Pereira Ribeiro Teodoro , Cid Naudi Silva Campos , Fabio Henrique Rojo Baio , Carlos Antônio da Silva Junior , Elber Vinicius Martins Silva , Luan Pereira de Oliveira , Paulo Eduardo Teodoro

Given the wide variety of genotypes, classifying them based on physiological traits can aid breeding efforts. This study aimed to identify the most effective machine learning algorithms and input variables for classifying soybean genotypes according to physiological features. A field experiment was conducted using a random complete block design with three replications, testing 32 soybean genotypes. Multispectral and hyperspectral reflectance data were collected and grouped into 20 representative wavelength ranges. The physiological traits measured included net photosynthesis, internal CO₂ concentration, stomatal conductance, and transpiration rates. Genotype groups based on physiological performance were classified with the k-means algorithm and principal component analysis, resulting in two clusters. K-means is an algorithm that groups data into k similar groups. It does this by finding patterns and placing each sample in the group whose center (centroid) it is closest to. These clusters served as output variables in machine learning models, with input variables including wavelengths and spectral band averages. A hyperspectral sensor, was employed to record leaf reflectance at wavelengths ranging from 450 to 824 nm in the laboratory under artificial lighting. The algorithms tested were artificial neural networks, J48 decision trees, REPTree, support vector machines (SVM), random forest, and logistic regression (LR) as a control. Model accuracy was assessed using correct classification percentage and F-score. SVM and RL stood out with accuracies above 0.6 in classifications for CC and an F-score exceeding 0.75. When using spectral bands as predictors, both showed similar performance, but with wavelengths as predictors, WL becomes a robust input for the models because its complete spectrum information provides important data for group classifications.

鉴于基因型的多样性，基于生理特征对它们进行分类有助于育种工作。本研究旨在确定最有效的机器学习算法和输入变量，以根据生理特征对大豆基因型进行分类。采用3个重复的随机完全区组设计，对32个大豆基因型进行了田间试验。收集多光谱和高光谱反射率数据，并将其分为20个代表性波长范围。测定的生理性状包括净光合作用、内部CO2浓度、气孔导度和蒸腾速率。采用k-means算法和主成分分析法对生理性能基因型组进行分类，得到2个聚类。k -means是一种将数据分成k个相似组的算法。它通过寻找模式并将每个样本放置在其中心（质心）最接近的组中来做到这一点。这些集群作为机器学习模型的输出变量，输入变量包括波长和光谱带平均值。采用高光谱传感器，在实验室人工光照下记录450 ~ 824 nm波长的叶片反射率。测试的算法有人工神经网络、J48决策树、REPTree、支持向量机（SVM）、随机森林和逻辑回归（LR）作为对照。采用正确分类百分比和f分数评估模型准确性。SVM和RL在CC分类中准确率在0.6以上，f值超过0.75。当使用光谱波段作为预测因子时，两者表现出相似的性能，但使用波长作为预测因子时，WL成为模型的稳健输入，因为其完整的光谱信息为群体分类提供了重要的数据。

