Computers and Electronics in Agriculture最新文献_第5页

Development of a Machine stereo vision-based autonomous navigation system for orchard speed sprayers 为果园快速喷雾器开发基于机器立体视觉的自主导航系统

IF 7.7 1区农林科学 Q1 AGRICULTURE, MULTIDISCIPLINARY

Computers and Electronics in Agriculture

Pub Date : 2024-11-19 DOI: 10.1016/j.compag.2024.109669

Victor Massaki Nakaguchi , R.M. Rasika D. Abeyrathna , Zifu Liu , Ryozo Noguchi , Tofael Ahamed

In orchards, radio frequency scintillation caused by high vegetation density can hinder the effectiveness of machinery auto guidance based on the Global Navigation Satellite System (GNSS). Signal interference not only leads to poor quality of machine operation but can also pose operational risks. In this work, we propose an alternative trajectory positioning system to supplement traditional GNSS-based navigation for machinery used in orchards. The aim of this research was to develop a deep learning machine stereo vision guidance system onboard an orchard speed sprayer. The developed system combines a collision avoidance methodology along with deep learning-driven machine vision for interrow positioning and a dead reckoning set of rules for alternating U-turns. The developed methodology was tested in 4 rows of an artificial orchard. The results show that it is possible for the embedded EfficientDet target detection algorithm to guide the equipment at 1, 1.5 and 2 km × h⁻¹ with minimum average root-mean-square errors (RMSEs) of 0.24 m, 0.20 m, and 0.31 m, respectively. When the system navigation was performed using YOLOv7, the minimum average RMSEs for each row were 0.40 m, 0.48 m, and 0.43 m, respectively, for the abovementioned speeds. The U-turn by dead reckoning showed minimum average RMSE values of 0.56 m, 0.22 m, and 0.35 m for row navigation based on EfficientDet at 1, 1.5 and 2 km × h^−1, respectively. For YOLOv7-based navigation in rows, the minimum average RMSE values at these speeds were 0.35 m, 0.81 m, and 0.44 m, respectively. This study contributes to the field by proposing an alternative navigation system for orchard machines that operates without the limitations associated with GNSS. In addition, our proposed guidance methodology introduces an RGB-D collision avoidance system with a demonstrated safety capacity for navigation under real scenario conditions.

在果园里，高植被密度造成的无线电频率闪烁会妨碍基于全球导航卫星系统（GNSS）的机械自动导航的有效性。信号干扰不仅会导致机器运行质量低下，还会带来操作风险。在这项工作中，我们提出了一种替代轨迹定位系统，以补充传统的基于全球导航卫星系统的果园机械导航。这项研究的目的是在果园快速喷雾器上开发一种深度学习机器立体视觉导航系统。开发的系统结合了避免碰撞方法、用于行间定位的深度学习驱动的机器视觉以及用于交替掉头的死算规则集。在人工果园的 4 行中对所开发的方法进行了测试。结果表明，嵌入式 EfficientDet 目标检测算法能够以 1、1.5 和 2 km × h-1 的速度引导设备，平均均方根误差（RMSE）最小值分别为 0.24 m、0.20 m 和 0.31 m。使用 YOLOv7 进行系统导航时，在上述速度下，每行的最小平均均方根误差分别为 0.40 米、0.48 米和 0.43 米。基于 EfficientDet 的行导航在 1、1.5 和 2 km × h-1 时，通过惯性导航进行 U 形转弯的最小平均有效误差值分别为 0.56 m、0.22 m 和 0.35 m。对于基于 YOLOv7 的行导航，在这些速度下的最小平均 RMSE 值分别为 0.35 m、0.81 m 和 0.44 m。本研究为果园机械提出了一种替代导航系统，其运行不受与全球导航卫星系统相关的限制，从而为该领域做出了贡献。此外，我们提出的制导方法还引入了 RGB-D 防撞系统，该系统在真实场景条件下的导航安全能力已得到证实。

{"title":"Development of a Machine stereo vision-based autonomous navigation system for orchard speed sprayers","authors":"Victor Massaki Nakaguchi , R.M. Rasika D. Abeyrathna , Zifu Liu , Ryozo Noguchi , Tofael Ahamed","doi":"10.1016/j.compag.2024.109669","DOIUrl":"10.1016/j.compag.2024.109669","url":null,"abstract":"<div><div>In orchards, radio frequency scintillation caused by high vegetation density can hinder the effectiveness of machinery auto guidance based on the Global Navigation Satellite System (GNSS). Signal interference not only leads to poor quality of machine operation but can also pose operational risks. In this work, we propose an alternative trajectory positioning system to supplement traditional GNSS-based navigation for machinery used in orchards. The aim of this research was to develop a deep learning machine stereo vision guidance system onboard an orchard speed sprayer. The developed system combines a collision avoidance methodology along with deep learning-driven machine vision for interrow positioning and a dead reckoning set of rules for alternating U-turns. The developed methodology was tested in 4 rows of an artificial orchard. The results show that it is possible for the embedded EfficientDet target detection algorithm to guide the equipment at 1, 1.5 and 2 km × h<sup>−1</sup> with minimum average root-mean-square errors (RMSEs) of 0.24 m, 0.20 m, and 0.31 m, respectively. When the system navigation was performed using YOLOv7, the minimum average RMSEs for each row were 0.40 m, 0.48 m, and 0.43 m, respectively, for the abovementioned speeds. The U-turn by dead reckoning showed minimum average RMSE values of 0.56 m, 0.22 m, and 0.35 m for row navigation based on EfficientDet at 1, 1.5 and 2 km × h<sup>−1,</sup> respectively. For YOLOv7-based navigation in rows, the minimum average RMSE values at these speeds were 0.35 m, 0.81 m, and 0.44 m, respectively. This study contributes to the field by proposing an alternative navigation system for orchard machines that operates without the limitations associated with GNSS. In addition, our proposed guidance methodology introduces an RGB-D collision avoidance system with a demonstrated safety capacity for navigation under real scenario conditions.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109669"},"PeriodicalIF":7.7,"publicationDate":"2024-11-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142699906","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A trajectory tracking control method for the discharge arm of the self-propelled forage harvester 自走式牧草收割机卸料臂的轨迹跟踪控制方法

IF 7.7 1区农林科学 Q1 AGRICULTURE, MULTIDISCIPLINARY

Computers and Electronics in Agriculture

Pub Date : 2024-11-18 DOI: 10.1016/j.compag.2024.109627

Lei Liu , Siyu Hou , Yuefeng Du , Guorun Li , Yucong Wang , Du Chen , Zhongxiang Zhu , Zhenghe Song , Xiaoyu Li

