Pub Date : 2024-12-15DOI: 10.1016/j.atech.2024.100719
Fengkai Tian , Jianfeng Zhou , Curtis J. Ransom , Noel Aloysius , Kenneth A. Sudduth
Chlorophyll is crucial for photosynthesis and impacts plant growth and yield in crops. Accurate estimation of plant health and fertilizer status is essential for effective nitrogen (N) management in corn. However, crop chlorophyll is primarily quantified using handheld sensors, which is time-consuming, labor-intensive, and of low spatial resolution. This study aimed to evaluate an airborne multispectral imaging system in estimating the chlorophyll content of corn leaves at four vegetative growth stages. Three replicates of 12 nitrogen rates (between 0 and 285 kg ha−1) were applied to corn at the V4 vegetative stage (i.e., with four established leaves). Soil apparent electrical conductivity (ECa) of all test plots was measured before planting and corn leaf chlorophyll content was measured using a commercial handheld chlorophyll meter at four vegetative stages (V8, V9, V11, and V12). A UAV-based multispectral camera collected imagery at the same time as manual readings. Machine learning models developed based on image features derived from UAV images were used to predict leaf chlorophyll content. Results showed that an epsilon support vector regression model built on imagery data across imagery data collected over four growth stages with a sequential forward feature selection achieved the best performance (R² = 0.87, MAE = 1.80, and RMSE = 2.26 SPAD units). There was no significant difference in the performance of models across the four growth stages. By utilizing the developed model, researchers and growers can effectively map the chlorophyll content of corn leaves at different growth stages, enabling them to make timely and informed management decisions.
{"title":"Estimating corn leaf chlorophyll content using airborne multispectral imagery and machine learning","authors":"Fengkai Tian , Jianfeng Zhou , Curtis J. Ransom , Noel Aloysius , Kenneth A. Sudduth","doi":"10.1016/j.atech.2024.100719","DOIUrl":"10.1016/j.atech.2024.100719","url":null,"abstract":"<div><div>Chlorophyll is crucial for photosynthesis and impacts plant growth and yield in crops. Accurate estimation of plant health and fertilizer status is essential for effective nitrogen (N) management in corn. However, crop chlorophyll is primarily quantified using handheld sensors, which is time-consuming, labor-intensive, and of low spatial resolution. This study aimed to evaluate an airborne multispectral imaging system in estimating the chlorophyll content of corn leaves at four vegetative growth stages. Three replicates of 12 nitrogen rates (between 0 and 285 kg ha<sup>−1</sup>) were applied to corn at the V4 vegetative stage (i.e., with four established leaves). Soil apparent electrical conductivity (EC<sub>a</sub>) of all test plots was measured before planting and corn leaf chlorophyll content was measured using a commercial handheld chlorophyll meter at four vegetative stages (V8, V9, V11, and V12). A UAV-based multispectral camera collected imagery at the same time as manual readings. Machine learning models developed based on image features derived from UAV images were used to predict leaf chlorophyll content. Results showed that an epsilon support vector regression model built on imagery data across imagery data collected over four growth stages with a sequential forward feature selection achieved the best performance (R² = 0.87, MAE = 1.80, and RMSE = 2.26 SPAD units). There was no significant difference in the performance of models across the four growth stages. By utilizing the developed model, researchers and growers can effectively map the chlorophyll content of corn leaves at different growth stages, enabling them to make timely and informed management decisions.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100719"},"PeriodicalIF":6.3,"publicationDate":"2024-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143182102","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Vegetable grafting is a sustainable and eco-friendly practice for vegetable producers to manage plant diseases, boost plant nutrients and water absorption, and increase fruit yields and quality. To better promote the practice among U.S. vegetable producers and enhance the research and extension efforts regarding vegetable grafting, an online multistakeholder vegetable grafting economic decision support system (DSS) is developed. Users can rely on the DSS to conduct partial-budget, sensitivity, and breakeven analyses to comprehensively assess the economic implications of using vegetable grafting in various farm operations. The DSS also provides an essential platform for researchers, extension personnel, and producers to share information on the economics of vegetable grafting. It can serve as a valuable tool for stakeholders to analyze the economic feasibility of vegetable grafting and promote connections between various stakeholders, including producers, extension personnel, and researchers.
