Pub Date : 2024-12-30DOI: 10.1016/j.atech.2024.100738
Alia Khalid , Muhammad Latif Anjum , Salman Naveed , Wajahat Hussain
Detecting weak wingbeats of a flying bug is a challenging problem in uncontrolled outdoor settings. In this work, we show that proper treatment of environmental noise is a key factor in robust acoustic classifier design and propose a novel environmental noise treatment method. Our proposed method generalizes over different classifiers and features. Our algorithm provides robust detection and classification of multiple bugs, over longest ranges reported, using simple microphones. In order to benchmark research in this area, we release a novel dataset containing acoustic data of four bugs (Guava fly, Melon fly, Blue bottle fly, and mosquitoes). We additionally investigate the feasibility of deploying our acoustic classifier on a noisy mobile platform, i.e., a drone. To this end, we expose the limitations of signal processing techniques to deal with loud drone noise. We demonstrate how soundproofing can be used to design acoustic sensing for drones.
{"title":"Whispers in the air: Designing acoustic classifiers to detect fruit flies from afar","authors":"Alia Khalid , Muhammad Latif Anjum , Salman Naveed , Wajahat Hussain","doi":"10.1016/j.atech.2024.100738","DOIUrl":"10.1016/j.atech.2024.100738","url":null,"abstract":"<div><div>Detecting weak wingbeats of a flying bug is a challenging problem in uncontrolled outdoor settings. In this work, we show that proper treatment of environmental noise is a key factor in robust acoustic classifier design and propose a novel environmental noise treatment method. Our proposed method generalizes over different classifiers and features. Our algorithm provides robust detection and classification of multiple bugs, over longest ranges reported, using simple microphones. In order to benchmark research in this area, we release a novel dataset containing acoustic data of four bugs (Guava fly, Melon fly, Blue bottle fly, and mosquitoes). We additionally investigate the feasibility of deploying our acoustic classifier on a noisy mobile platform, i.e., a drone. To this end, we expose the limitations of signal processing techniques to deal with loud drone noise. We demonstrate how soundproofing can be used to design acoustic sensing for drones.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100738"},"PeriodicalIF":6.3,"publicationDate":"2024-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143182768","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 : 2024-12-29DOI: 10.1016/j.atech.2024.100749
Praveen Sankarasubramanian
The rapid advancement in communication technology has sparked a transformative wave across various domains, significantly enhancing comfort and convenience in daily life. Addressing the escalating global demand for food, coupled with the need to alleviate the efforts of farmers, technology, particularly the Internet of Things (IoT), has emerged as a pivotal force. Precisely predicting variations in climatestrictures, ground conditions, and dirt attributes has emerged as a formidable challenge in the realm of agricultural IoT. In this paper, we introduce a smart optimal prediction model for sensors based on IoT-enabled precision agriculture. Initially, we enhance the THAM index (temperature, humidity, air- and water-quality measurement) by using the modified Wild Geese (MWG) algorithm to predict environmental conditions accurately. The deployment of IoT sensor nodes using quantum deep reinforcement learning (QDRL) to determine the idealamount of devices required for effective coverage of the target agricultural field to improving communication. Furthermore, we compute the production yield rate, consider various attributes such as fertilizer regulatory measures, temperature quotient, and agronomy by using the improved prairie dog optimization (IPDO) algorithm. Finally, we assess the performance of MWG-QDRL-IPDO model using test samples collected from the Meteorology Bureau through the related sensor middleware. Our findings reveal a checking efficacy of 96.35 %, even with a reduced amount of devices covering a hugezone. Similarly, the accuracy of IoT sensor node deployment reaches 91.47 %, contributive to reduce the irrelevant data generation and processing time.
