Pub Date : 2024-11-09DOI: 10.1016/j.compag.2024.109632
Jin Gu , Bin Zhang , Yu Wang , Yawei Zhang
Landscaping is an important way to realize carbon neutralization. The prospect of automatic trimming technology in the horticulture industry has received much attention in recent years. Compared with manual trimming, robots still have a large gap in trimming efficiency and functional integrity. The purpose of this study is to accurately obtain the shape parameters of a hedge by reconstructing its three-dimensional model, enabling the robot to have the complete ability to automate trimming, and improving the efficiency of trimming robot. Firstly, a trimming robot prototype system was constructed by using three-dimensional vision detection technology and autonomous motion control technology. Then, we studied the adaptive template matching method which was used for hedge detection, and the three-dimensional reconstruction method based on curvature feature similarity was used to obtain the position and shape parameters of hedge. We propose an adaptive Ant Colony Optimization trajectory planning method combined with point cloud classification strategy that can improve the efficiency of trimming robot. The results of tests show that the mean absolute value of measurement error of the hand-eye system is 3.7 mm, the mean value of the positioning error of the visual recognition is 2.1 mm, and the mean value of the positioning error of the trimming robot system is 3.8 mm. The trimming robot realized the automatic trimming operation of spherical hedge model and actual hedge in laboratory. During the actual trimming test, it demonstrated an average error of 8.2 mm, and its efficiency and reliability in trimming surpassed manual trimming methods. The research suggests that with the continuous improvement of robot technology, the use of trimming robot system in the horticulture industry will gradually become a reality.
{"title":"Hedge three-dimensional reconstruction and motion control technology for trimming robot","authors":"Jin Gu , Bin Zhang , Yu Wang , Yawei Zhang","doi":"10.1016/j.compag.2024.109632","DOIUrl":"10.1016/j.compag.2024.109632","url":null,"abstract":"<div><div>Landscaping is an important way to realize carbon neutralization. The prospect of automatic trimming technology in the horticulture industry has received much attention in recent years. Compared with manual trimming, robots still have a large gap in trimming efficiency and functional integrity. The purpose of this study is to accurately obtain the shape parameters of a hedge by reconstructing its three-dimensional model, enabling the robot to have the complete ability to automate trimming, and improving the efficiency of trimming robot. Firstly, a trimming robot prototype system was constructed by using three-dimensional vision detection technology and autonomous motion control technology. Then, we studied the adaptive template matching method which was used for hedge detection, and the three-dimensional reconstruction method based on curvature feature similarity was used to obtain the position and shape parameters of hedge. We propose an adaptive Ant Colony Optimization trajectory planning method combined with point cloud classification strategy that can improve the efficiency of trimming robot. The results of tests show that the mean absolute value of measurement error of the hand-eye system is 3.7 mm, the mean value of the positioning error of the visual recognition is 2.1 mm, and the mean value of the positioning error of the trimming robot system is 3.8 mm. The trimming robot realized the automatic trimming operation of spherical hedge model and actual hedge in laboratory. During the actual trimming test, it demonstrated an average error of 8.2 mm, and its efficiency and reliability in trimming surpassed manual trimming methods. The research suggests that with the continuous improvement of robot technology, the use of trimming robot system in the horticulture industry will gradually become a reality.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109632"},"PeriodicalIF":7.7,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142661817","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}
Pub Date : 2024-11-09DOI: 10.1016/j.compag.2024.109595
Weihong Ma , Xingmeng Wang , Simon X. Yang , Xianglong Xue , Mingyu Li , Rong Wang , Ligen Yu , Lepeng Song , Qifeng Li
Daily inspections of individual laying hens in large-scale egg farms are both labor-intensive and time-consuming, requiring farm staff to manually check each caged hen and promptly remove any deceased birds to prevent the spread of disease within the battery cages. To streamline this process, a specialized robot has been developed to enhance inspection efficiency, reduce manual labor, and enable rapid identification of dead hens. This inspection robot integrates cutting-edge technologies such as deep learning for real-time detection and identification, QR code-based positioning for precise localization, and autonomous navigation for seamless movement through the farm. It automates the otherwise tedious inspection process by visualizing and pinpointing the location of dead hens within the cages. In experimental tests, the robot achieved a detection accuracy of 90.61 % by incorporating a supplementary lighting system, setting an inspection speed of 9 m per minute, and fine-tuning the inspection algorithm with a probability value parameter of 0.48 and an area ratio parameter of 0.05. Additionally, the robot demonstrated a low false detection rate of 0.14 % and a minimal obvious false detection rate of 0.06 %. Compared to traditional manual inspection methods, this robotic system not only automates the task but also significantly reduces labor requirements and improves the overall management efficiency of large-scale egg farms. With its high accuracy and speed, the robot presents a viable solution for modern poultry operations, ensuring timely removal of dead hens and contributing to better farm hygiene and animal welfare.