{"title":"Classification of soybean genotypes based on physiological clustering (PCA + k-means) integrated with VIS-NIR hyperspectral data and machine learning models","authors":"Regimar Garcia dos Santos , Dthenifer Cordeiro Santana , Larissa Pereira Ribeiro Teodoro , Cid Naudi Silva Campos , Fabio Henrique Rojo Baio , Carlos Antônio da Silva Junior , Elber Vinicius Martins Silva , Luan Pereira de Oliveira , Paulo Eduardo Teodoro","doi":"10.1016/j.atech.2025.101702","DOIUrl":"10.1016/j.atech.2025.101702","url":null,"abstract":"<div><div>Given the wide variety of genotypes, classifying them based on physiological traits can aid breeding efforts. This study aimed to identify the most effective machine learning algorithms and input variables for classifying soybean genotypes according to physiological features. A field experiment was conducted using a random complete block design with three replications, testing 32 soybean genotypes. Multispectral and hyperspectral reflectance data were collected and grouped into 20 representative wavelength ranges. The physiological traits measured included net photosynthesis, internal CO<sub>2</sub> concentration, stomatal conductance, and transpiration rates. Genotype groups based on physiological performance were classified with the k-means algorithm and principal component analysis, resulting in two clusters. K-means is an algorithm that groups data into k similar groups. It does this by finding patterns and placing each sample in the group whose center (centroid) it is closest to. These clusters served as output variables in machine learning models, with input variables including wavelengths and spectral band averages. A hyperspectral sensor, was employed to record leaf reflectance at wavelengths ranging from 450 to 824 nm in the laboratory under artificial lighting. The algorithms tested were artificial neural networks, J48 decision trees, REPTree, support vector machines (SVM), random forest, and logistic regression (LR) as a control. Model accuracy was assessed using correct classification percentage and F-score. SVM and RL stood out with accuracies above 0.6 in classifications for CC and an F-score exceeding 0.75. When using spectral bands as predictors, both showed similar performance, but with wavelengths as predictors, WL becomes a robust input for the models because its complete spectrum information provides important data for group classifications.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101702"},"PeriodicalIF":5.7,"publicationDate":"2025-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145925720","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Real-time sunflower detection using semi-supervised and self-supervised deep learning for precision agriculture 基于半监督和自监督深度学习的精准农业实时向日葵检测

IF 5.7 Q1 AGRICULTURAL ENGINEERING

Smart agricultural technology

Pub Date : 2025-12-07 DOI: 10.1016/j.atech.2025.101684

Fathhur Rahaman Sams, Sanjana Kazi Supti, Shayma Binte Hamid, Radin Junayed, K.M. Fahim A Bari, Md Junaeid Ali, Raiyan Gani, Karib Shams, Mohammad Rifat Ahmmad Rashid, Raihan Ul Islam

Accurate sunflower head detection is essential for precision agriculture, supporting timely monitoring and yield estimation. However, reliable detection under UAV settings remains challenging due to annotation scarcity, variable field conditions, and inconsistent localization across flowering stages. This study presents a unified framework that evaluates supervised, semi-supervised, and self-supervised learning strategies on UAV imagery collected under real field conditions. In the supervised setting, YOLOv12s achieved the strongest performance (mAP@50  ≈  93 %), with stable convergence and focused visual attention, while RF-DETR showed lower recall and weaker localization. To reduce annotation requirements, a Pseudo-STAC teacher–student approach was evaluated across varying labeled-to-unlabeled ratios. Teacher models maintained high accuracy even with limited supervision (mAP@50 = 88.5–91.6 %), while student models approached teacher-level performance when 20–30 % of images were labeled. At extremely low label ratios (10 %), instability from pseudo-label noise was observed, though confidence-adaptive filtering alleviated some of these effects. Self-supervised learning (SSL) using DINOv2-style and BYOL pretraining further strengthened representation quality, consistently producing mAP@50 scores above 91 %. SSL-enhanced YOLOv12s generated compact and discriminative embeddings and exhibited smoother optimization, confirmed through loss curves, clustering analyses, and XAI visualizations. Finally, a real-time Streamlit application was developed, enabling image, video, and live-camera detection at up to 22 FPS, demonstrating the practical deployment potential of the proposed framework. This work demonstrates the potential of semi- and self-supervised learning to reduce annotation costs, enhance generalization, and deliver interpretable real-time solutions for precision agriculture.

准确的葵花籽头检测对精准农业至关重要，支持及时监测和产量估算。然而，在无人机设置下的可靠检测仍然具有挑战性，因为注释稀缺性，多变的现场条件和开花阶段不一致的定位。本研究提出了一个统一的框架，用于评估在真实现场条件下收集的无人机图像的监督、半监督和自监督学习策略。在监督设置下，YOLOv12s的表现最强（mAP@50 ≈ 93%），收敛稳定，视觉注意力集中，而RF-DETR的召回率较低，定位能力较弱。为了减少注释需求，伪stac师生方法在不同的标记与未标记比率上进行了评估。即使在有限的监督下，教师模型也保持了很高的准确性（mAP@50 = 88.5 - 91.6%），而学生模型在20 - 30%的图像被标记时，表现接近教师水平。在极低的标签比率（10%）下，观察到伪标签噪声的不稳定性，尽管自适应置信滤波减轻了其中的一些影响。使用dinov2风格和BYOL预训练的自监督学习（Self-supervised learning， SSL）进一步增强了表征质量，mAP@50得分始终在91%以上。ssl增强的YOLOv12s生成紧凑和判别嵌入，并通过损失曲线、聚类分析和XAI可视化证实了更平滑的优化。最后，开发了实时Streamlit应用程序，支持高达22 FPS的图像、视频和实时摄像机检测，展示了所提出框架的实际部署潜力。这项工作证明了半监督学习和自监督学习在降低注释成本、增强泛化和为精准农业提供可解释的实时解决方案方面的潜力。