{"title":"Enhancing precision in agriculture: A smart predictive model for optimal sensor selection through IoT integration","authors":"Praveen Sankarasubramanian","doi":"10.1016/j.atech.2024.100749","DOIUrl":"10.1016/j.atech.2024.100749","url":null,"abstract":"<div><div>The rapid advancement in communication technology has sparked a transformative wave across various domains, significantly enhancing comfort and convenience in daily life. Addressing the escalating global demand for food, coupled with the need to alleviate the efforts of farmers, technology, particularly the Internet of Things (IoT), has emerged as a pivotal force. Precisely predicting variations in climatestrictures, ground conditions, and dirt attributes has emerged as a formidable challenge in the realm of agricultural IoT. In this paper, we introduce a smart optimal prediction model for sensors based on IoT-enabled precision agriculture. Initially, we enhance the THAM index (temperature, humidity, air- and water-quality measurement) by using the modified Wild Geese (MWG) algorithm to predict environmental conditions accurately. The deployment of IoT sensor nodes using quantum deep reinforcement learning (QDRL) to determine the idealamount of devices required for effective coverage of the target agricultural field to improving communication. Furthermore, we compute the production yield rate, consider various attributes such as fertilizer regulatory measures, temperature quotient, and agronomy by using the improved prairie dog optimization (IPDO) algorithm. Finally, we assess the performance of MWG-QDRL-IPDO model using test samples collected from the Meteorology Bureau through the related sensor middleware. Our findings reveal a checking efficacy of 96.35 %, even with a reduced amount of devices covering a hugezone. Similarly, the accuracy of IoT sensor node deployment reaches 91.47 %, contributive to reduce the irrelevant data generation and processing time.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100749"},"PeriodicalIF":6.3,"publicationDate":"2024-12-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143181122","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 : 2024-12-28DOI: 10.1016/j.atech.2024.100763
Dimas Firmanda Al Riza , Lucky Candra Musahada , Romzi Izzudin Aufa , Mochamad Bagus Hermanto , Hermawan Nugroho , Yusuf Hendrawan
Fruits detection and counting is an important task for yield prediction that could be achieved by computer vision. The ability to locate and count the fruits could also help the harvesting robot to do a picking task. YOLO is one of the deep learning models which is popular and widely used for object detection and has good performance in detection speed and precision. In the citrus counting task, the label could be set as a single label or multi-label which shows different citrus maturity. The performance of the deep learning model could be different with a different number of labels. Furthermore, there are several types of YOLOv7 models with different sizes and purposes which could also have different performances to do a similar task. This study aims to compare the performance of different kinds of YOLOv7-based deep learning models for citrus fruit detection and counting. Case study on the citrus cv. Batu 55 trees have been carried out. The results show that the original YOLOv7 achieved the best performance both on single and double labels compared to the tiny and X versions of YOLOv7. The YOLOv7 could reach mAP50 of 0.906, a precision of 0.85, a sensitivity of 0.825, and an F1-score of 0.837, while for the counting task, the model has a good performance with R2 of 0.966.
{"title":"Comparative study of citrus fruits (Citrus reticulata Blanco cv. Batu 55) detection and counting with single and double labels based on convolutional neural network using YOLOv7","authors":"Dimas Firmanda Al Riza , Lucky Candra Musahada , Romzi Izzudin Aufa , Mochamad Bagus Hermanto , Hermawan Nugroho , Yusuf Hendrawan","doi":"10.1016/j.atech.2024.100763","DOIUrl":"10.1016/j.atech.2024.100763","url":null,"abstract":"<div><div>Fruits detection and counting is an important task for yield prediction that could be achieved by computer vision. The ability to locate and count the fruits could also help the harvesting robot to do a picking task. YOLO is one of the deep learning models which is popular and widely used for object detection and has good performance in detection speed and precision. In the citrus counting task, the label could be set as a single label or multi-label which shows different citrus maturity. The performance of the deep learning model could be different with a different number of labels. Furthermore, there are several types of YOLOv7 models with different sizes and purposes which could also have different performances to do a similar task. This study aims to compare the performance of different kinds of YOLOv7-based deep learning models for citrus fruit detection and counting. Case study on the citrus cv. Batu 55 trees have been carried out. The results show that the original YOLOv7 achieved the best performance both on single and double labels compared to the tiny and X versions of YOLOv7. The YOLOv7 could reach mAP50 of 0.906, a precision of 0.85, a sensitivity of 0.825, and an F1-score of 0.837, while for the counting task, the model has a good performance with R<sup>2</sup> of 0.966.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100763"},"PeriodicalIF":6.3,"publicationDate":"2024-12-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143182763","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 : 2024-12-28DOI: 10.1016/j.atech.2024.100764