{"title":"Autonomous inspection robot for dead laying hens in caged layer house","authors":"Weihong Ma , Xingmeng Wang , Simon X. Yang , Xianglong Xue , Mingyu Li , Rong Wang , Ligen Yu , Lepeng Song , Qifeng Li","doi":"10.1016/j.compag.2024.109595","DOIUrl":"10.1016/j.compag.2024.109595","url":null,"abstract":"<div><div>Daily inspections of individual laying hens in large-scale egg farms are both labor-intensive and time-consuming, requiring farm staff to manually check each caged hen and promptly remove any deceased birds to prevent the spread of disease within the battery cages. To streamline this process, a specialized robot has been developed to enhance inspection efficiency, reduce manual labor, and enable rapid identification of dead hens. This inspection robot integrates cutting-edge technologies such as deep learning for real-time detection and identification, QR code-based positioning for precise localization, and autonomous navigation for seamless movement through the farm. It automates the otherwise tedious inspection process by visualizing and pinpointing the location of dead hens within the cages. In experimental tests, the robot achieved a detection accuracy of 90.61 % by incorporating a supplementary lighting system, setting an inspection speed of 9 m per minute, and fine-tuning the inspection algorithm with a probability value parameter of 0.48 and an area ratio parameter of 0.05. Additionally, the robot demonstrated a low false detection rate of 0.14 % and a minimal obvious false detection rate of 0.06 %. Compared to traditional manual inspection methods, this robotic system not only automates the task but also significantly reduces labor requirements and improves the overall management efficiency of large-scale egg farms. With its high accuracy and speed, the robot presents a viable solution for modern poultry operations, ensuring timely removal of dead hens and contributing to better farm hygiene and animal welfare.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109595"},"PeriodicalIF":7.7,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142661762","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}
Pub Date : 2024-11-09DOI: 10.1016/j.compag.2024.109548
Shaoning Pang , Shyh Wei Teng , Manzur Murshed , Cuong Van Bui , Priyabrata Karmakar , Yanyu Li , Hao Lin
The integration of blockchain technology in agricultural traceability has shown immense potential, yet its widespread adoption faces significant roadblocks. Using bulk product traceability as a foundational reference, this paper presents a comprehensive evaluation framework for Blockchain-based Agricultural Traceability. The framework accentuates product identification and data traceability across the supply chain, addressing traceability disconnections caused by bulk product blending. It dives into depth levels from adoption decision-making to system design, development, and deployment, emphasizing the critical aspects of traceability compliance and standardization. As a result, we identified the obstacles to adopting agricultural digital traceability and pave the pathway to traceability system deployment. We examined the barriers to implementing digital traceability of agricultural products, taking the Australian grain supply chain as an example. Our findings reveal that lack of standardization and participation barriers are the primary challenges in implementing digital traceability for agricultural products. Our paper offers insights and recommendations for researchers, industry practitioners, and business owners to overcome these challenges and enable digital traceability of agricultural products in global supply chains.
{"title":"A survey on evaluation of blockchain-based agricultural traceability","authors":"Shaoning Pang , Shyh Wei Teng , Manzur Murshed , Cuong Van Bui , Priyabrata Karmakar , Yanyu Li , Hao Lin","doi":"10.1016/j.compag.2024.109548","DOIUrl":"10.1016/j.compag.2024.109548","url":null,"abstract":"<div><div>The integration of blockchain technology in agricultural traceability has shown immense potential, yet its widespread adoption faces significant roadblocks. Using bulk product traceability as a foundational reference, this paper presents a comprehensive evaluation framework for Blockchain-based Agricultural Traceability. The framework accentuates product identification and data traceability across the supply chain, addressing traceability disconnections caused by bulk product blending. It dives into depth levels from adoption decision-making to system design, development, and deployment, emphasizing the critical aspects of traceability compliance and standardization. As a result, we identified the obstacles to adopting agricultural digital traceability and pave the pathway to traceability system deployment. We examined the barriers to implementing digital traceability of agricultural products, taking the Australian grain supply chain as an example. Our findings reveal that lack of standardization and participation barriers are the primary challenges in implementing digital traceability for agricultural products. Our paper offers insights and recommendations for researchers, industry practitioners, and business owners to overcome these challenges and enable digital traceability of agricultural products in global supply chains.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109548"},"PeriodicalIF":7.7,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142661680","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-09DOI: 10.1016/j.compag.2024.109608
Linxiao Miao , Peng Wang , Haifeng Cao , Zhenqing Zhao , Zhenbang Hu , Qingshan Chen , Dawei Xin , Rongsheng Zhu
Accurately determining the stage of crop development holds significant importance for field crop management. With the advancement of smart agriculture, an increasing number of Internet of Things (IoT) devices are being integrated into agricultural production, enabling more efficient acquisition of high-precision crop images. Currently, research on detecting crop growth stages based on IoT device images remains relatively scarce. Most existing studies rely on a single network model for detection, often encountering issues such as low accuracy and overfitting. Therefore, in this study, we collected maize images using IoT devices and constructed an integrated deep learning model by utilizing four convolutional neural networks (CNNs) to detect the growth period of maize in real time. Additionally, we implemented several improvements on these four CNNs and subsequently tested the performance of the ensemble model on the maize dataset. Regarding the ensemble strategy for the ensemble model, we proposed a dynamic weighted voting method, building upon the original voting approach, which can mitigate model training fluctuations and expedite model convergence. Ultimately, we manually simulated various lighting conditions to assess their impact on the ensemble model. Experimental results demonstrate that the ensemble deep model proposed in this paper represents a robust method for detecting maize growth stages, achieving an accuracy rate of 0.976 on the maize dataset, effectively facilitating high-precision detection of maize growth stages in complex backgrounds.