{"title":"Making grafting accessible to vegetable producers: An online multistakeholder economic decision-support system for vegetable grafting","authors":"Yefan Nian , Zhifeng Gao , Xin Zhao , Junhong Chen","doi":"10.1016/j.atech.2024.100723","DOIUrl":"10.1016/j.atech.2024.100723","url":null,"abstract":"<div><div>Vegetable grafting is a sustainable and eco-friendly practice for vegetable producers to manage plant diseases, boost plant nutrients and water absorption, and increase fruit yields and quality. To better promote the practice among U.S. vegetable producers and enhance the research and extension efforts regarding vegetable grafting, an online multistakeholder vegetable grafting economic decision support system (DSS) is developed. Users can rely on the DSS to conduct partial-budget, sensitivity, and breakeven analyses to comprehensively assess the economic implications of using vegetable grafting in various farm operations. The DSS also provides an essential platform for researchers, extension personnel, and producers to share information on the economics of vegetable grafting. It can serve as a valuable tool for stakeholders to analyze the economic feasibility of vegetable grafting and promote connections between various stakeholders, including producers, extension personnel, and researchers.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100723"},"PeriodicalIF":6.3,"publicationDate":"2024-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143182104","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-15DOI: 10.1016/j.atech.2024.100717
Alexander Steiger , Muhammad Qaswar , Ralf Bill , Abdul M. Mouazen , Görres Grenzdörffer
Understanding within-field variability is essential for optimizing crop yields, enhancing resource efficiency, and reducing environmental impacts in agriculture. Precision soil maps are crucial tools for achieving these goals. This study assessed the accuracy and practicality of the Stenon FarmLab, a handheld real-time soil sensor, by comparing its measurements to high-resolution soil maps generated using a Vis-NIR on-line multi-sensor platform. The experiment was conducted on a potato field in northeast Germany, with on-line sensing completed in August 2022 and Stenon sampling conducted monthly from November 2022 to September 2023. Calibration and validation of the Vis-NIR on-line multi-sensor data were performed using laboratory analyses and partial least squares regression. The R²-values varied across soil properties, ranging from 0.68 to 0.97 for calibration and 0.64 to 0.88 for validation, with mean R²-values of 0.81 and 0.72, respectively. Based on this, high-resolution soil property and nutrient maps were generated using ordinary kriging, effectively capturing spatial heterogeneity. The Stenon FarmLab showed significant variability in measurement accuracy and consistency. Temporal trends in mineralized nitrogen (Nmin), soil organic carbon (SOC), and pH were deemed implausible, as they could not be explained by natural soil processes or management practices. Correlation analyses between the two systems for stable soil properties resulted in R²-values of 0.29 for SOC, 0.41 for pH, and 0.50 for soil texture. While the Stenon FarmLab provides a rapid and convenient method for in-field soil analysis, this study revealed significant limitations in the accuracy of its measurements for most soil parameters.