{"title":"Real-time sunflower detection using semi-supervised and self-supervised deep learning for precision agriculture","authors":"Fathhur Rahaman Sams, Sanjana Kazi Supti, Shayma Binte Hamid, Radin Junayed, K.M. Fahim A Bari, Md Junaeid Ali, Raiyan Gani, Karib Shams, Mohammad Rifat Ahmmad Rashid, Raihan Ul Islam","doi":"10.1016/j.atech.2025.101684","DOIUrl":"10.1016/j.atech.2025.101684","url":null,"abstract":"<div><div>Accurate sunflower head detection is essential for precision agriculture, supporting timely monitoring and yield estimation. However, reliable detection under UAV settings remains challenging due to annotation scarcity, variable field conditions, and inconsistent localization across flowering stages. This study presents a unified framework that evaluates supervised, semi-supervised, and self-supervised learning strategies on UAV imagery collected under real field conditions. In the supervised setting, YOLOv12s achieved the strongest performance (mAP@50  ≈  93 %), with stable convergence and focused visual attention, while RF-DETR showed lower recall and weaker localization. To reduce annotation requirements, a Pseudo-STAC teacher–student approach was evaluated across varying labeled-to-unlabeled ratios. Teacher models maintained high accuracy even with limited supervision (mAP@50 = 88.5–91.6 %), while student models approached teacher-level performance when 20–30 % of images were labeled. At extremely low label ratios (10 %), instability from pseudo-label noise was observed, though confidence-adaptive filtering alleviated some of these effects. Self-supervised learning (SSL) using DINOv2-style and BYOL pretraining further strengthened representation quality, consistently producing mAP@50 scores above 91 %. SSL-enhanced YOLOv12s generated compact and discriminative embeddings and exhibited smoother optimization, confirmed through loss curves, clustering analyses, and XAI visualizations. Finally, a real-time Streamlit application was developed, enabling image, video, and live-camera detection at up to 22 FPS, demonstrating the practical deployment potential of the proposed framework. This work demonstrates the potential of semi- and self-supervised learning to reduce annotation costs, enhance generalization, and deliver interpretable real-time solutions for precision agriculture.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101684"},"PeriodicalIF":5.7,"publicationDate":"2025-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145790644","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

EPAF-DETR:Efficient transformer-based model for abnormal fish behavior detection under water quality anomalies EPAF-DETR：基于变压器的水质异常下鱼类异常行为检测的高效模型

IF 5.7 Q1 AGRICULTURAL ENGINEERING

Smart agricultural technology

Pub Date : 2025-12-05 DOI: 10.1016/j.atech.2025.101681

Xun Chen , Chuang Wang , Siyu Wu , Xinjian Ou

In aquaculture management, the prompt recognition of abnormal fish behaviors induced by external stimuli or diseases is crucial for enhancing breeding efficiency and securing economic returns for farmers. Nevertheless, monitoring such behaviors remains challenging owing to intricate aquatic environments, frequent occlusions, and considerable visual similarities across different behavioral categories. To tackle these issues, this study introduces an Efficient Parallel Attention Fusion Detection Transformer, designated as EPAF-DETR, which is developed to achieve high-precision and robust object detection. By integrating EfficientViT as the backbone network, the computational complexity of the model is significantly reduced. Combined with an adaptive sparse self-attention mechanism and a spatially enhanced feedforward network, an improved AIFI module is introduced to strengthen feature extraction capabilities. Furthermore, Multi-Level Hierarchical Attention Fusion module is designed to enhance the original cross-scale feature fusion component in RT-DETR, enhancing the salience of critical features and further improving detection accuracy. Finally, by incorporating Matchability-Aware Loss function, the model is guided to place greater emphasis on matching low-quality features.These architectural advancements considerably boost the model’s adaptability in demanding underwater settings and augment its capacity to discriminate fine-grained behavioral characteristics of fish. Experimental outcomes indicate that EPAF-DETR attains detection performance while reducing computational costs, achieving an average F1-score of 94 % and a mAP of 95.7 %. In conclusion, the proposed approach effectively addresses detection difficulties in complex aquaculture environments, enabling accurate and reliable identification of anomalous fish behaviors.