{"title":"A high-precision automatic diagnosis method of maize developmental stage based on ensemble deep learning with IoT devices","authors":"Linxiao Miao , Peng Wang , Haifeng Cao , Zhenqing Zhao , Zhenbang Hu , Qingshan Chen , Dawei Xin , Rongsheng Zhu","doi":"10.1016/j.compag.2024.109608","DOIUrl":"10.1016/j.compag.2024.109608","url":null,"abstract":"<div><div>Accurately determining the stage of crop development holds significant importance for field crop management. With the advancement of smart agriculture, an increasing number of Internet of Things (IoT) devices are being integrated into agricultural production, enabling more efficient acquisition of high-precision crop images. Currently, research on detecting crop growth stages based on IoT device images remains relatively scarce. Most existing studies rely on a single network model for detection, often encountering issues such as low accuracy and overfitting. Therefore, in this study, we collected maize images using IoT devices and constructed an integrated deep learning model by utilizing four convolutional neural networks (CNNs) to detect the growth period of maize in real time. Additionally, we implemented several improvements on these four CNNs and subsequently tested the performance of the ensemble model on the maize dataset. Regarding the ensemble strategy for the ensemble model, we proposed a dynamic weighted voting method, building upon the original voting approach, which can mitigate model training fluctuations and expedite model convergence. Ultimately, we manually simulated various lighting conditions to assess their impact on the ensemble model. Experimental results demonstrate that the ensemble deep model proposed in this paper represents a robust method for detecting maize growth stages, achieving an accuracy rate of 0.976 on the maize dataset, effectively facilitating high-precision detection of maize growth stages in complex backgrounds.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109608"},"PeriodicalIF":7.7,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142661778","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}
Multiple agricultural machines for cooperative operations are a technology that utilizes multiple agricultural machines to work together to complete specific agricultural tasks. This technology can improve productivity and reduce labor intensity, which can help alleviate the impact of population aging on agriculture. However, there remains a challenge in the practical application of this technology in terms of adaptability, operational accuracy, and operational efficiency. This review aims to systematically overview the status of multiple agricultural machines for cooperative operations through multiple machines planning technology, multiple machines communication technology, and multiple machines cooperative control. Taking the harvester-grain truck owner from the cooperative harvesting scene and the homogeneous multi-machine cooperation scene of the same kind of operation as an example, this work summarizes the typical operation scenarios and application areas of multi-machine cooperation in agriculture. Additionally, this review discusses the challenges and future directions of multiple agricultural machines for cooperative operations.