The cooperative operation between the self-propelled forage harvester and the trailer aims to achieve precise and automatic forage unloading. The discharge arm serves as the core structure for conveying forage, and the precision and speed of its motion control are crucial factors influencing the efficiency of forage harvesting. In this context, we proposed a trajectory tracking control method for the discharge arm based on an improved particle swarm optimization (IPSO)-PID controller. Firstly, we utilized the Denavit-Hartenberg (D-H) model of the discharge arm for a positive kinematics solution and the geometric resolution and linear fitting method for the inverse kinematics solution. Secondly, we employed the polynomial interpolation method for trajectory planning on the joint space of the discharge arm, and the PSO algorithm for time-optimal trajectory planning. Then, we designed an IPSO-PID trajectory tracking control algorithm and built an Amesim-Simulink co-simulation model for multiple simulation experiments. Finally, we conducted several performance tests of the discharge arm automatic control system in the workstation and the field, respectively. The experiment results indicate that the performance of the IPSO-PID controller exceeds that of all other controllers, which can meet the discharge arm’s motion control accuracy and speed requirements. Our research results are of great significance for improving the productivity and automation process of the self-propelled forage harvester and provide valuable references for research on automatic and precise control of material loading in other agricultural cooperative harvesting.

自走式牧草收割机与拖车之间的协同作业旨在实现牧草的精确自动卸载。卸料臂作为输送牧草的核心结构，其运动控制的精度和速度是影响牧草收割效率的关键因素。为此，我们提出了一种基于改进型粒子群优化（IPSO）-PID 控制器的卸料臂轨迹跟踪控制方法。首先，我们利用卸料臂的 Denavit-Hartenberg (D-H) 模型进行正运动学求解，并利用几何分辨率和线性拟合方法进行反运动学求解。其次，我们采用多项式插值法对放电臂的关节空间进行轨迹规划，并采用 PSO 算法进行时间最优轨迹规划。然后，我们设计了 IPSO-PID 轨迹跟踪控制算法，并建立了 Amesim-Simulink 协同仿真模型，进行了多次仿真实验。最后，我们分别在工作站和现场对放电臂自动控制系统进行了多次性能测试。实验结果表明，IPSO-PID 控制器的性能超过了其他所有控制器，可以满足放电臂的运动控制精度和速度要求。我们的研究成果对提高自走式牧草收割机的生产效率和自动化进程具有重要意义，也为其他农业合作收割中物料装载的自动精确控制研究提供了有价值的参考。

{"title":"A trajectory tracking control method for the discharge arm of the self-propelled forage harvester","authors":"Lei Liu , Siyu Hou , Yuefeng Du , Guorun Li , Yucong Wang , Du Chen , Zhongxiang Zhu , Zhenghe Song , Xiaoyu Li","doi":"10.1016/j.compag.2024.109627","DOIUrl":"10.1016/j.compag.2024.109627","url":null,"abstract":"<div><div>The cooperative operation between the self-propelled forage harvester and the trailer aims to achieve precise and automatic forage unloading. The discharge arm serves as the core structure for conveying forage, and the precision and speed of its motion control are crucial factors influencing the efficiency of forage harvesting. In this context, we proposed a trajectory tracking control method for the discharge arm based on an improved particle swarm optimization (IPSO)-PID controller. Firstly, we utilized the Denavit-Hartenberg (D-H) model of the discharge arm for a positive kinematics solution and the geometric resolution and linear fitting method for the inverse kinematics solution. Secondly, we employed the polynomial interpolation method for trajectory planning on the joint space of the discharge arm, and the PSO algorithm for time-optimal trajectory planning. Then, we designed an IPSO-PID trajectory tracking control algorithm and built an Amesim-Simulink co-simulation model for multiple simulation experiments. Finally, we conducted several performance tests of the discharge arm automatic control system in the workstation and the field, respectively. The experiment results indicate that the performance of the IPSO-PID controller exceeds that of all other controllers, which can meet the discharge arm’s motion control accuracy and speed requirements. Our research results are of great significance for improving the productivity and automation process of the self-propelled forage harvester and provide valuable references for research on automatic and precise control of material loading in other agricultural cooperative harvesting.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109627"},"PeriodicalIF":7.7,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142699945","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","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 maize lodging types using UAV-SAR remote sensing data and machine learning methods 利用无人机-合成孔径雷达遥感数据和机器学习方法对玉米宿根类型进行分类

IF 7.7 1区农林科学 Q1 AGRICULTURE, MULTIDISCIPLINARY

Computers and Electronics in Agriculture

Pub Date : 2024-11-18 DOI: 10.1016/j.compag.2024.109637

Dashuai Wang , Minghu Zhao , Zhuolin Li , Xiaohu Wu , Nan Li , Decheng Li , Sheng Xu , Xiaoguang Liu

Lodging seriously threatens maize quality and yield and inevitably increases management and harvest costs. Timely collection of crop lodging information plays a pivotal role in the post-disaster assessment and agricultural insurance claims. Although spaceborne radar and optical remote sensing have unparalleled advantages in obtaining large-scale agricultural information, their response capacity to sudden natural maize lodging disasters is insufficient due to the limited spatial–temporal resolution of the satellite data. In recent years, the widespread application of unmanned aerial vehicles (UAVs) based optical remote sensing in precision agriculture has provided an effective alternative to spaceborne remote sensing. However, optical sensing can only effectively reveal the reflectance spectral characteristics of lodging maize under good lighting conditions. This work proposes a novel maize lodging classification method based on UAV synthetic aperture radar (UAV-SAR) and machine learning to circumvent the limitations of spaceborne and UAV-based remote sensing in monitoring maize lodging. Firstly, the raw radar remote sensing data of our study area containing lodging and non-lodging maize plants at the maturity stage is collected by the custom-built X-band and Ku-band UAV-SAR systems. Secondly, the corresponding backscattering coefficients and radar vegetation indices in each lodging type are extracted through radiation calibration and band math. Subsequently, the impacts of radar parameters (bands, polarizations, and observation orientations) and lodging types on backscattering coefficients are comprehensively analyzed. Fourthly, we applied the recursive feature elimination (RFE) algorithm to identify significant feature subsets and constructed multiple datasets using ten filter scales. Finally, five machine learning models (XGBoost, LDA, RF, KNN, and ANN) are trained and tested based on these materials. The classification results under different filter scales and feature combinations show that ANN achieves the best performance with an overall accuracy of 98.26 % and a Kappa coefficient of 0.982. This is the first innovative study successfully introducing cutting-edge UAV-SAR into maize lodging monitoring. Following spaceborne optical, spaceborne radar, and UAV-based optical remote sensing technologies, UAV-SAR holds great potential as the fourth practical means for collecting high-resolution agricultural information.