在水产养殖管理中，及时识别外部刺激或疾病引起的鱼类异常行为对于提高养殖效率和确保养殖户的经济回报至关重要。然而，由于复杂的水生环境，频繁的闭塞，以及不同行为类别之间相当大的视觉相似性，监测这些行为仍然具有挑战性。为了解决这些问题，本研究引入了一种高效并行注意力融合检测变压器，称为EPAF-DETR，用于实现高精度和鲁棒的目标检测。通过集成effentvit作为骨干网，大大降低了模型的计算复杂度。结合自适应稀疏自关注机制和空间增强前馈网络，引入改进的AIFI模块增强特征提取能力。此外，设计了多层次注意融合模块，增强RT-DETR中原有的跨尺度特征融合组件，增强关键特征的显著性，进一步提高检测精度。最后，通过引入匹配感知损失函数，引导模型更加重视匹配低质量特征。这些结构上的进步大大提高了模型在苛刻的水下环境中的适应性，并增强了它区分鱼类细粒度行为特征的能力。实验结果表明，EPAF-DETR在降低计算成本的同时达到了检测性能，平均f1得分为94%，mAP为95.7%。综上所述，该方法有效解决了复杂水产养殖环境中的检测困难，能够准确可靠地识别鱼类的异常行为。

{"title":"EPAF-DETR:Efficient transformer-based model for abnormal fish behavior detection under water quality anomalies","authors":"Xun Chen , Chuang Wang , Siyu Wu , Xinjian Ou","doi":"10.1016/j.atech.2025.101681","DOIUrl":"10.1016/j.atech.2025.101681","url":null,"abstract":"<div><div>In aquaculture management, the prompt recognition of abnormal fish behaviors induced by external stimuli or diseases is crucial for enhancing breeding efficiency and securing economic returns for farmers. Nevertheless, monitoring such behaviors remains challenging owing to intricate aquatic environments, frequent occlusions, and considerable visual similarities across different behavioral categories. To tackle these issues, this study introduces an Efficient Parallel Attention Fusion Detection Transformer, designated as EPAF-DETR, which is developed to achieve high-precision and robust object detection. By integrating EfficientViT as the backbone network, the computational complexity of the model is significantly reduced. Combined with an adaptive sparse self-attention mechanism and a spatially enhanced feedforward network, an improved AIFI module is introduced to strengthen feature extraction capabilities. Furthermore, Multi-Level Hierarchical Attention Fusion module is designed to enhance the original cross-scale feature fusion component in RT-DETR, enhancing the salience of critical features and further improving detection accuracy. Finally, by incorporating Matchability-Aware Loss function, the model is guided to place greater emphasis on matching low-quality features.These architectural advancements considerably boost the model’s adaptability in demanding underwater settings and augment its capacity to discriminate fine-grained behavioral characteristics of fish. Experimental outcomes indicate that EPAF-DETR attains detection performance while reducing computational costs, achieving an average F1-score of 94 % and a mAP of 95.7 %. In conclusion, the proposed approach effectively addresses detection difficulties in complex aquaculture environments, enabling accurate and reliable identification of anomalous fish behaviors.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101681"},"PeriodicalIF":5.7,"publicationDate":"2025-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145737823","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

RAIL-WG : Robotic imitation learning for waypoint generation in agricultural autonomous driving RAIL-WG：农业自动驾驶中航路点生成的机器人模仿学习