{"title":"Research progress of multiple agricultural machines for cooperative operations: A review","authors":"Wenbo Wei , Maohua Xiao , Hui Wang , Yejun Zhu , Chenshuo Xie , Guosheng Geng","doi":"10.1016/j.compag.2024.109628","DOIUrl":"10.1016/j.compag.2024.109628","url":null,"abstract":"<div><div>Multiple agricultural machines for cooperative operations are a technology that utilizes multiple agricultural machines to work together to complete specific agricultural tasks. This technology can improve productivity and reduce labor intensity, which can help alleviate the impact of population aging on agriculture. However, there remains a challenge in the practical application of this technology in terms of adaptability, operational accuracy, and operational efficiency. This review aims to systematically overview the status of multiple agricultural machines for cooperative operations through multiple machines planning technology, multiple machines communication technology, and multiple machines cooperative control. Taking the harvester-grain truck owner from the cooperative harvesting scene and the homogeneous multi-machine cooperation scene of the same kind of operation as an example, this work summarizes the typical operation scenarios and application areas of multi-machine cooperation in agriculture. Additionally, this review discusses the challenges and future directions of multiple agricultural machines for cooperative operations.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109628"},"PeriodicalIF":7.7,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142661818","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}
Pub Date : 2024-11-09DOI: 10.1016/j.compag.2024.109598
Yuchao Wang , Chunhai Fu , Ruiyu Huang , Kelin Tong , Yong He , Lijia Xu
In complex greenhouse orchard environments, reasonable path planning algorithms are crucial for ensuring efficient and high-quality operation of mobile robots. The unstructured layouts of greenhouse orchard environments, which feature many irregular obstacles, pose high demands on navigation accuracy. Ideal path planning algorithms need to plan a safe and efficient navigation path in complex environments. In this paper, we propose a path planning fusion algorithm which integrates improved A* algorithm and Fuzzy Dynamic Window Approach (FDWA) algorithm. Firstly, the A* algorithm that introduces the rate of environmental obstacles is designed for generating global paths in greenhouses. The search strategy can be changed according to the number of environmental obstacles. Then, a rule to optimize the search neighborhood is proposed to adjust the search neighborhood to five-neighborhood, which improves the node search efficiency. Further, a local path planning strategy incorporating fuzzy control is proposed to enable the robot to maintain a safe distance from obstacles and improve the stability of obstacle avoidance. Finally, the effectiveness of proposed algorithm is verified via the simulated environment and actual greenhouse, respectively. The simulation results show that, the improved A* algorithm reduces the critical turning points and total steering angle by a maximum of 40%. The actual greenhouse experimental results show that, in three different paths, the proposed fusion algorithm reduces the distance deviation by a maximum of 31.8% and the heading angle deviation by a maximum of 28.6%, while increasing the safety distance by up to 30%.
{"title":"Path planning for mobile robots in greenhouse orchards based on improved A* and fuzzy DWA algorithms","authors":"Yuchao Wang , Chunhai Fu , Ruiyu Huang , Kelin Tong , Yong He , Lijia Xu","doi":"10.1016/j.compag.2024.109598","DOIUrl":"10.1016/j.compag.2024.109598","url":null,"abstract":"<div><div>In complex greenhouse orchard environments, reasonable path planning algorithms are crucial for ensuring efficient and high-quality operation of mobile robots. The unstructured layouts of greenhouse orchard environments, which feature many irregular obstacles, pose high demands on navigation accuracy. Ideal path planning algorithms need to plan a safe and efficient navigation path in complex environments. In this paper, we propose a path planning fusion algorithm which integrates improved A* algorithm and Fuzzy Dynamic Window Approach (FDWA) algorithm. Firstly, the A* algorithm that introduces the rate of environmental obstacles is designed for generating global paths in greenhouses. The search strategy can be changed according to the number of environmental obstacles. Then, a rule to optimize the search neighborhood is proposed to adjust the search neighborhood to five-neighborhood, which improves the node search efficiency. Further, a local path planning strategy incorporating fuzzy control is proposed to enable the robot to maintain a safe distance from obstacles and improve the stability of obstacle avoidance. Finally, the effectiveness of proposed algorithm is verified via the simulated environment and actual greenhouse, respectively. The simulation results show that, the improved A* algorithm reduces the critical turning points and total steering angle by a maximum of 40%. The actual greenhouse experimental results show that, in three different paths, the proposed fusion algorithm reduces the distance deviation by a maximum of 31.8% and the heading angle deviation by a maximum of 28.6%, while increasing the safety distance by up to 30%.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109598"},"PeriodicalIF":7.7,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142661769","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}