{"title":"Comparing the handheld Stenon FarmLab soil sensor with a Vis-NIR multi-sensor soil sensing platform","authors":"Alexander Steiger , Muhammad Qaswar , Ralf Bill , Abdul M. Mouazen , Görres Grenzdörffer","doi":"10.1016/j.atech.2024.100717","DOIUrl":"10.1016/j.atech.2024.100717","url":null,"abstract":"<div><div>Understanding within-field variability is essential for optimizing crop yields, enhancing resource efficiency, and reducing environmental impacts in agriculture. Precision soil maps are crucial tools for achieving these goals. This study assessed the accuracy and practicality of the Stenon FarmLab, a handheld real-time soil sensor, by comparing its measurements to high-resolution soil maps generated using a Vis-NIR on-line multi-sensor platform. The experiment was conducted on a potato field in northeast Germany, with on-line sensing completed in August 2022 and Stenon sampling conducted monthly from November 2022 to September 2023. Calibration and validation of the Vis-NIR on-line multi-sensor data were performed using laboratory analyses and partial least squares regression. The R²-values varied across soil properties, ranging from 0.68 to 0.97 for calibration and 0.64 to 0.88 for validation, with mean R²-values of 0.81 and 0.72, respectively. Based on this, high-resolution soil property and nutrient maps were generated using ordinary kriging, effectively capturing spatial heterogeneity. The Stenon FarmLab showed significant variability in measurement accuracy and consistency. Temporal trends in mineralized nitrogen (N<sub>min</sub>), soil organic carbon (SOC), and pH were deemed implausible, as they could not be explained by natural soil processes or management practices. Correlation analyses between the two systems for stable soil properties resulted in R²-values of 0.29 for SOC, 0.41 for pH, and 0.50 for soil texture. While the Stenon FarmLab provides a rapid and convenient method for in-field soil analysis, this study revealed significant limitations in the accuracy of its measurements for most soil parameters.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100717"},"PeriodicalIF":6.3,"publicationDate":"2024-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143183123","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-13DOI: 10.1016/j.atech.2024.100710
Mireille van Hilten , Mark Ryan , Vincent Blok , Nina de Roo
As Artificial Intelligence (AI) continues to emerge in various sectors, ethical frameworks and guidelines aim to contribute to responsible AI development. While AI ethics has gained prominence in addressing broader societal concerns, existing regulations and guidelines often lack specificity for certain domain-specific applications (e.g. agri-food). AI is rapidly developed and deployed throughout the agri-food sector, but there is little practical guidance on how to do this responsibly. This study examines if the agri-food sector needs domain-specific guidance for the development and use of responsible AI and, if so, what it could look like. This research proposes it does and aims to fill this gap by introducing a novel approach for responsible AI in agri-food: the ethical, legal, and social aspects (ELSA) Scan. This assessment comprises 25 targeted questions aimed at identifying ELSA considerations. These questions were developed and based on 23 ELSA aspects of AI in agri-food literature and from testing in two case studies (arable and dairy farming). The ELSA Scan provides a clear and implementable approach for identifying ELSA in the development and use of AI in agri-food with AI developers and organisations.
{"title":"Ethical, Legal and Social Aspects (ELSA) for AI: An assessment tool for Agri-food","authors":"Mireille van Hilten , Mark Ryan , Vincent Blok , Nina de Roo","doi":"10.1016/j.atech.2024.100710","DOIUrl":"10.1016/j.atech.2024.100710","url":null,"abstract":"<div><div>As Artificial Intelligence (AI) continues to emerge in various sectors, ethical frameworks and guidelines aim to contribute to responsible AI development. While AI ethics has gained prominence in addressing broader societal concerns, existing regulations and guidelines often lack specificity for certain domain-specific applications (e.g. agri-food). AI is rapidly developed and deployed throughout the agri-food sector, but there is little practical guidance on how to do this responsibly. This study examines if the agri-food sector needs domain-specific guidance for the development and use of responsible AI and, if so, what it could look like. This research proposes it does and aims to fill this gap by introducing a novel approach for responsible AI in agri-food: the ethical, legal, and social aspects (ELSA) Scan. This assessment comprises 25 targeted questions aimed at identifying ELSA considerations. These questions were developed and based on 23 ELSA aspects of AI in agri-food literature and from testing in two case studies (arable and dairy farming). The ELSA Scan provides a clear and implementable approach for identifying ELSA in the development and use of AI in agri-food with AI developers and organisations.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100710"},"PeriodicalIF":6.3,"publicationDate":"2024-12-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143182764","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-12DOI: 10.1016/j.atech.2024.100708
Giulio Calderone, Massimo Vincenzo Ferro, Pietro Catania