棉铃虫严重威胁玉米的质量和产量，并不可避免地增加管理和收获成本。及时收集作物冻害信息对灾后评估和农业保险理赔起着至关重要的作用。虽然星载雷达和光学遥感在获取大尺度农业信息方面具有无可比拟的优势，但由于卫星数据的时空分辨率有限，对突发性玉米自然纹枯病灾害的响应能力不足。近年来，基于无人机（UAVs）的光学遥感在精准农业中的广泛应用为空间遥感提供了有效的替代方案。然而，光学传感只能在良好的光照条件下有效地揭示玉米宿存的反射光谱特征。本研究提出了一种基于无人机合成孔径雷达（UAV-SAR）和机器学习的新型玉米冻害分类方法，以规避空间遥感和无人机遥感在监测玉米冻害方面的局限性。首先，定制的 X 波段和 Ku 波段无人机合成孔径雷达系统采集了研究区域内玉米成熟期结瘤和不结瘤植株的原始雷达遥感数据。其次，通过辐射校准和波段数学运算，提取每种宿存类型相应的后向散射系数和雷达植被指数。随后，全面分析了雷达参数（波段、极化和观测方向）和宿主类型对反向散射系数的影响。第四，我们应用递归特征消除（RFE）算法识别重要特征子集，并使用十种滤波器尺度构建了多个数据集。最后，基于这些材料对五种机器学习模型（XGBoost、LDA、RF、KNN 和 ANN）进行了训练和测试。不同过滤尺度和特征组合下的分类结果表明，ANN 的整体准确率为 98.26%，Kappa 系数为 0.982，表现最佳。这是首次将前沿的无人机合成孔径雷达成功引入玉米生育期监测的创新研究。继空间光学、空间雷达和基于无人机的光学遥感技术之后，无人机-合成孔径雷达作为收集高分辨率农业信息的第四种实用手段具有巨大潜力。

{"title":"Classification of maize lodging types using UAV-SAR remote sensing data and machine learning methods","authors":"Dashuai Wang , Minghu Zhao , Zhuolin Li , Xiaohu Wu , Nan Li , Decheng Li , Sheng Xu , Xiaoguang Liu","doi":"10.1016/j.compag.2024.109637","DOIUrl":"10.1016/j.compag.2024.109637","url":null,"abstract":"<div><div>Lodging seriously threatens maize quality and yield and inevitably increases management and harvest costs. Timely collection of crop lodging information plays a pivotal role in the post-disaster assessment and agricultural insurance claims. Although spaceborne radar and optical remote sensing have unparalleled advantages in obtaining large-scale agricultural information, their response capacity to sudden natural maize lodging disasters is insufficient due to the limited spatial–temporal resolution of the satellite data. In recent years, the widespread application of unmanned aerial vehicles (UAVs) based optical remote sensing in precision agriculture has provided an effective alternative to spaceborne remote sensing. However, optical sensing can only effectively reveal the reflectance spectral characteristics of lodging maize under good lighting conditions. This work proposes a novel maize lodging classification method based on UAV synthetic aperture radar (UAV-SAR) and machine learning to circumvent the limitations of spaceborne and UAV-based remote sensing in monitoring maize lodging. Firstly, the raw radar remote sensing data of our study area containing lodging and non-lodging maize plants at the maturity stage is collected by the custom-built X-band and Ku-band UAV-SAR systems. Secondly, the corresponding backscattering coefficients and radar vegetation indices in each lodging type are extracted through radiation calibration and band math. Subsequently, the impacts of radar parameters (bands, polarizations, and observation orientations) and lodging types on backscattering coefficients are comprehensively analyzed. Fourthly, we applied the recursive feature elimination (RFE) algorithm to identify significant feature subsets and constructed multiple datasets using ten filter scales. Finally, five machine learning models (XGBoost, LDA, RF, KNN, and ANN) are trained and tested based on these materials. The classification results under different filter scales and feature combinations show that ANN achieves the best performance with an overall accuracy of 98.26 % and a Kappa coefficient of 0.982. This is the first innovative study successfully introducing cutting-edge UAV-SAR into maize lodging monitoring. Following spaceborne optical, spaceborne radar, and UAV-based optical remote sensing technologies, UAV-SAR holds great potential as the fourth practical means for collecting high-resolution agricultural information.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109637"},"PeriodicalIF":7.7,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142699944","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Aquatic plants detection in crab ponds using UAV hyperspectral imagery combined with transformer-based semantic segmentation model 利用无人机高光谱图像结合基于变换器的语义分割模型检测蟹塘中的水生植物

IF 7.7 1区农林科学 Q1 AGRICULTURE, MULTIDISCIPLINARY

Computers and Electronics in Agriculture

Pub Date : 2024-11-18 DOI: 10.1016/j.compag.2024.109656

Zijian Yu , Tingyu Xie , Qibing Zhu , Peiyu Dai , Xing Mao , Ni Ren , Xin Zhao , Xinnian Guo

Aquatic plants provide habitat and food for Chinese mitten crab growth, the identification of aquatic plant species and monitoring of their coverage can provide basic information for the management of aquatic plants, which can help to improve the efficiency of aquaculture. In this study, to address the time-consuming and labour-intensive nature of traditional aquatic plant monitoring in crab ponds relying on manual observation, a classification method for aquatic plant species using unmanned aerial vehicle and hyperspectral imagery (UAV-HSI) technology, combined with an improved semantic segmentation model named SpectralUFormer was reported for the first time. The UAV-HSI data provides a high-quality data source for automatic aquatic plants detection, and the proposed SpectralUFormer integrates hybrid attention block and hybrid cascaded upsampler. Specifically, the hybrid attention block aggregates abundant spectral features in the encoder. In the decoder part, the hybrid cascaded upsampler is designed by incorporating PixelShuffle and G-L MLP block, which together perform the importance calculation and alignment of feature weights. Experimental results show that the SpectralUFormer achieves high-precision classification of aquatic plant species, with an overall accuracy of 93.15% and a Kappa coefficient of 89.14%. This study offers a feasible approach for the automatic identification of aquatic plant species in crab ponds and the estimation of their coverage.