Connor C. Mullins, Travis J. Esau, Qamar U. Zaman, Patrick J. Hennessy
This research developed a workflow to assess the viability of AI-generated imagery in training machine learning models for detecting ripe wild blueberries (Vaccinium angustifolium Ait.), hair fescue weeds (Festuca filiformis Pourr.), and red leaf disease (Exobasidium vaccinii). Ground truth images were collected and augmented with AI-generated variations using DALL-E 2 to expand the dataset. Models were trained on three datasets: ground truth, generated, and a combination (40% generated images). Evaluation metrics included precision, recall, mAP50, and mAP50–95, analyzed using ANOVA multiple mean comparisons and Tukey's HSD test (α = 0.05). For ripe wild blueberries, combination models achieved the highest performance across all metrics (mAP50: 0.834), significantly outperforming the ground truth model (mAP50: 0.806) in terms of mAP50–95 (0.478 compared to 0.424). For hair fescue weeds, the combination dataset outperformed others with the highest mAP50 (0.983), closely followed by the ground truth dataset (mAP50: 0.969). In detecting red leaf disease, the combination dataset showed the best performance (mAP50: 0.848 ± 0.140, mAP50–95: 0.607 ± 0.219), compared to the ground truth (mAP50: 0.615 ± 0.092, mAP50–95: 0.417 ± 0.045) and generated datasets (mAP50: 0.245 ± 0.088, mAP50–95: 0.144 ± 0.059). Models trained solely on generated images showed significantly lower performance across all categories except the precision metric for red leaf, where performance was comparable to ground truth. This indicated that while AI-generated images can augment datasets and improve generalization, they cannot fully replace ground truth data while maintaining model performance. Integrating AI-generated images with real-world data significantly improved model performance, reduced labor-intensive data collection processes, and provided a more diverse and comprehensive dataset for training, underscoring the importance of a balanced approach to optimizing data collection protocols for wild blueberry cultivation.
{"title":"Optimizing data collection requirements for machine learning models in wild blueberry automation through the application of DALL-E 2","authors":"Connor C. Mullins, Travis J. Esau, Qamar U. Zaman, Patrick J. Hennessy","doi":"10.1016/j.atech.2024.100764","DOIUrl":"10.1016/j.atech.2024.100764","url":null,"abstract":"<div><div>This research developed a workflow to assess the viability of AI-generated imagery in training machine learning models for detecting ripe wild blueberries (<em>Vaccinium angustifolium</em> Ait.), hair fescue weeds (<em>Festuca filiformis</em> Pourr.), and red leaf disease (<em>Exobasidium vaccinii</em>). Ground truth images were collected and augmented with AI-generated variations using DALL-E 2 to expand the dataset. Models were trained on three datasets: ground truth, generated, and a combination (40% generated images). Evaluation metrics included precision, recall, mAP<sub>50</sub>, and mAP<sub>50–95</sub>, analyzed using ANOVA multiple mean comparisons and Tukey's HSD test (α = 0.05). For ripe wild blueberries, combination models achieved the highest performance across all metrics (mAP<sub>50</sub>: 0.834), significantly outperforming the ground truth model (mAP<sub>50</sub>: 0.806) in terms of mAP<sub>50–95</sub> (0.478 compared to 0.424). For hair fescue weeds, the combination dataset outperformed others with the highest mAP<sub>50</sub> (0.983), closely followed by the ground truth dataset (mAP<sub>50</sub>: 0.969). In detecting red leaf disease, the combination dataset showed the best performance (mAP<sub>50</sub>: 0.848 ± 0.140, mAP<sub>50–95</sub>: 0.607 ± 0.219), compared to the ground truth (mAP<sub>50</sub>: 0.615 ± 0.092, mAP<sub>50–95</sub>: 0.417 ± 0.045) and generated datasets (mAP<sub>50</sub>: 0.245 ± 0.088, mAP<sub>50–95</sub>: 0.144 ± 0.059). Models trained solely on generated images showed significantly lower performance across all categories except the precision metric for red leaf, where performance was comparable to ground truth. This indicated that while AI-generated images can augment datasets and improve generalization, they cannot fully replace ground truth data while maintaining model performance. Integrating AI-generated images with real-world data significantly improved model performance, reduced labor-intensive data collection processes, and provided a more diverse and comprehensive dataset for training, underscoring the importance of a balanced approach to optimizing data collection protocols for wild blueberry cultivation.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100764"},"PeriodicalIF":6.3,"publicationDate":"2024-12-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143182766","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 : 2024-12-27DOI: 10.1016/j.atech.2024.100758
Lianghai Yang , Jing Yan , Xinyue Cao , Huiru Li , Binjie Ge , JiaXin He , Zhechen Qi , Xiaoling Yan