In a context of global population growth, modern agriculture is required to optimize crop production efficiency. Precision agriculture (PA) can contribute to this by optimizing production and quality standards. PA has implemented the use of advanced technologies such as unmanned aerial system (UAS) that enable the implementation of resource use efficiency. UAS due to their feasibility and versatility, play a key role in orchard monitoring and plant protection product (PPP) distribution. The aim of this review is to explore the crucial role of unmanned aerial spraying systems (UASS) as innovative tools to optimize management, resource efficiency and sustainability of orchards. The review involved the analysis of a range of studies published in the period of 2019–2024. This study provides a comprehensive review of the scientific literature, focusing on the effect of aerial spray drift, considering the main parameters that could influence the application of PPP with UASS. A comparison of the operational parameters of UASS and their different configurations was examined in depth. The efficiency of droplet deposition quality of UASS on different canopy layers of trees was also examined, exploring the findings of these systems in orchards. The comparison between UASS and traditional orchard spraying methods was explored, and global legislative regulations and future perspectives were discussed. The literature highlights that using a flight height of 2 m above the canopy level (ACL), the downwash effect ensures efficient deposition of PPP while effectively limiting drift. The results show that approximately 60% of UASS spraying treatments in orchards use hydraulic nozzles, significantly more common than centrifugal nozzles. The 4-rotors UASS performed better than the 6-rotors and 8-rotors UASS in terms of deposition efficiency due to the 1.5 m working width corresponding to the canopy diameter, minimizing drift and ground losses while optimizing lateral coverage.
Although UASS application of PPP still has to overcome several challenges, the review highlighted the importance of aerial spraying applications in orchards and the huge potential of this technology.
{"title":"A systematic literature review on recent unmanned aerial spraying systems applications in orchards","authors":"Giulio Calderone, Massimo Vincenzo Ferro, Pietro Catania","doi":"10.1016/j.atech.2024.100708","DOIUrl":"10.1016/j.atech.2024.100708","url":null,"abstract":"<div><div>In a context of global population growth, modern agriculture is required to optimize crop production efficiency. Precision agriculture (PA) can contribute to this by optimizing production and quality standards. PA has implemented the use of advanced technologies such as unmanned aerial system (UAS) that enable the implementation of resource use efficiency. UAS due to their feasibility and versatility, play a key role in orchard monitoring and plant protection product (PPP) distribution. The aim of this review is to explore the crucial role of unmanned aerial spraying systems (UASS) as innovative tools to optimize management, resource efficiency and sustainability of orchards. The review involved the analysis of a range of studies published in the period of 2019–2024. This study provides a comprehensive review of the scientific literature, focusing on the effect of aerial spray drift, considering the main parameters that could influence the application of PPP with UASS. A comparison of the operational parameters of UASS and their different configurations was examined in depth. The efficiency of droplet deposition quality of UASS on different canopy layers of trees was also examined, exploring the findings of these systems in orchards. The comparison between UASS and traditional orchard spraying methods was explored, and global legislative regulations and future perspectives were discussed. The literature highlights that using a flight height of 2 m above the canopy level (ACL), the downwash effect ensures efficient deposition of PPP while effectively limiting drift. The results show that approximately 60% of UASS spraying treatments in orchards use hydraulic nozzles, significantly more common than centrifugal nozzles. The 4-rotors UASS performed better than the 6-rotors and 8-rotors UASS in terms of deposition efficiency due to the 1.5 m working width corresponding to the canopy diameter, minimizing drift and ground losses while optimizing lateral coverage.</div><div>Although UASS application of PPP still has to overcome several challenges, the review highlighted the importance of aerial spraying applications in orchards and the huge potential of this technology.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100708"},"PeriodicalIF":6.3,"publicationDate":"2024-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143182109","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this pilot research, a smart monitoring and management system based on Internet of Things (IoT) was developed and realized for apple orchards to monitor the orchard environmental conditions and to forecast the most important disease and pest, as well as to manage irrigation and fertilization. The architecture of the IoT was considered as four layers including perception layer, transport layer, processing layer and application layer. Environmental data in apple orchard was collected on-line with wireless weather and soil sensors (perception layer) and sent to gateway through LoRa radio protocol and then from the gateway to the network server (transport layer) and provided to the software for users. Methods for forecasting apple powdery mildew disease and apple codling moth, as well as thresholds for starting and stopping irrigation were determined and used as bases for decision-making in the system