Pub Date : 2024-11-09DOI: 10.1016/j.compag.2024.109621
Yang Liu , Mingjia Liu , Guohui Liu , Hong Sun , Lulu An , Ruomei Zhao , Weijie Tang , Fangkui Zhao , Xiaojing Yan , Yuntao Ma , Minzan Li
The wheat powdery mildew (WPM) always alters the pigment and structure of the leaf and canopy, disrupting crop growth. A challenge on the WPM monitoring is the limited capability of unmanned aerial vehicle (UAV)-based canopy images to directly indicate complex infection symptoms. However, the WPM infection markedly changed canopy chlorophyll content (CCC), which encompassed both leaf and canopy attributes, and this change was relatively easy to capture by UAV remote sensing. Thus, this study aimed to estimate CCC to indirectly explore WPM using different scales of UAV image features. UAV-based winter wheat canopy images were acquired continuously in the field during the early, middle, and late infection stages after being artificially inoculated with fungal pathogens at the Institute of Plant Protection, Chinese Academy of Agricultural Sciences, Xinxiang, China, in 2022. The study evaluated the potential of spectral (Spe) and textural (Tex) features and their combination to estimate CCC and characterize WPM dynamic. Considering the impacts of spatial scales, the selected Spe and Tex textures were calculated from images with 1, 2, 5, 10, 15, and 20 cm spatial resolution. The changes in different types of features under WPM stress and their response to CCC were analyzed. Three regression methods, including extreme gradient boosting regression (XGBR), multilayer perceptron regression (MLPR), and partial least squares regression (PLSR) were used to estimate CCC based on the acquired sensitive features and track the infection status. Results showed that the image spatial resolution barely affected the Spe performance while notably affecting the Tex performance. The performance of estimating CCC under WPM stress was superior for Tex (ranging from 1 to 20 cm spatial resolution imagery) compared to Spe features. The best modeling result was the combination of Spe with Tex features from 1 and 10 cm (R2 = 0.82, RMSE = 28.49 mg/L, NRMSE = 12.38 %), which could be related to information captured from different viewpoints. Although finer spatial resolution was advantageous for capturing the complex symptoms caused by WPM, it increased the burden on UAV missions. UAV multispectral imagery with the 10 cm spatial resolution using XGBR (R2 = 0.74, RMSE = 33.48 mg/L, NRMSE = 14.55 %) might be used as an optimization scheme for estimating CCC and exploring WPM stress, as it decreased the cost associated with data processing and time in the actual operation. This study indirectly characterizes the condition of WPM infection by estimating CCC, which provides promising and valuable insights for disease management and control in the field.
{"title":"Estimating canopy chlorophyll content of powdery mildew stressed winter wheat by different spatial resolutions of UAV-imagery","authors":"Yang Liu , Mingjia Liu , Guohui Liu , Hong Sun , Lulu An , Ruomei Zhao , Weijie Tang , Fangkui Zhao , Xiaojing Yan , Yuntao Ma , Minzan Li","doi":"10.1016/j.compag.2024.109621","DOIUrl":"10.1016/j.compag.2024.109621","url":null,"abstract":"<div><div>The wheat powdery mildew (WPM) always alters the pigment and structure of the leaf and canopy, disrupting crop growth. A challenge on the WPM monitoring is the limited capability of unmanned aerial vehicle (UAV)-based canopy images to directly indicate complex infection symptoms. However, the WPM infection markedly changed canopy chlorophyll content (CCC), which encompassed both leaf and canopy attributes, and this change was relatively easy to capture by UAV remote sensing. Thus, this study aimed to estimate CCC to indirectly explore WPM using different scales of UAV image features. UAV-based winter wheat canopy images were acquired continuously in the field during the early, middle, and late infection stages after being artificially inoculated with fungal pathogens at the Institute of Plant Protection, Chinese Academy of Agricultural Sciences, Xinxiang, China, in 2022. The study evaluated the potential of spectral (Spe) and textural (Tex) features and their combination to estimate CCC and characterize WPM dynamic. Considering the impacts of spatial scales, the selected Spe and Tex textures were calculated from images with 1, 2, 5, 10, 15, and 20 cm spatial resolution. The changes in different types of features under WPM stress and their response to CCC were analyzed. Three regression methods, including extreme gradient boosting regression (XGBR), multilayer perceptron regression (MLPR), and partial least squares regression (PLSR) were used to estimate CCC based on the acquired sensitive features and track the infection status. Results showed that the image spatial resolution barely affected the Spe performance while notably affecting the Tex performance. The performance of estimating CCC under WPM stress was superior for Tex (ranging from 1 to 20 cm spatial resolution imagery) compared to Spe features. The best modeling result was the combination of Spe with Tex features from 1 and 10 cm (R<sup>2</sup> = 0.82, RMSE = 28.49 mg/L, NRMSE = 12.38 %), which could be related to information captured from different viewpoints. Although finer spatial resolution was advantageous for capturing the complex symptoms caused by WPM, it increased the burden on UAV missions. UAV multispectral imagery with the 10 cm spatial resolution using XGBR (R<sup>2</sup> = 0.74, RMSE = 33.48 mg/L, NRMSE = 14.55 %) might be used as an optimization scheme for estimating CCC and exploring WPM stress, as it decreased the cost associated with data processing and time in the actual operation. This study indirectly characterizes the condition of WPM infection by estimating CCC, which provides promising and valuable insights for disease management and control in the field.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109621"},"PeriodicalIF":7.7,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142661777","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}