Invasive weed seeds pose a huge threat to local ecosystems, and it is of great significance to accurately classify invasive weed seeds. Leveraging the rapid advancements in deep learning, various methods have become potential solutions to this problem. In this study, we constructed a large dataset of invasive weed seeds in China and proposed a novel approach to address the identification of species caused by the high similarity among species within the same genus, utilizing Hierarchical Bilinear Pooling (HBP) with ResNet50 as the backbone network. To validate the efficacy of our method, we conducted comparative experiments with classic models in the field of fine-grained recognition. Our evaluation encompassed overall benchmark performance, classification for similar species within the genus, and the classification of species of different sizes. The results demonstrated the HBP-ResNet50 model achieved an outstanding overall benchmark performance accuracy of 99.1 %. Even in Amaranthus and Euphorbia which have highly similar seed morphology, it can achieve high accuracy of 97.94 % and 96.19 %, respectively. The model achieved high accuracy across different sizes of seeds, especially reaching an astonishing 99.18 % in the medium size (1–5 mm). These exceptional results establish the superior performance of HBP-ResNet50. This research has greatly improved the detection efficiency and accuracy, helps curtailing the proliferation of invasive weed seeds, and reduces damage to agricultural ecosystems and economic property losses. The success of our work encourages the future application of this method in the classification of plants, insects, and other relevant fields.
{"title":"High-accuracy classification of invasive weed seeds with highly similar morphologies: Utilizing hierarchical bilinear pooling for fine-grained image classification","authors":"Lianghai Yang , Jing Yan , Xinyue Cao , Huiru Li , Binjie Ge , JiaXin He , Zhechen Qi , Xiaoling Yan","doi":"10.1016/j.atech.2024.100758","DOIUrl":"10.1016/j.atech.2024.100758","url":null,"abstract":"<div><div>Invasive weed seeds pose a huge threat to local ecosystems, and it is of great significance to accurately classify invasive weed seeds. Leveraging the rapid advancements in deep learning, various methods have become potential solutions to this problem. In this study, we constructed a large dataset of invasive weed seeds in China and proposed a novel approach to address the identification of species caused by the high similarity among species within the same genus, utilizing Hierarchical Bilinear Pooling (HBP) with ResNet50 as the backbone network. To validate the efficacy of our method, we conducted comparative experiments with classic models in the field of fine-grained recognition. Our evaluation encompassed overall benchmark performance, classification for similar species within the genus, and the classification of species of different sizes. The results demonstrated the HBP-ResNet50 model achieved an outstanding overall benchmark performance accuracy of 99.1 %. Even in <em>Amaranthus</em> and <em>Euphorbia</em> which have highly similar seed morphology, it can achieve high accuracy of 97.94 % and 96.19 %, respectively. The model achieved high accuracy across different sizes of seeds, especially reaching an astonishing 99.18 % in the medium size (1–5 mm). These exceptional results establish the superior performance of HBP-ResNet50. This research has greatly improved the detection efficiency and accuracy, helps curtailing the proliferation of invasive weed seeds, and reduces damage to agricultural ecosystems and economic property losses. The success of our work encourages the future application of this method in the classification of plants, insects, and other relevant fields.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100758"},"PeriodicalIF":6.3,"publicationDate":"2024-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143181124","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 : 2024-12-27DOI: 10.1016/j.atech.2024.100748
Kwee Kim Teo , Nurul Fazmidar Binti Mohd Noor , Shivakumara Palaiahnakote , Mohamad Nizam Bin Ayub
Oil palm tree detection and classification from coconut and banana trees is vital for increasing the production of oil palm businesses globally, particularly in Malaysia. Since oil palm, coconut, and banana trees share common characteristics such as tree shape and structure, classification is challenging. Further, this work considers images captured by drones, which adds complexity to the classification problem. Unlike most existing methods that primarily detect oil palm trees, the proposed work aims to detect and classify multiple tree types. Inspired by the success of the Segment Anything Model (SAM), a generalized model for object segmentation, we adapted SAM for detecting and localizing oil palm, coconut, and banana trees in drone images. Similarly, motivated by the efficiency and effective feature extraction of EfficientNetB3, we integrated it for the classification task. The proposed model combines SAM for detection and EfficientNetB3 for classification in an end-to-end architecture. To evaluate its performance, we conducted experiments on a dataset collected from a Malaysian drone services company, featuring frames captured across diverse locations. Results demonstrate that the proposed method significantly outperforms state-of-the-art approaches. For detection, the proposed SAM achieves F1-scores of 97 %, 89 %, and 91 % for oil palm, coconut, and banana trees, respectively. For classification, the proposed model achieves F1 scores of 92 %, 88 %, and 91 % for oil palm, coconut, and banana trees, respectively. The results show that the proposed method is superior to the existing methods.