水生植物为中华绒螯蟹的生长提供了栖息地和食物，对水生植物种类的鉴定和覆盖范围的监测可为水生植物的管理提供基础信息，有助于提高养殖效率。本研究针对传统蟹塘水生植物监测依赖人工观测耗时耗力的特点，首次报道了利用无人机和高光谱成像（UAV-HSI）技术，结合改进的语义分割模型 SpectralUFormer，对水生植物种类进行分类的方法。无人机-高光谱成像数据为水生植物的自动检测提供了高质量的数据源，而所提出的 SpectralUFormer 则集成了混合注意力块和混合级联上采样器。具体来说，混合注意力块在编码器中汇聚了丰富的光谱特征。在解码器部分，混合级联上采样器的设计结合了 PixelShuffle 和 G-L MLP 模块，共同完成重要度计算和特征权重的校准。实验结果表明，SpectralUFormer 实现了高精度的水生植物物种分类，总体准确率为 93.15%，Kappa 系数为 89.14%。这项研究为自动识别蟹塘中的水生植物物种和估计其覆盖范围提供了一种可行的方法。

{"title":"Aquatic plants detection in crab ponds using UAV hyperspectral imagery combined with transformer-based semantic segmentation model","authors":"Zijian Yu , Tingyu Xie , Qibing Zhu , Peiyu Dai , Xing Mao , Ni Ren , Xin Zhao , Xinnian Guo","doi":"10.1016/j.compag.2024.109656","DOIUrl":"10.1016/j.compag.2024.109656","url":null,"abstract":"<div><div>Aquatic plants provide habitat and food for Chinese mitten crab growth, the identification of aquatic plant species and monitoring of their coverage can provide basic information for the management of aquatic plants, which can help to improve the efficiency of aquaculture. In this study, to address the time-consuming and labour-intensive nature of traditional aquatic plant monitoring in crab ponds relying on manual observation, a classification method for aquatic plant species using unmanned aerial vehicle and hyperspectral imagery (UAV-HSI) technology, combined with an improved semantic segmentation model named SpectralUFormer was reported for the first time. The UAV-HSI data provides a high-quality data source for automatic aquatic plants detection, and the proposed SpectralUFormer integrates hybrid attention block and hybrid cascaded upsampler. Specifically, the hybrid attention block aggregates abundant spectral features in the encoder. In the decoder part, the hybrid cascaded upsampler is designed by incorporating PixelShuffle and G-L MLP block, which together perform the importance calculation and alignment of feature weights. Experimental results show that the SpectralUFormer achieves high-precision classification of aquatic plant species, with an overall accuracy of 93.15% and a Kappa coefficient of 89.14%. This study offers a feasible approach for the automatic identification of aquatic plant species in crab ponds and the estimation of their coverage.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109656"},"PeriodicalIF":7.7,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142699947","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Construction and validation of a mathematical model for the pressure subsidence of mixed crop straw in Shajiang black soil 沙江黑土中混合作物秸秆压力沉降数学模型的构建与验证

IF 7.7 1区农林科学 Q1 AGRICULTURE, MULTIDISCIPLINARY

Computers and Electronics in Agriculture

Pub Date : 2024-11-18 DOI: 10.1016/j.compag.2024.109649

Dongbo Xie , Zhiqiang Li , Ce Liu , Gang Zhao , Liqing Chen

The soil properties of mixed crop straw do not enable conventional pressure subsidence models to characterize the relationship between straw amount and pressure-bearing properties accurately. Based on the distribution of straw in the field, this study explored the effect of the amount of surface straw cover on the pressure subsidence relationship in Shajiang black soil. The quadratic rotated orthogonal combination test was used to quantify the mathematical relationships of Shajiang black soil pressure subsidence modeling with the amount of surface straw cover (SSC) and mass mixing ratio of soil to straw (MSS). Then, using the weighted least squares method, the pressure subsidence parameters (cohesive deformation modulus, friction deformation modulus, and subsidence index) were obtained, and the Bekker model was modified to construct a pressure subsidence model for the straw-containing soil. Finally, the modified model was verified under conditions of a water content of 18 %, the SSC of 2.5 t·ha⁻¹, and the MSS of 2.5 %. Results showed that the proposed pressure subsidence model predicted the value with a relative error of 2.21 % compared with the experimental measurements. The model’s predicted value accuracy improved by 10.65 % compared to the conventional model. From these results, this study proposes that a mixed crop straw Shajiang black soil pressure subsidence model can predict the soil’s internal stress transfer and stress–strain conditions.

混合作物秸秆的土壤特性使得传统的压力沉降模型无法准确表征秸秆量与承压特性之间的关系。本研究根据田间秸秆的分布情况，探讨了地表秸秆覆盖量对沙江黑土压力沉降关系的影响。通过二次旋转正交组合检验，量化了沙江黑土压力下沉模型与地表秸秆覆盖量（SSC）和土壤与秸秆质量混合比（MSS）的数学关系。然后，利用加权最小二乘法得到了压力下沉参数（粘聚变形模量、摩擦变形模量和下沉指数），并对 Bekker 模型进行了修正，构建了含秸秆土壤的压力下沉模型。最后，在含水量为 18%、SSC 为 2.5 吨-公顷-1 和 MSS 为 2.5% 的条件下对修改后的模型进行了验证。结果表明，与实验测量结果相比，建议的压力沉降模型预测值的相对误差为 2.21%。与传统模型相比，该模型的预测值精度提高了 10.65%。从这些结果来看，本研究提出的混合作物秸秆沙江黑土压力沉降模型可以预测土壤的内应力传递和应力应变状况。