software. Moreover, data processing and required analysis were carried out in the system software along with the presentation of meteorological information, phenology of apple fruit growth stages, and scientific and technical instructions for fertilization in apple orchards (processing layer). A dashboard was also created to visually display the results (application layer). The results showed that the smart system is able to inform the user for the best spraying times and practical recommendations to control the biological threats. Therefore, the losses caused by the disease and pest will be reduced and consequently, the yield will be improved. Evaluation results indicated that the system based on the determined forecasting methods can reduce the number of spraying times (twice instead of at least three times to control apple powdery mildew, and twice instead of four times to control apple codling moth). Therefore, the consumption of fungicides and pesticides is reduced (>33 % and up to 50 %, respectively) which will improve the quality of apples in terms of chemical residues. Moreover, this smart system can help for optimal use of agricultural water according to the tree need with optimal management and irrigation scheduling based on the set thresholds for starting and stopping irrigation. On the other hand, applying fertilization recommendations provided in the system is based on different stages of phenology that helps for optimal fertilizer consumption. In conclusion, the use of this system can help the user to reduce production costs and to increase the quantity and quality of the product by providing timely warnings and practical recommendations regarding spraying, irrigation and fertilization. Considering the effectiveness and technical capabilities, it is recommended to implement the smart monitoring and management system in apple orchards and develop it to manage other basic challenges in such orchards.
{"title":"Internet of things-based smart system for apple orchards monitoring and management","authors":"Bahareh Jamshidi , Hossein Khabbaz Jolfaee , Kazem Mohammadpour , Mohsen Seilsepour , Hossein Dehghanisanij , Hassan Hajnajari , Hossein Farazmand , Alireza Atri","doi":"10.1016/j.atech.2024.100715","DOIUrl":"10.1016/j.atech.2024.100715","url":null,"abstract":"<div><div>In this pilot research, a smart monitoring and management system based on Internet of Things (IoT) was developed and realized for apple orchards to monitor the orchard environmental conditions and to forecast the most important disease and pest, as well as to manage irrigation and fertilization. The architecture of the IoT was considered as four layers including perception layer, transport layer, processing layer and application layer. Environmental data in apple orchard was collected on-line with wireless weather and soil sensors (perception layer) and sent to gateway through LoRa radio protocol and then from the gateway to the network server (transport layer) and provided to the software for users. Methods for forecasting apple powdery mildew disease and apple codling moth, as well as thresholds for starting and stopping irrigation were determined and used as bases for decision-making in the system software. Moreover, data processing and required analysis were carried out in the system software along with the presentation of meteorological information, phenology of apple fruit growth stages, and scientific and technical instructions for fertilization in apple orchards (processing layer). A dashboard was also created to visually display the results (application layer). The results showed that the smart system is able to inform the user for the best spraying times and practical recommendations to control the biological threats. Therefore, the losses caused by the disease and pest will be reduced and consequently, the yield will be improved. Evaluation results indicated that the system based on the determined forecasting methods can reduce the number of spraying times (twice instead of at least three times to control apple powdery mildew, and twice instead of four times to control apple codling moth). Therefore, the consumption of fungicides and pesticides is reduced (>33 % and up to 50 %, respectively) which will improve the quality of apples in terms of chemical residues. Moreover, this smart system can help for optimal use of agricultural water according to the tree need with optimal management and irrigation scheduling based on the set thresholds for starting and stopping irrigation. On the other hand, applying fertilization recommendations provided in the system is based on different stages of phenology that helps for optimal fertilizer consumption. In conclusion, the use of this system can help the user to reduce production costs and to increase the quantity and quality of the product by providing timely warnings and practical recommendations regarding spraying, irrigation and fertilization. Considering the effectiveness and technical capabilities, it is recommended to implement the smart monitoring and management system in apple orchards and develop it to manage other basic challenges in such orchards.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100715"},"PeriodicalIF":6.3,"publicationDate":"2024-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143183125","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-11DOI: 10.1016/j.atech.2024.100705
Maichol Dadi , Annalisa Franco , Giuseppe Sangiorgi , Silvio Salvi , Alessandra Lumini