{"title":"An efficientnet-based model for classification of oil palm, coconut and banana trees in drone images","authors":"Kwee Kim Teo , Nurul Fazmidar Binti Mohd Noor , Shivakumara Palaiahnakote , Mohamad Nizam Bin Ayub","doi":"10.1016/j.atech.2024.100748","DOIUrl":"10.1016/j.atech.2024.100748","url":null,"abstract":"<div><div>Oil palm tree detection and classification from coconut and banana trees is vital for increasing the production of oil palm businesses globally, particularly in Malaysia. Since oil palm, coconut, and banana trees share common characteristics such as tree shape and structure, classification is challenging. Further, this work considers images captured by drones, which adds complexity to the classification problem. Unlike most existing methods that primarily detect oil palm trees, the proposed work aims to detect and classify multiple tree types. Inspired by the success of the Segment Anything Model (SAM), a generalized model for object segmentation, we adapted SAM for detecting and localizing oil palm, coconut, and banana trees in drone images. Similarly, motivated by the efficiency and effective feature extraction of EfficientNetB3, we integrated it for the classification task. The proposed model combines SAM for detection and EfficientNetB3 for classification in an end-to-end architecture. To evaluate its performance, we conducted experiments on a dataset collected from a Malaysian drone services company, featuring frames captured across diverse locations. Results demonstrate that the proposed method significantly outperforms state-of-the-art approaches. For detection, the proposed SAM achieves F1-scores of 97 %, 89 %, and 91 % for oil palm, coconut, and banana trees, respectively. For classification, the proposed model achieves F1 scores of 92 %, 88 %, and 91 % for oil palm, coconut, and banana trees, respectively. The results show that the proposed method is superior to the existing methods.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100748"},"PeriodicalIF":6.3,"publicationDate":"2024-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143182761","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}
Smallholder farmers are critical to global food production and natural resource management. Due to increased competition for water resources and variability in rainfall due to climate change, chronic irrigation water scarcity is rising particularly in drought-prone regions. Improving the awareness of climatic risk to yields and incomes is critical to sustainable agricultural intensification. However, adopting a new technology represents a certain level of risk for the farmers, who invest time and economic resources in changing their practices. We have developed a mobile application, currently for cotton, that would allow farmers to actualize the risk of growing cotton. By implementing a sociohydrological dynamic model with a kernel principal component analysis structural error model, the software provides a risk forecast of the yield and profit the user can expect at the end of the season. The mobile app not only processes social and agricultural information provided by the user but also retrieves and continually updates climate datasets from the web, as well as market prices. The users can request the execution of the sociohydrological model to the servers from their own mobile devices. By following an agile methodology, the mobile app has been tested with ∼100 farmers in order to get feedback from real users; this brought the opportunity to redesign the functionality based on the correct understanding of information and, a fast and clear management of the tool and helping in the adoption of the technology. This was combined with existing knowledge around communicating risk by using multiple modes of communication - text, graphics, sound and video - all of which were implemented to reinforce the knowledge communicated and ensure sufficient redundancy. This turned out to be beneficial for farmers with low prior knowledge and higher acceptability of the mobile app by the users as evidenced through feedback rounds with them. This study exemplifies an approach to address the gap in communicating risks in agriculture using a user-friendly mobile application.