{"title":"Construction and validation of a mathematical model for the pressure subsidence of mixed crop straw in Shajiang black soil","authors":"Dongbo Xie , Zhiqiang Li , Ce Liu , Gang Zhao , Liqing Chen","doi":"10.1016/j.compag.2024.109649","DOIUrl":"10.1016/j.compag.2024.109649","url":null,"abstract":"<div><div>The soil properties of mixed crop straw do not enable conventional pressure subsidence models to characterize the relationship between straw amount and pressure-bearing properties accurately. Based on the distribution of straw in the field, this study explored the effect of the amount of surface straw cover on the pressure subsidence relationship in Shajiang black soil. The quadratic rotated orthogonal combination test was used to quantify the mathematical relationships of Shajiang black soil pressure subsidence modeling with the amount of surface straw cover (SSC) and mass mixing ratio of soil to straw (MSS). Then, using the weighted least squares method, the pressure subsidence parameters (cohesive deformation modulus, friction deformation modulus, and subsidence index) were obtained, and the Bekker model was modified to construct a pressure subsidence model for the straw-containing soil. Finally, the modified model was verified under conditions of a water content of 18 %, the SSC of 2.5 t·ha<sup>−1</sup>, and the MSS of 2.5 %. Results showed that the proposed pressure subsidence model predicted the value with a relative error of 2.21 % compared with the experimental measurements. The model’s predicted value accuracy improved by 10.65 % compared to the conventional model. From these results, this study proposes that a mixed crop straw Shajiang black soil pressure subsidence model can predict the soil’s internal stress transfer and stress–strain conditions.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109649"},"PeriodicalIF":7.7,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142661677","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Canopy structure dynamics constraints and time sequence alignment for improving retrieval of rice leaf area index from multi-temporal Sentinel-1 imagery 利用冠层结构动力学约束和时序配准改进多时相 Sentinel-1 图像的水稻叶面积指数检索

IF 7.7 1区农林科学 Q1 AGRICULTURE, MULTIDISCIPLINARY

Computers and Electronics in Agriculture

Pub Date : 2024-11-18 DOI: 10.1016/j.compag.2024.109658

Yu Liu , Bo Wang , Junfeng Tao , Sijing Tian , Qinghong Sheng , Jun Li , Shuwei Wang , Xiaoli Liu , Honglin He

Due to the limited availability of in-situ observation data, most existing leaf area index (LAI) inversion models do not fully leverage temporal information. Furthermore, the phenological evolution of crops can result in unstable and inaccurate retrieval outcomes. To address these challenges, this study proposes a novel framework for LAI inversion based on Sentinel-1. First, the constrained canopy structure dynamic hierarchical linear model (CSDHLM) is constructed, which integrates canopy dynamics information and temporal constraints. Second, the microwave scattering characteristics at various crop growth stages used to develop the phenological segment dynamic time warping (PSDTW). The PSDTW aims to address the challenges posed by inconsistent phenological dynamics across different plots. The quantitative evaluation results indicate that CSDHLM more accurately captures the temporal changes of LAI (R² = 0.7688, RMSE = 0.8742) compared to hierarchical linear model (R² = 0.7234, RMSE = 0.9561) and gaussian process regression (R² = 0.7143, RMSE = 0.9717). Additionally, the LAI inversion results obtained by combining CSDHLM and PSDTW have greater robustness (R² = 0.7332, RMSE = 1.4032) across diverse agricultural scenarios. This study emphasizes the importance of phenological information in estimating rice LAI, and the proposed framework is capable of generating long-term rice LAI maps with high resolution, demonstrating significant potential for agricultural applications at the regional scale.

由于现场观测数据有限，现有的叶面积指数（LAI）反演模型大多不能充分利用时间信息。此外，作物的物候演变也会导致检索结果不稳定、不准确。为了应对这些挑战，本研究提出了一种基于 Sentinel-1 的新型 LAI 反演框架。首先，构建了约束冠层结构动态分层线性模型（CSDHLM），该模型整合了冠层动态信息和时间约束。其次，利用作物不同生长阶段的微波散射特征，建立物候区段动态时间扭曲（PSDTW）。PSDTW 旨在解决不同地块物候动态不一致带来的挑战。定量评估结果表明，与分层线性模型（R2 = 0.7234，RMSE = 0.9561）和高斯过程回归（R2 = 0.7143，RMSE = 0.9717）相比，CSDHLM 能更准确地捕捉 LAI 的时间变化（R2 = 0.7688，RMSE = 0.8742）。此外，结合 CSDHLM 和 PSDTW 得出的 LAI 反演结果在不同农业情景下具有更强的鲁棒性（R2 = 0.7332，RMSE = 1.4032）。本研究强调了物候信息在估算水稻LAI中的重要性，所提出的框架能够生成高分辨率的长期水稻LAI图，在区域尺度的农业应用中具有巨大潜力。

{"title":"Canopy structure dynamics constraints and time sequence alignment for improving retrieval of rice leaf area index from multi-temporal Sentinel-1 imagery","authors":"Yu Liu , Bo Wang , Junfeng Tao , Sijing Tian , Qinghong Sheng , Jun Li , Shuwei Wang , Xiaoli Liu , Honglin He","doi":"10.1016/j.compag.2024.109658","DOIUrl":"10.1016/j.compag.2024.109658","url":null,"abstract":"<div><div>Due to the limited availability of in-situ observation data, most existing leaf area index (LAI) inversion models do not fully leverage temporal information. Furthermore, the phenological evolution of crops can result in unstable and inaccurate retrieval outcomes. To address these challenges, this study proposes a novel framework for LAI inversion based on Sentinel-1. First, the constrained canopy structure dynamic hierarchical linear model (CSDHLM) is constructed, which integrates canopy dynamics information and temporal constraints. Second, the microwave scattering characteristics at various crop growth stages used to develop the phenological segment dynamic time warping (PSDTW). The PSDTW aims to address the challenges posed by inconsistent phenological dynamics across different plots. The quantitative evaluation results indicate that CSDHLM more accurately captures the temporal changes of LAI (R<sup>2</sup> = 0.7688, RMSE = 0.8742) compared to hierarchical linear model (R<sup>2</sup> = 0.7234, RMSE = 0.9561) and gaussian process regression (R<sup>2</sup> = 0.7143, RMSE = 0.9717). Additionally, the LAI inversion results obtained by combining CSDHLM and PSDTW have greater robustness (R<sup>2</sup> = 0.7332, RMSE = 1.4032) across diverse agricultural scenarios. This study emphasizes the importance of phenological information in estimating rice LAI, and the proposed framework is capable of generating long-term rice LAI maps with high resolution, demonstrating significant potential for agricultural applications at the regional scale.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109658"},"PeriodicalIF":7.7,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142700044","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Estimation of crop leaf area index based on Sentinel-2 images and PROSAIL-Transformer coupling model 基于哨兵-2 图像和 PROSAIL-Transformer 耦合模型的作物叶面积指数估算

IF 7.7 1区农林科学 Q1 AGRICULTURE, MULTIDISCIPLINARY

Computers and Electronics in Agriculture

Pub Date : 2024-11-17 DOI: 10.1016/j.compag.2024.109663

Tianjiao Liu , Si-Bo Duan , Niantang Liu , Baoan Wei , Juntao Yang , Jiankui Chen , Li Zhang