IF 5.7 Q1 AGRICULTURAL ENGINEERING

Smart agricultural technology

Pub Date : 2025-12-04 DOI: 10.1016/j.atech.2025.101682

Sun Ho Jang, Yong Jun Lee, Woo Jin Ahn, Myo Taeg Lim

Waypoint generation is a critical component of autonomous navigation, directly affecting trajectory accuracy, operational efficiency, and system robustness. Traditional fixed-interval strategies are computationally simple but lack adaptability to dynamic environments, whereas reinforcement learning (RL) methods often face unstable training and limited generalization. To overcome these challenges, we introduce robotic imitation learning for waypoint generation in agricultural autonomous driving (RAIL-WG), an LSTM-based imitation learning framework trained on expert demonstrations. Using the GROW dataset, which contains large-scale, high-resolution GPS trajectories from real-world orchard operations, RAIL-WG learns curvature-adaptive waypoint placement that balances density between straight and curved paths. Extensive simulations and field experiments show that RAIL-WG consistently outperforms both fixed-interval and RL-based baselines in trajectory tracking accuracy, computational efficiency, and smoothness. Beyond agricultural applications, the proposed framework demonstrates strong potential as a generalizable AI model for waypoint optimization, applicable to diverse autonomous systems such as mobile robots, UAVs, and ground vehicles operating in unstructured environments. This versatility highlights RAIL-WG as a scalable solution for adaptive navigation across heterogeneous domains.

路点生成是自主导航的关键组成部分，直接影响轨迹精度、操作效率和系统鲁棒性。传统的固定区间策略计算简单，但缺乏对动态环境的适应性，而强化学习（RL）方法往往面临训练不稳定和泛化有限的问题。为了克服这些挑战，我们引入了用于农业自动驾驶航路点生成的机器人模仿学习（RAIL-WG），这是一种基于lstm的模仿学习框架，经过专家演示训练。使用GROW数据集，其中包含来自真实果园操作的大规模、高分辨率GPS轨迹，RAIL-WG学习曲率自适应路点放置，平衡直线和弯曲路径之间的密度。大量的模拟和现场实验表明，RAIL-WG在轨迹跟踪精度、计算效率和平滑性方面始终优于固定间隔和基于rl的基线。除了农业应用之外，所提出的框架还显示出强大的潜力，可以作为一种通用的人工智能模型，用于航路点优化，适用于各种自主系统，如移动机器人、无人机和在非结构化环境中运行的地面车辆。这种多功能性突出了rails - wg作为跨异构域自适应导航的可伸缩解决方案。

{"title":"RAIL-WG : Robotic imitation learning for waypoint generation in agricultural autonomous driving","authors":"Sun Ho Jang, Yong Jun Lee, Woo Jin Ahn, Myo Taeg Lim","doi":"10.1016/j.atech.2025.101682","DOIUrl":"10.1016/j.atech.2025.101682","url":null,"abstract":"<div><div>Waypoint generation is a critical component of autonomous navigation, directly affecting trajectory accuracy, operational efficiency, and system robustness. Traditional fixed-interval strategies are computationally simple but lack adaptability to dynamic environments, whereas reinforcement learning (RL) methods often face unstable training and limited generalization. To overcome these challenges, we introduce robotic imitation learning for waypoint generation in agricultural autonomous driving (RAIL-WG), an LSTM-based imitation learning framework trained on expert demonstrations. Using the GROW dataset, which contains large-scale, high-resolution GPS trajectories from real-world orchard operations, RAIL-WG learns curvature-adaptive waypoint placement that balances density between straight and curved paths. Extensive simulations and field experiments show that RAIL-WG consistently outperforms both fixed-interval and RL-based baselines in trajectory tracking accuracy, computational efficiency, and smoothness. Beyond agricultural applications, the proposed framework demonstrates strong potential as a generalizable AI model for waypoint optimization, applicable to diverse autonomous systems such as mobile robots, UAVs, and ground vehicles operating in unstructured environments. This versatility highlights RAIL-WG as a scalable solution for adaptive navigation across heterogeneous domains.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"13 ","pages":"Article 101682"},"PeriodicalIF":5.7,"publicationDate":"2025-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145737821","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0