{"title":"Makara: A tool for cotton farmers to evaluate risk to income","authors":"Mario Alberto Ponce-Pacheco , Soham Adla , Ramesh Guntha , Aiswarya Aravindakshan , Maya Presannakumar , Ashray Tyagi , Anukool Nagi , Prashant Pastore , Saket Pande","doi":"10.1016/j.atech.2024.100759","DOIUrl":"10.1016/j.atech.2024.100759","url":null,"abstract":"<div><div>Smallholder farmers are critical to global food production and natural resource management. Due to increased competition for water resources and variability in rainfall due to climate change, chronic irrigation water scarcity is rising particularly in drought-prone regions. Improving the awareness of climatic risk to yields and incomes is critical to sustainable agricultural intensification. However, adopting a new technology represents a certain level of risk for the farmers, who invest time and economic resources in changing their practices. We have developed a mobile application, currently for cotton, that would allow farmers to actualize the risk of growing cotton. By implementing a sociohydrological dynamic model with a kernel principal component analysis structural error model, the software provides a risk forecast of the yield and profit the user can expect at the end of the season. The mobile app not only processes social and agricultural information provided by the user but also retrieves and continually updates climate datasets from the web, as well as market prices. The users can request the execution of the sociohydrological model to the servers from their own mobile devices. By following an agile methodology, the mobile app has been tested with ∼100 farmers in order to get feedback from real users; this brought the opportunity to redesign the functionality based on the correct understanding of information and, a fast and clear management of the tool and helping in the adoption of the technology. This was combined with existing knowledge around communicating risk by using multiple modes of communication - text, graphics, sound and video - all of which were implemented to reinforce the knowledge communicated and ensure sufficient redundancy. This turned out to be beneficial for farmers with low prior knowledge and higher acceptability of the mobile app by the users as evidenced through feedback rounds with them. This study exemplifies an approach to address the gap in communicating risks in agriculture using a user-friendly mobile application.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100759"},"PeriodicalIF":6.3,"publicationDate":"2024-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143181950","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 : 2024-12-27DOI: 10.1016/j.atech.2024.100756
Wenli Xue , Ewaut Kissel , András Tóth , Raphael Pilloni , Vincent Vadez
Sorghum genotypes vary in their response to higher sowing density, but the traits explaining these variations are unknown. In the present study, a 3D-imaging based approach identified the phenotypic traits responsible for the genetic variation in sorghum's response to high sowing density. Twenty sorghum genotypes, some varying in their response to density, were grown and 3D-images were collected weekly between weeks 4–6. From these images, 80 phenotypic traits, including 33 architectural and 47 multispectral, were extracted. The within-genotype means of these 80 traits, and two indicators of the sowing density response (Biomass ratio (Br) and Transpiration ratio (Tr)), measured in a previous study with 13 common genotypes, were used in a Spearman correlation analysis. Seventeen and four traits were strongly correlated with Br and Tr, respectively. The majority of these traits, predominantly architectural, strongly suggest that, under high sowing density, a fuller light interception, having more leaf area in the lower canopy, lead to a larger Br, while more vertically aligned leaves favour larger Tr values, which related to higher water use efficiency in another study. Furthermore, a Principal Component Analysis (PCA) indicated traits contributing to better photosynthesis could be used to estimate Br. Similarly, a combination of traits relating to leaf angle were good indicators of the genetic variation in Tr values. These results provide insights about the strategies some sorghum genotypes have developed to thrive under higher sowing density and that could be used as biomarkers for the breeding of density-resistant cultivars.
{"title":"Identifying phenotypic markers explaining positive sorghum response to sowing density using 3D-imaging","authors":"Wenli Xue , Ewaut Kissel , András Tóth , Raphael Pilloni , Vincent Vadez","doi":"10.1016/j.atech.2024.100756","DOIUrl":"10.1016/j.atech.2024.100756","url":null,"abstract":"<div><div>Sorghum genotypes vary in their response to higher sowing density, but the traits explaining these variations are unknown. In the present study, a 3D-imaging based approach identified the phenotypic traits responsible for the genetic variation in sorghum's response to high sowing density. Twenty sorghum genotypes, some varying in their response to density, were grown and 3D-images were collected weekly between weeks 4–6. From these images, 80 phenotypic traits, including 33 architectural and 47 multispectral, were extracted. The within-genotype means of these 80 traits, and two indicators of the sowing density response (Biomass ratio (Br) and Transpiration ratio (Tr)), measured in a previous study with 13 common genotypes, were used in a Spearman correlation analysis. Seventeen and four traits were strongly