Pub Date : 2024-11-09DOI: 10.1016/j.compag.2024.109600
Hanhui Jiang , Liguo Jiang , Leilei He , Bryan Gilbert Murengami , Xudong Jing , Paula A. Misiewicz , Fernando Auat Cheein , Longsheng Fu
Yield information of root crops guides precision agriculture efforts and optimizes resource allocation. Predicting root crops prior to harvest is crucial to crop management and planning and requires obtaining root crop yield without damaging them. Non-destructive access to yield of root crops is challenging because of the edible portion of the crops being located underground, which impacts precision agriculture technology application. Remote sensing provides a possible way to solve this problem. There are no review reports on yield prediction for root crops using remote sensing, though root crops share the same growth characteristic of producing edible parts underground, which makes their yield prediction techniques similar. In this work, a total of 49 sources on the use of remote sensing techniques for yield prediction of root crops in field were collected, analyzed and discussed from the aspects of remote sensing platforms, input features and modelling methods. In terms of usage counts of remote sensing platforms, ground penetrating radars that are directly exposed to edible parts of root crops have the potential to be applied to root crop yield predictions, while spaceborne platforms are the current trend, accounting for 51 %. Feature combination from environment and crop itself is beneficial to crop yield prediction models, particularly the processed-based crop models. It is recommended to collect data time after ensuring specific root data types. Additionally, full-cycle data is suggested to be used to increase robustness of root crop yield prediction models. The result showed that plant-by-plant detection was only applied to radar-based platforms while spectral-based platforms are still in plot level, which further investigated that improving accuracy of root crop yield prediction through individual above ground phenotypic traits. The review is intended to summarize the development of root crop yield prediction using remote sensing and put forward further for further improvement.
{"title":"Yield prediction of root crops in field using remote sensing: A comprehensive review","authors":"Hanhui Jiang , Liguo Jiang , Leilei He , Bryan Gilbert Murengami , Xudong Jing , Paula A. Misiewicz , Fernando Auat Cheein , Longsheng Fu","doi":"10.1016/j.compag.2024.109600","DOIUrl":"10.1016/j.compag.2024.109600","url":null,"abstract":"<div><div>Yield information of root crops guides precision agriculture efforts and optimizes resource allocation. Predicting root crops prior to harvest is crucial to crop management and planning and requires obtaining root crop yield without damaging them. Non-destructive access to yield of root crops is challenging because of the edible portion of the crops being located underground, which impacts precision agriculture technology application. Remote sensing provides a possible way to solve this problem. There are no review reports on yield prediction for root crops using remote sensing, though root crops share the same growth characteristic of producing edible parts underground, which makes their yield prediction techniques similar. In this work, a total of 49 sources on the use of remote sensing techniques for yield prediction of root crops in field were collected, analyzed and discussed from the aspects of remote sensing platforms, input features and modelling methods. In terms of usage counts of remote sensing platforms, ground penetrating radars that are directly exposed to edible parts of root crops have the potential to be applied to root crop yield predictions, while spaceborne platforms are the current trend, accounting for 51 %. Feature combination from environment and crop itself is beneficial to crop yield prediction models, particularly the processed-based crop models. It is recommended to collect data time after ensuring specific root data types. Additionally, full-cycle data is suggested to be used to increase robustness of root crop yield prediction models. The result showed that plant-by-plant detection was only applied to radar-based platforms while spectral-based platforms are still in plot level, which further investigated that improving accuracy of root crop yield prediction through individual above ground phenotypic traits. The review is intended to summarize the development of root crop yield prediction using remote sensing and put forward further for further improvement.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109600"},"PeriodicalIF":7.7,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142661357","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}
Pub Date : 2024-11-09DOI: 10.1016/j.compag.2024.109626
Chengkun Zhai , Caiyun Lu , Hongwen Li , Jin He , Qingjie Wang , Fangle Chang , Jinshuo Bi , Zhengyang Wu