Plant phenotyping is essential in agricultural research for identifying resilient traits critical for global food security. Analyzing root growth quantitatively is increasingly vital for evaluating a plant's resilience to abiotic stresses and its efficiency in nutrient and water absorption. However, extracting features from root images poses significant challenges due to the complexity of root structures, variations in size, background noise, occlusions, clutter, and inconsistent lighting conditions. In this study, we introduce “RootTracer” a software tool that offers a variety of functionalities. RootTracer enables users to quickly and easily create RSML files that capture the structure of a root system by inputting the image to be analyzed and marking or modifying key points within the image. Additionally, it allows for the modification of previously created RSML files (using any state-of-the-art creation tool) through an intuitive and user-friendly interface. The program also provides the capability to automatically extract various plant and root measurements from the RSML file. Furthermore, we present a new image dataset, named TILLMore CDC (Compact Disk Case), that includes ground truth annotations manually generated with the support of RootTracer, designed to advance the development of automated root recognition systems. This dataset, which will be publicly released, can be used by researchers to train machine learning models for accurate root image analysis, helping to overcome the challenges posed by complex root structures and varied imaging conditions. By leveraging this dataset, we aim to enhance the accuracy and robustness of root phenotyping methods, thereby contributing to the broader field of plant phenotyping and agricultural research. The RootTracer tool and the TILLMore CDC dataset are available on GitHub.
{"title":"RootTracer: An intuitive solution for root image annotation","authors":"Maichol Dadi , Annalisa Franco , Giuseppe Sangiorgi , Silvio Salvi , Alessandra Lumini","doi":"10.1016/j.atech.2024.100705","DOIUrl":"10.1016/j.atech.2024.100705","url":null,"abstract":"<div><div>Plant phenotyping is essential in agricultural research for identifying resilient traits critical for global food security. Analyzing root growth quantitatively is increasingly vital for evaluating a plant's resilience to abiotic stresses and its efficiency in nutrient and water absorption. However, extracting features from root images poses significant challenges due to the complexity of root structures, variations in size, background noise, occlusions, clutter, and inconsistent lighting conditions. In this study, we introduce “RootTracer” a software tool that offers a variety of functionalities. RootTracer enables users to quickly and easily create RSML files that capture the structure of a root system by inputting the image to be analyzed and marking or modifying key points within the image. Additionally, it allows for the modification of previously created RSML files (using any state-of-the-art creation tool) through an intuitive and user-friendly interface. The program also provides the capability to automatically extract various plant and root measurements from the RSML file. Furthermore, we present a new image dataset, named TILLMore CDC (Compact Disk Case), that includes ground truth annotations manually generated with the support of RootTracer, designed to advance the development of automated root recognition systems. This dataset, which will be publicly released, can be used by researchers to train machine learning models for accurate root image analysis, helping to overcome the challenges posed by complex root structures and varied imaging conditions. By leveraging this dataset, we aim to enhance the accuracy and robustness of root phenotyping methods, thereby contributing to the broader field of plant phenotyping and agricultural research. The RootTracer tool and the TILLMore CDC dataset are available on GitHub.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100705"},"PeriodicalIF":6.3,"publicationDate":"2024-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143183189","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-11DOI: 10.1016/j.atech.2024.100714
M.D. Rakesh, M. Jeevankumar, S.B. Rudraswamy
This work presents the implementation of deep learning models for classifying root crop leaves, specifically beetroot, potato, radish, and sweet potato. Applying ResNet50 and DenseNet121 architectures, the work demonstrates the classification based on a comprehensive dataset of over 2,500 images collected from various locations across Karnataka, India. Both models exhibited good performance, with ResNet50 achieving 99.60 % accuracy and DenseNet121 attaining 97.60 %. The models maintained high precision, recall, and F1 scores across all classes, using CPU. A key achievement was the successful deployment of these models on a Raspberry Pi 4B, with ResNet50 maintaining its high accuracy with 99.60 % and DenseNet121 achieving 96.81 % accuracy on this resource constrained device. The practical applicability was further validated through field testing, where the Raspberry Pi 4B setup was mounted on a vehicle with the webcam to capture root crop leaves in real time and successfully tested in actual agricultural field. This demonstrated the system's viability for real-time crop classification. The outcomes highlight the promise of deep learning models in agriculture technology by providing a dependable, effective, and portable method for classifying root crop leaves. The consistent high accuracy maintained across different hardware platforms and in real-world conditions demonstrates the robustness and versatility of the developed models.