Accurate estimation of leaf area index (LAI) is hindered by challenges in capturing crop-specific spectral variability and integrating complex model-data relationships. To address these issues, this study proposes a novel framework based on Sentinel-2 images, coupling the PROSAIL physical model with a Transformer-based deep learning model. This framework incorporates three key features contributing to its effectiveness. Firstly, Sentinel-2 reflectance was generated using the PROSAIL model and refined through sample matching to ensure optimal alignment with Sentinel-2 imagery specific to each crop type. Secondly, the Maximum Information Coefficient (MIC) and Recursive Feature Elimination (RFE) were employed to identify the most relevant spectral feature combinations for different crop categories. Thirdly, a PROSAIL-Transformer coupling model was constructed based on selected feature combinations to generate accurate Sentinel-2 LAI products. To validate the proposed approach, field crop LAI measurements were collected at five plots within the study area. Quantitative assessments demonstrate a coefficient of determination (R²) of 0.87, root mean square error (RMSE) of 0.48, and mean absolute error (MAE) of 0.36. The proposed framework enables the production of time-series LAI maps at fine resolution, facilitating dynamic crop monitoring and management in areas of high spatial heterogeneity.

叶面积指数（LAI）的精确估算受到捕捉作物特定光谱变异性和整合复杂的模型-数据关系等挑战的阻碍。为解决这些问题，本研究提出了一种基于哨兵-2 图像的新型框架，将 PROSAIL 物理模型与基于 Transformer 的深度学习模型相结合。该框架包含三个有助于提高其有效性的关键特征。首先，使用 PROSAIL 模型生成哨兵-2 反射率，并通过样本匹配进行改进，以确保与哨兵-2 图像针对每种作物类型进行最佳匹配。其次，利用最大信息系数（MIC）和递归特征消除（RFE）来确定与不同作物类别最相关的光谱特征组合。第三，根据选定的特征组合构建 PROSAIL-Transformer 耦合模型，生成准确的 Sentinel-2 LAI 产品。为了验证所提出的方法，在研究区域内的五个地块收集了田间作物 LAI 测量数据。定量评估表明，确定系数 (R2) 为 0.87，均方根误差 (RMSE) 为 0.48，平均绝对误差 (MAE) 为 0.36。所提出的框架能够绘制精细分辨率的时间序列 LAI 地图，有助于在空间异质性较高的地区对作物进行动态监测和管理。

{"title":"Estimation of crop leaf area index based on Sentinel-2 images and PROSAIL-Transformer coupling model","authors":"Tianjiao Liu , Si-Bo Duan , Niantang Liu , Baoan Wei , Juntao Yang , Jiankui Chen , Li Zhang","doi":"10.1016/j.compag.2024.109663","DOIUrl":"10.1016/j.compag.2024.109663","url":null,"abstract":"<div><div>Accurate estimation of leaf area index (LAI) is hindered by challenges in capturing crop-specific spectral variability and integrating complex model-data relationships. To address these issues, this study proposes a novel framework based on Sentinel-2 images, coupling the PROSAIL physical model with a Transformer-based deep learning model. This framework incorporates three key features contributing to its effectiveness. Firstly, Sentinel-2 reflectance was generated using the PROSAIL model and refined through sample matching to ensure optimal alignment with Sentinel-2 imagery specific to each crop type. Secondly, the Maximum Information Coefficient (MIC) and Recursive Feature Elimination (RFE) were employed to identify the most relevant spectral feature combinations for different crop categories. Thirdly, a PROSAIL-Transformer coupling model was constructed based on selected feature combinations to generate accurate Sentinel-2 LAI products. To validate the proposed approach, field crop LAI measurements were collected at five plots within the study area. Quantitative assessments demonstrate a coefficient of determination (R<sup>2</sup>) of 0.87, root mean square error (RMSE) of 0.48, and mean absolute error (MAE) of 0.36. The proposed framework enables the production of time-series LAI maps at fine resolution, facilitating dynamic crop monitoring and management in areas of high spatial heterogeneity.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109663"},"PeriodicalIF":7.7,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142661679","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Fish feeding behavior recognition using time-domain and frequency-domain signals fusion from six-axis inertial sensors 利用六轴惯性传感器的时域和频域信号融合识别鱼类摄食行为

IF 7.7 1区农林科学 Q1 AGRICULTURE, MULTIDISCIPLINARY

Computers and Electronics in Agriculture

Pub Date : 2024-11-17 DOI: 10.1016/j.compag.2024.109652

Pingchuan Ma , Xinting Yang , Weichen Hu , Tingting Fu , Chao Zhou

In aquaculture, real-time recognition of fish feeding activities is important for enhancing feed conversion rate and reducing production costs. Therefore, this study uses a six-axis inertial sensor to collect water surface fluctuation caused by fish feeding, and proposes a time-domain and frequency-domain fusion model (TFFormer) for fish feeding behavior recognition, and identifies the feeding intensity of fish as four categories: Strong, Medium, Weak, and None. The implementation details are as follows: Firstly, the data collected by the six-axis inertial sensor is preprocessed using a sliding window to obtain time series data, and perform Fourier transform on it to obtain the frequency domain sequence. Then, the transformer is used to unify the time domain and frequency domain features respectively. A Mutual Promotion Unit (MPU) is established based on cross self-attention and a feedforward neural network (FFN). By integrating with a Global multimodal fusion (G) module, MPU establishes a global–local interactive learning framework to extract features from temporal and frequency domains, resulting in temporal-frequency interaction features. Finally, the introduction of supervised contrastive loss function supervises the training process, enhancing the accuracy of fish school feeding intensity classification. Experimental results demonstrate that the proposed TFFormer model effectively processes both temporal and frequency signals, achieving an accuracy of 91.52%, a 5.56% improvement over the baseline model and provides technical support for the development of intelligent feeding machines.