In the context of precision agriculture, real-time monitoring of maize seeding parameters is of great significance for evaluating seeding situations and ensuring seeding quality. At present, seeding monitoring mainly uses the through beam photoelectric (TBP) method, which is susceptible to dust and can only be used at the upper part of the seed tube, affecting monitoring accuracy. For this purpose, this study developed a maize seeding parameter monitoring system based on near-infrared diffusion emission-diffuse reflectance (NIRDE-DR), which utilizes the diffusion emission effect of NIR rays to form a three-dimensional monitoring area for maize seeds without missed monitoring. When maize seeds with uneven surfaces enter the monitoring area, the diffuse reflectance effect of the seeds on NIR rays is utilized to change the electrical signal of the monitoring system, and the recognition of falling seeds is achieved by processing the electrical signal. NIRDE-DR takes advantage of the small size of dust particles, which are difficult to form a reflective area, effectively avoiding dust interference. Therefore, it can perform high-precision monitoring at the end of the seed tube. The NIR spectrum of coated maize seeds was measured, and the NIR wavenumber with the lowest absorbance and strongest reflection ability of maize seeds was determined as the target wavenumber of the monitoring system. The impact of the horizontal distance from the monitoring surface to the inner wall of the seed tube (HD) on seeding monitoring was clarified. The value of HD in the developed seeding parameter monitoring system was determined, so that when the NIR rays are emitted into the seed tube, they can cover the entire cross-section of the end of the seed tube without being reflected by dust, avoiding missed monitoring and false monitoring. A signal shielding filtering algorithm based on sawtooth wave shielding was proposed. In regard to the characteristic of high-frequency sawtooth wave in the signal generated by seeds passing through the monitoring area, the first rising edge of the signal is used as the seed recognition signal. By analyzing the duration of high-frequency sawtooth wave and the interval between adjacent seeds, the shielding time of the interference signal is determined to achieve effective noise reduction. Performance evaluation test in the bench results showed that NIRDE-DR has a better recognition effect on maize seeds than TBP. Performance evaluation test in the field showed that at a seeding speed of 6–14 km/h, the maximum monitoring error of the developed system for seeding quantity was 7.98 %, and the maximum monitoring error for seeding qualified rate was 7.69 %. The developed seeding parameter monitoring system has good performance, providing a reference for the advancement of seeding parameter monitoring technology at the end of the seed tube.
在精准农业背景下，玉米播种参数的实时监测对于评估播种情况、确保播种质量具有重要意义。目前，播种监测主要采用透射光电法（TBP），该方法易受灰尘影响，且只能在种子管上部使用，影响监测精度。为此，本研究开发了一种基于近红外扩散发射-漫反射（NIRDE-DR）的玉米播种参数监测系统，利用近红外射线的扩散发射效应，形成玉米种子的三维监测区域，无遗漏监测。当表面凹凸不平的玉米种子进入监测区域时，利用种子对近红外射线的漫反射效应改变监测系统的电信号，通过处理电信号实现对掉落种子的识别。NIRDE-DR 利用灰尘颗粒小，难以形成反射区的特点，有效避免了灰尘干扰。因此，它可以在种子管末端进行高精度监测。测量了包衣玉米种子的近红外光谱，确定了玉米种子吸光率最低、反射能力最强的近红外波长作为监测系统的目标波长。明确了监测面到种子管内壁的水平距离（HD）对播种监测的影响。确定了所开发的播种参数监测系统的 HD 值，使近红外射线射入播种管时，能覆盖播种管末端的整个横截面而不被灰尘反射，避免漏测和误测。提出了一种基于锯齿波屏蔽的信号屏蔽滤波算法。针对种子通过监测区域时产生的信号中存在高频锯齿波的特点，将信号的第一个上升沿作为种子识别信号。通过分析高频锯齿波的持续时间和相邻种子之间的间隔，确定干扰信号的屏蔽时间，从而实现有效降噪。台架性能评估测试结果表明，NIRDE-DR 对玉米种子的识别效果优于 TBP。田间性能评估测试表明，在播种速度为 6-14 km/h 时，所开发系统对播种量的最大监测误差为 7.98%，对播种合格率的最大监测误差为 7.69%。所开发的播种参数监测系统性能良好，为种子管末端播种参数监测技术的进步提供了参考。
{"title":"A precise maize seeding parameter monitoring system at the end of seed tube: Improving monitoring accuracy using near-infrared diffusion emission-diffuse reflectance (NIRDE-DR)","authors":"Chengkun Zhai , Caiyun Lu , Hongwen Li , Jin He , Qingjie Wang , Fangle Chang , Jinshuo Bi , Zhengyang Wu","doi":"10.1016/j.compag.2024.109626","DOIUrl":"10.1016/j.compag.2024.109626","url":null,"abstract":"<div><div>In the context of precision agriculture, real-time monitoring of maize seeding parameters is of great significance for evaluating seeding situations and ensuring seeding quality. At present, seeding monitoring mainly uses the through beam photoelectric (TBP) method, which is susceptible to dust and can only be used at the upper part of the seed tube, affecting monitoring accuracy. For this purpose, this study developed a maize seeding parameter monitoring system based on near-infrared diffusion emission-diffuse reflectance (NIRDE-DR), which utilizes the diffusion emission effect of NIR rays to form a three-dimensional monitoring area for maize seeds without missed monitoring. When maize seeds with uneven surfaces enter the monitoring area, the diffuse reflectance effect of the seeds on NIR rays is utilized to change the electrical signal of the monitoring system, and the recognition of falling seeds is achieved by processing the electrical signal. NIRDE-DR takes advantage of the small size of dust particles, which are difficult to form a reflective area, effectively avoiding dust interference. Therefore, it can perform high-precision monitoring at the end of the seed tube. The NIR spectrum of coated maize seeds was measured, and the NIR wavenumber with the lowest absorbance and strongest reflection ability of maize seeds was determined as the target wavenumber of the monitoring system. The impact of the horizontal distance from the monitoring surface to the inner wall of the seed tube (HD) on seeding monitoring was clarified. The