{"title":"Implementation of real time root crop leaf classification using CNN on raspberry-Pi microprocessor","authors":"M.D. Rakesh, M. Jeevankumar, S.B. Rudraswamy","doi":"10.1016/j.atech.2024.100714","DOIUrl":"10.1016/j.atech.2024.100714","url":null,"abstract":"<div><div>This work presents the implementation of deep learning models for classifying root crop leaves, specifically beetroot, potato, radish, and sweet potato. Applying ResNet50 and DenseNet121 architectures, the work demonstrates the classification based on a comprehensive dataset of over 2,500 images collected from various locations across Karnataka, India. Both models exhibited good performance, with ResNet50 achieving 99.60 % accuracy and DenseNet121 attaining 97.60 %. The models maintained high precision, recall, and F1 scores across all classes, using CPU. A key achievement was the successful deployment of these models on a Raspberry Pi 4B, with ResNet50 maintaining its high accuracy with 99.60 % and DenseNet121 achieving 96.81 % accuracy on this resource constrained device. The practical applicability was further validated through field testing, where the Raspberry Pi 4B setup was mounted on a vehicle with the webcam to capture root crop leaves in real time and successfully tested in actual agricultural field. This demonstrated the system's viability for real-time crop classification. The outcomes highlight the promise of deep learning models in agriculture technology by providing a dependable, effective, and portable method for classifying root crop leaves. The consistent high accuracy maintained across different hardware platforms and in real-world conditions demonstrates the robustness and versatility of the developed models.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100714"},"PeriodicalIF":6.3,"publicationDate":"2024-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143183193","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-11DOI: 10.1016/j.atech.2024.100713
Alexandra J. Harland , Francisco J. Novais , Obioha N. Durunna , Carolyn J. Fitzsimmons , John S. Church , Edward W. Bork
Virtual fence (VF) technology uses GPS-enabled collars to manage cattle movement through audio cues and electrical pulses, offering a potential alternative to traditional physical fencing. The performance of Nofence VF collars was evaluated while in operational mode and deployed on cattle grazing within the northern temperate climate of central Alberta, Canada. Technical parameters such as network connectivity, collar failures, battery performance, and solar charging capabilities of the VF collars were evaluated across four grazing trials, three conducted in summer and one in winter. The network connection intervals, defined as the time between successive connection events, ranged from 8.1 (± 6.2) to 9.4 (± 5.4) minutes throughout the trials, remaining well within the optimal 15-minute interval, highlighting the favourable interactivity with end-users. Poor network connections occurred less than 1 % of the time, demonstrating robust coverage across the entire area. Fourteen collars experienced a network connection failure that did not persist after a manual reset. Four cattle physically lost their collars, which were then recovered and promptly redeployed. Although the mean solar charging rate was lower during the winter trial (3.1 ± 10.8 mA h-1) than the summer trials (7.9 ± 18.0 to 12.4 ± 22.1 mA h-1), mean battery charge remained greater than 96 % for all trials, even during winter when daylight was limited. While reliable cellular network access is crucial, these results indicate that Nofence VF collars can effectively function in diverse environmental conditions, and may be suitable for broader adoption by cattle producers grazing in relatively cold climates, including those of western Canada.