在水产养殖中，鱼类摄食活动的实时识别对于提高饲料转化率和降低生产成本非常重要。因此，本研究利用六轴惯性传感器采集鱼类摄食引起的水面波动，并提出了一种用于识别鱼类摄食行为的时域和频域融合模型（TFFormer），并将鱼类的摄食强度识别为四个类别：鱼类摄食强度分为四类：强、中、弱和无。具体实现过程如下：首先，利用滑动窗口对六轴惯性传感器采集的数据进行预处理，得到时间序列数据，并对其进行傅里叶变换，得到频域序列。然后，利用变换器分别统一时域和频域特征。在交叉自注意和前馈神经网络（FFN）的基础上建立互促单元（MPU）。通过与全局多模态融合（G）模块整合，MPU 建立了一个全局-局部交互式学习框架，从时域和频域提取特征，形成时频交互特征。最后，引入监督对比损失函数对训练过程进行监督，提高了鱼群摄食强度分类的准确性。实验结果表明，所提出的 TFFormer 模型能有效处理时域和频域信号，准确率达到 91.52%，比基线模型提高了 5.56%，为开发智能饲喂机提供了技术支持。

{"title":"Fish feeding behavior recognition using time-domain and frequency-domain signals fusion from six-axis inertial sensors","authors":"Pingchuan Ma , Xinting Yang , Weichen Hu , Tingting Fu , Chao Zhou","doi":"10.1016/j.compag.2024.109652","DOIUrl":"10.1016/j.compag.2024.109652","url":null,"abstract":"<div><div>In aquaculture, real-time recognition of fish feeding activities is important for enhancing feed conversion rate and reducing production costs. Therefore, this study uses a six-axis inertial sensor to collect water surface fluctuation caused by fish feeding, and proposes a time-domain and frequency-domain fusion model (TFFormer) for fish feeding behavior recognition, and identifies the feeding intensity of fish as four categories: Strong, Medium, Weak, and None. The implementation details are as follows: Firstly, the data collected by the six-axis inertial sensor is preprocessed using a sliding window to obtain time series data, and perform Fourier transform on it to obtain the frequency domain sequence. Then, the transformer is used to unify the time domain and frequency domain features respectively. A Mutual Promotion Unit (MPU) is established based on cross self-attention and a feedforward neural network (FFN). By integrating with a Global multimodal fusion (G) module, MPU establishes a global–local interactive learning framework to extract features from temporal and frequency domains, resulting in temporal-frequency interaction features. Finally, the introduction of supervised contrastive loss function supervises the training process, enhancing the accuracy of fish school feeding intensity classification. Experimental results demonstrate that the proposed TFFormer model effectively processes both temporal and frequency signals, achieving an accuracy of 91.52%, a 5.56% improvement over the baseline model and provides technical support for the development of intelligent feeding machines.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109652"},"PeriodicalIF":7.7,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142661678","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Design, integration, and field evaluation of a selective harvesting robot for broccoli 西兰花选择性收获机器人的设计、集成和实地评估

IF 7.7 1区农林科学 Q1 AGRICULTURE, MULTIDISCIPLINARY

Computers and Electronics in Agriculture

Pub Date : 2024-11-16 DOI: 10.1016/j.compag.2024.109654

Shuo Kang , Sifang Long , Dongfang Li , Jiali Fan , Dongdong Du , Jun Wang

The agronomic characteristics of broccoli necessitate selective harvesting in multiple batches, highlighting an urgent need for a selective harvesting robot to alleviate labour constraints. However, current research has inadequately addressed the problems of maturity identification of broccoli heads, fast and safe movement of the manipulator, and efficient and stable end-effector. Therefore, we proposed a semantic segmentation network called Broccoli Segmentation (BroSeg) for the mature identification and localisation of broccoli. BroSeg incorporated a lightweight backbone network, attention mechanisms, densely connected atrous spatial pyramid pooling, and a post-processing module. BroSeg achieved a Mean Intersection over Union (mIoU) of 58.92 % and a mean category prediction accuracy of 81.63 %. Using a collaborative simulation based on the Robot Operating System (ROS) and conducting comparative experiments, we selected the Batch Informed Trees (BIT*) algorithm that was most suitable for broccoli harvesting tasks. The effectiveness of the proposed method was validated through collaborative simulation and field experiments. Based on morphological analysis and cutting experiments, we designed an integrated gripper-cutting end-effector that mimics human hand-pinching for broccoli harvesting. The success rate of field harvesting reaches 86.96 %. This research integrates the functionalities of perception, manipulation, and cognition to construct a broccoli selective harvesting robot. Field experiments demonstrate a selective harvesting success rate of 63.16 %, with an average time of 11.9 s, validating the effectiveness and potential of the system.

由于西兰花的农艺特性，有必要进行多批次选择性收获，因此迫切需要一种选择性收获机器人来缓解劳动力限制。然而，目前的研究尚未充分解决西兰花头部的成熟度识别、机械手的快速安全移动以及高效稳定的末端执行器等问题。因此，我们提出了一种名为 "西兰花分割（Broccoli Segmentation，BroSeg）"的语义分割网络，用于西兰花的成熟识别和定位。BroSeg 包含一个轻量级骨干网络、注意力机制、密集连接的无齿空间金字塔池和一个后处理模块。BroSeg 的平均联合交叉率（mIoU）为 58.92%，平均类别预测准确率为 81.63%。通过基于机器人操作系统（ROS）的协作模拟和对比实验，我们选择了最适合西兰花收获任务的批量信息树（BIT*）算法。通过协同模拟和现场实验，验证了所提方法的有效性。在形态分析和切割实验的基础上，我们设计了一种集成式机械手切割末端执行器，可模仿人手夹持西兰花收割。田间收割的成功率达到 86.96%。这项研究集成了感知、操纵和认知功能，构建了一种西兰花选择性收获机器人。现场实验表明，选择性收获的成功率为 63.16%，平均时间为 11.9 秒，验证了该系统的有效性和潜力。

{"title":"Design, integration, and field evaluation of a selective harvesting robot for broccoli","authors":"Shuo Kang , Sifang Long , Dongfang Li , Jiali Fan , Dongdong Du , Jun Wang","doi":"10.1016/j.compag.2024.109654","DOIUrl":"10.1016/j.compag.2024.109654","url":null,"abstract":"<div><div>The agronomic characteristics of broccoli necessitate selective harvesting in multiple batches, highlighting an urgent need for a selective harvesting robot to alleviate labour constraints. However, current research has inadequately addressed the problems of maturity identification of broccoli heads, fast and safe movement of the manipulator, and efficient and stable end-effector. Therefore, we proposed a semantic segmentation network called Broccoli Segmentation (BroSeg) for the mature identification and localisation of broccoli. BroSeg incorporated a lightweight backbone network, attention mechanisms, densely connected atrous spatial pyramid pooling, and a post-processing module. BroSeg achieved a Mean Intersection over Union (mIoU) of 58.92 % and a mean category prediction accuracy of 81.63 %. Using a collaborative simulation based on the Robot Operating System (ROS) and conducting comparative experiments, we selected the Batch Informed Trees (BIT*) algorithm that was most suitable for broccoli harvesting tasks. The effectiveness of the proposed method was validated through collaborative simulation and field experiments. Based on morphological analysis and cutting experiments, we designed an integrated gripper-cutting end-effector that mimics human hand-pinching for broccoli harvesting. The success rate of field harvesting reaches 86.96 %. This research integrates the functionalities of perception, manipulation, and cognition to construct a broccoli selective harvesting robot. Field experiments demonstrate a selective harvesting success rate of 63.16 %, with an average time of 11.9 s, validating the effectiveness and potential of the system.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109654"},"PeriodicalIF":7.7,"publicationDate":"2024-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142661675","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A Novel Behavior Detection Method for Sows and Piglets during Lactation Based on an Inspection Robot 基于检测机器人的哺乳期母猪和仔猪行为检测新方法