value of HD in the developed seeding parameter monitoring system was determined, so that when the NIR rays are emitted into the seed tube, they can cover the entire cross-section of the end of the seed tube without being reflected by dust, avoiding missed monitoring and false monitoring. A signal shielding filtering algorithm based on sawtooth wave shielding was proposed. In regard to the characteristic of high-frequency sawtooth wave in the signal generated by seeds passing through the monitoring area, the first rising edge of the signal is used as the seed recognition signal. By analyzing the duration of high-frequency sawtooth wave and the interval between adjacent seeds, the shielding time of the interference signal is determined to achieve effective noise reduction. Performance evaluation test in the bench results showed that NIRDE-DR has a better recognition effect on maize seeds than TBP. Performance evaluation test in the field showed that at a seeding speed of 6–14 km/h, the maximum monitoring error of the developed system for seeding quantity was 7.98 %, and the maximum monitoring error for seeding qualified rate was 7.69 %. The developed seeding parameter monitoring system has good performance, providing a reference for the advancement of seeding parameter monitoring technology at the end of the seed tube.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109626"},"PeriodicalIF":7.7,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142661761","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}
Pub Date : 2024-11-09DOI: 10.1016/j.compag.2024.109606
Malte von Bloh , David Lobell , Senthold Asseng
Research on yield predictions is dominated by two approaches: machine learning and process-based models. Machine learning has shown impressive results in capturing complex relationships but is often limited by data availability in agriculture. Conversely, process-based models, with over 60 years of research history, simulate crop growth processes using biophysical equations. Here, we present a method to transfer domain knowledge from the Decision Support System for Agrotechnology Transfer framework (DSSAT) using the Nwheat crop simulation process-model into neural networks and random forest for predicting wheat yield at field scale. Expanding the feature and distribution space involved simulating crop parameters and synthetic samples through the utilization of observed and historical weather recordings, as well as future climate projections. We demonstrated that neural networks can learn both general crop growth and yield processes and then effectively adapt to regional, field-specific growth patterns using synthetic and high-resolution field data. This approach boosts overall performance and reduces model error by 8 % compared to a purely data-centric model without process-knowledge transfer and solely trained on observed field data and features. Synthetic samples generated from warmer conditions were the greatest driver for improvements and we showed that the climate scenario for data generation is more important than the actual synthetic data set size. The proposed method shows the potential of combining process-based and machine-learning models, highlighting the potential to leverage the strengths of both methods in a collaborative manner.
{"title":"Knowledge informed hybrid machine learning in agricultural yield prediction","authors":"Malte von Bloh , David Lobell , Senthold Asseng","doi":"10.1016/j.compag.2024.109606","DOIUrl":"10.1016/j.compag.2024.109606","url":null,"abstract":"<div><div>Research on yield predictions is dominated by two approaches: machine learning and process-based models. Machine learning has shown impressive results in capturing complex relationships but is often limited by data availability in agriculture. Conversely, process-based models, with over 60 years of research history, simulate crop growth processes using biophysical equations. Here, we present a method to transfer domain knowledge from the Decision Support System for Agrotechnology Transfer framework (DSSAT) using the Nwheat crop simulation process-model into neural networks and random forest for predicting wheat yield at field scale. Expanding the feature and distribution space involved simulating crop parameters and synthetic samples through the utilization of observed and historical weather recordings, as well as future climate projections. We demonstrated that neural networks can learn both general crop growth and yield processes and then effectively adapt to regional, field-specific growth patterns using synthetic and high-resolution field data. This approach boosts overall performance and reduces model error by 8 % compared to a purely data-centric model without process-knowledge transfer and solely trained on observed field data and features. Synthetic samples generated from warmer conditions were the greatest driver for improvements and we showed that the climate scenario for data generation is more important than the actual synthetic data set size. The proposed method shows the potential of combining process-based and machine-learning models, highlighting the potential to leverage the strengths of both methods in a collaborative manner.</div></div>","PeriodicalId":50627,"journal":{"name":"Computers and Electronics in Agriculture","volume":"227 ","pages":"Article 109606"},"PeriodicalIF":7.7,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142661768","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","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