{"title":"Evaluation of the technical performance of the Nofence virtual fencing system in Alberta, Canada","authors":"Alexandra J. Harland , Francisco J. Novais , Obioha N. Durunna , Carolyn J. Fitzsimmons , John S. Church , Edward W. Bork","doi":"10.1016/j.atech.2024.100713","DOIUrl":"10.1016/j.atech.2024.100713","url":null,"abstract":"<div><div>Virtual fence (VF) technology uses GPS-enabled collars to manage cattle movement through audio cues and electrical pulses, offering a potential alternative to traditional physical fencing. The performance of Nofence VF collars was evaluated while in operational mode and deployed on cattle grazing within the northern temperate climate of central Alberta, Canada. Technical parameters such as network connectivity, collar failures, battery performance, and solar charging capabilities of the VF collars were evaluated across four grazing trials, three conducted in summer and one in winter. The network connection intervals, defined as the time between successive connection events, ranged from 8.1 (± 6.2) to 9.4 (± 5.4) minutes throughout the trials, remaining well within the optimal 15-minute interval, highlighting the favourable interactivity with end-users. Poor network connections occurred less than 1 % of the time, demonstrating robust coverage across the entire area. Fourteen collars experienced a network connection failure that did not persist after a manual reset. Four cattle physically lost their collars, which were then recovered and promptly redeployed. Although the mean solar charging rate was lower during the winter trial (3.1 ± 10.8 mA h<sup>-1</sup>) than the summer trials (7.9 ± 18.0 to 12.4 ± 22.1 mA h<sup>-1</sup>), mean battery charge remained greater than 96 % for all trials, even during winter when daylight was limited. While reliable cellular network access is crucial, these results indicate that Nofence VF collars can effectively function in diverse environmental conditions, and may be suitable for broader adoption by cattle producers grazing in relatively cold climates, including those of western Canada.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100713"},"PeriodicalIF":6.3,"publicationDate":"2024-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143183191","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-11DOI: 10.1016/j.atech.2024.100707
Panagiotis Karagiannis , Panagiotis Kotsaris , Vangelis Xanthakis , Panagiotis Vasilaros , George Michalos , Sotiris Makris , Frits K. van Evert , Ard T. Nieuwenhuizen , Spyros Fountas , George Chryssolouris
This paper focuses on the design and implementation of an intelligent system for planning operations of an agricultural robot. The aim of this intelligent system is to choose automatically the appropriate resources to execute certain agricultural operations, based on the user's preferences, as well as to define and schedule their sequence. The practical background is discussed, with descriptions of the solution space of agricultural applications and the way they can be restrained, through certain assumptions and decisions for minimizing the computing effort. The effectiveness of the intelligent system is demonstrated by comparing specific KPIs that have been calculated in the solution which would probably have been selected by a farmer and the solution proposed by the system, in different scenarios.
{"title":"On an intelligent system to plan agricultural operations","authors":"Panagiotis Karagiannis , Panagiotis Kotsaris , Vangelis Xanthakis , Panagiotis Vasilaros , George Michalos , Sotiris Makris , Frits K. van Evert , Ard T. Nieuwenhuizen , Spyros Fountas , George Chryssolouris","doi":"10.1016/j.atech.2024.100707","DOIUrl":"10.1016/j.atech.2024.100707","url":null,"abstract":"<div><div>This paper focuses on the design and implementation of an intelligent system for planning operations of an agricultural robot. The aim of this intelligent system is to choose automatically the appropriate resources to execute certain agricultural operations, based on the user's preferences, as well as to define and schedule their sequence. The practical background is discussed, with descriptions of the solution space of agricultural applications and the way they can be restrained, through certain assumptions and decisions for minimizing the computing effort. The effectiveness of the intelligent system is demonstrated by comparing specific KPIs that have been calculated in the solution which would probably have been selected by a farmer and the solution proposed by the system, in different scenarios.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100707"},"PeriodicalIF":6.3,"publicationDate":"2024-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143183195","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号