IF 7.7 1区农林科学 Q1 AGRICULTURE, MULTIDISCIPLINARY

Computers and Electronics in Agriculture

Pub Date : 2024-11-16 DOI: 10.1016/j.compag.2024.109613

Jie Zhou , Luo Liu , Tao Jiang , Haonan Tian , Mingxia Shen , Longshen Liu

Accurately identifying behaviors exhibited by lactating sows and piglets is crucial for maintaining swine health and preventing farming crises. In the absence of dedicated swine behavior monitoring systems and the challenges of implementing cloud-based automated monitoring in large-scale farming, this study proposes a method utilizing inspection robots to detect behaviors of lactating sows and piglets. The inspection robot initially serves as a data acquisition and storage tool, collecting behavioral data such as sows postures (standing, sitting, lateral recumbency, and sternal recumbency) and activities of piglet groups (resting, suckling, and active behavior) within confined pens. The YOLOv8 series algorithms are then employed to identify static postures of sows, while the Temporal Shift Module (TSM) is used to recognize dynamic behaviors within piglet groups. These models are fine-tuned and deployed on the Jetson Nano edge computing platform. Experimental results show that YOLOv8n accurately identifies sow postures with a mean Average Precision (mAP) @0.5 of 97.08% and a frame rate of 36.4 FPS at an image resolution of 480 × 288, following TensorRT acceleration. For piglet behavior recognition, the TSM model, using ResNet50 as the backbone network, achieves a Top-1 accuracy of 93.63% in recognizing piglet behaviors. Replacing ResNet50 with MobileNetv2 slightly reduces the Top-1 accuracy to 90.81%; however, there is a significant improvement in inference speed on Jetson Nano for a single video clip with a processing duration of 542.51 ms, representing more than a 20-fold enhancement compared to TSM_ResNet50. The Kappa consistency analysis reveals moderate behavioral coherence among sows in different pens and piglet groups. The study offers insights into automated detection of behaviors lactating sows and piglets within large-scale intensive farming systems.

准确识别哺乳母猪和仔猪的行为对维护猪群健康和预防养殖危机至关重要。由于缺乏专用的猪群行为监测系统，且在大规模养殖中实施基于云的自动监测存在挑战，本研究提出了一种利用检测机器人检测哺乳母猪和仔猪行为的方法。检测机器人最初作为数据采集和存储工具，收集行为数据，如母猪在密闭猪圈内的姿势（站立、坐立、侧卧、胸卧）和仔猪群体的活动（休息、吸吮和活动行为）。然后，利用 YOLOv8 系列算法识别母猪的静态姿势，同时利用时移模块 (TSM) 识别仔猪群的动态行为。这些模型经过微调后部署在 Jetson Nano 边缘计算平台上。实验结果表明，YOLOv8n 能准确识别母猪姿态，平均精度 (mAP) @0.5 为 97.08%，在图像分辨率为 480 × 288 的情况下，经过 TensorRT 加速后的帧速率为 36.4 FPS。在仔猪行为识别方面，使用 ResNet50 作为骨干网络的 TSM 模型在识别仔猪行为方面达到了 93.63% 的 Top-1 准确率。用 MobileNetv2 代替 ResNet50 后，Top-1 准确率略有下降，为 90.81%；但是，在 Jetson Nano 上对处理时长为 542.51 毫秒的单个视频片段的推理速度有了显著提高，与 TSM_ResNet50 相比提高了 20 倍以上。Kappa 一致性分析表明，不同猪栏和仔猪组的母猪行为具有适度的一致性。这项研究为在大规模集约化养殖系统中自动检测哺乳母猪和仔猪的行为提供了启示。

{"title":"A Novel Behavior Detection Method for Sows and Piglets during Lactation Based on an Inspection Robot","authors":"Jie Zhou , Luo Liu , Tao Jiang , Haonan Tian , Mingxia Shen , Longshen Liu","doi":"10.1016/j.compag.2024.109613","DOIUrl":"10.1016/j.compag.2024.109613","url":null,"abstract":"<div><div>Accurately identifying behaviors exhibited by lactating sows and piglets is crucial for maintaining swine health and preventing farming crises. In the absence of dedicated swine behavior monitoring systems and the challenges of implementing cloud-based automated monitoring in large-scale farming, this study proposes a method utilizing inspection robots to detect behaviors of lactating sows and piglets. The inspection robot initially serves as a data acquisition and storage tool, collecting behavioral data such as sows postures (standing, sitting, lateral recumbency, and sternal recumbency) and activities of piglet groups (resting, suckling, and active behavior) within confined pens. The YOLOv8 series algorithms are then employed to identify static postures of sows, while the Temporal Shift Module (TSM) is used to recognize dynamic behaviors within piglet groups. These models are fine-tuned and deployed on the Jetson Nano edge computing platform. Experimental results show that YOLOv8n accurately identifies sow postures with a mean Average Precision (mAP) @0.5 of 97.08% and a frame rate of 36.4 FPS at an image resolution of 480 × 288, following TensorRT acceleration. For piglet behavior recognition, the TSM model, using ResNet50 as the backbone network, achieves a Top-1 accuracy of 93.63% in recognizing piglet behaviors. Replacing ResNet50 with MobileNetv2 slightly reduces the Top-1 accuracy to 90.81%; however, there is a significant improvement in inference speed on Jetson Nano for a single video clip with a processing duration of 542.51 ms, representing more than a 20-fold enhancement compared to TSM_ResNet50. The Kappa consistency analysis reveals moderate behavioral coherence among sows in different pens and piglet groups. The study offers insights into automated detection of behaviors lactating sows and piglets within large-scale intensive farming systems.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109613"},"PeriodicalIF":7.7,"publicationDate":"2024-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142661727","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0