correlated with Br and Tr, respectively. The majority of these traits, predominantly architectural, strongly suggest that, under high sowing density, a fuller light interception, having more leaf area in the lower canopy, lead to a larger Br, while more vertically aligned leaves favour larger Tr values, which related to higher water use efficiency in another study. Furthermore, a Principal Component Analysis (PCA) indicated traits contributing to better photosynthesis could be used to estimate Br. Similarly, a combination of traits relating to leaf angle were good indicators of the genetic variation in Tr values. These results provide insights about the strategies some sorghum genotypes have developed to thrive under higher sowing density and that could be used as biomarkers for the breeding of density-resistant cultivars.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100756"},"PeriodicalIF":6.3,"publicationDate":"2024-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143181130","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 : 2024-12-27DOI: 10.1016/j.atech.2024.100739
Darwin Laura , Elsa Pilar Urrutia , Franklin Salazar , Jeanette Ureña , Rodrigo Moreno , Gustavo Machado , Maria Cazorla-Logroño , Santiago Altamirano
Broccoli is one of Ecuador's main agricultural products and is exported worldwide. To ensure high-quality production, routine inspections are necessary to counteract pathogens and diseases. This study presents an aerial remote sensing system to monitor broccoli crops using pre-programmed flight plans to assess crop health and enable timely treatments. The system leverages the YOLO v5x algorithm for deep learning under various production conditions. An autonomous drone, equipped with GPS for grid flight planning, captures high-definition images every 2 seconds. These images, tagged with geolocation data, are processed through a Python-based graphical interface. The results are stored in a database to improve the system's accuracy in detecting false positives and negatives.
The aerial remote sensing system successfully monitored broccoli crops, identifying areas affected by pathogens and diseases. The YOLO v5x algorithm demonstrated high accuracy in image analysis, reducing false detections. The system's autonomous drone efficiently covered large crop areas, providing precise geolocation data for targeted interventions. The collected data, stored in a centralized database, facilitated continuous improvement of the detection algorithm, ensuring reliable pathogen control and maintaining high production quality.
{"title":"Aerial remote sensing system to control pathogens and diseases in broccoli crops with the use of artificial vision","authors":"Darwin Laura , Elsa Pilar Urrutia , Franklin Salazar , Jeanette Ureña , Rodrigo Moreno , Gustavo Machado , Maria Cazorla-Logroño , Santiago Altamirano","doi":"10.1016/j.atech.2024.100739","DOIUrl":"10.1016/j.atech.2024.100739","url":null,"abstract":"<div><div>Broccoli is one of Ecuador's main agricultural products and is exported worldwide. To ensure high-quality production, routine inspections are necessary to counteract pathogens and diseases. This study presents an aerial remote sensing system to monitor broccoli crops using pre-programmed flight plans to assess crop health and enable timely treatments. The system leverages the YOLO v5x algorithm for deep learning under various production conditions. An autonomous drone, equipped with GPS for grid flight planning, captures high-definition images every 2 seconds. These images, tagged with geolocation data, are processed through a Python-based graphical interface. The results are stored in a database to improve the system's accuracy in detecting false positives and negatives.</div><div>The aerial remote sensing system successfully monitored broccoli crops, identifying areas affected by pathogens and diseases. The YOLO v5x algorithm demonstrated high accuracy in image analysis, reducing false detections. The system's autonomous drone efficiently covered large crop areas, providing precise geolocation data for targeted interventions. The collected data, stored in a centralized database, facilitated continuous improvement of the detection algorithm, ensuring reliable pathogen control and maintaining high production quality.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100739"},"PeriodicalIF":6.3,"publicationDate":"2024-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143181126","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 : 2024-12-27DOI: 10.1016/j.atech.2024.100762
Simone Figorilli , Lavinia Moscovini , Simone Vasta , Francesco Tocci , Simona Violino , Dyan Abraham , Solomon Pascal , Kelvin Benjamin , Roberto Sandoval , Raisa Spencer , Corrado Costa , Antonio Scarfone , Luciano Ortenzi , Federico Pallottino
Recently banana plantations have been affected by the Black Sigatoka Disease (BSD), producing streaks, lesions and yellow and brown spots on the leaves until the appearance of entire dead parts. The disease causes reductions in yield making it essential to assess infection by monitoring plants status and implementing agronomical measures. This work aims to develop a physical field device to identify the BSD presence. It consists in a 3D printed prototype embedding a smartphone acquiring and processing banana leaves images. An advanced Artificial Intelligence model was trained and implemented for real-time processing. The algorithm is a Convolutional Neural Network (CNN) able to classify the samples into 6 classes representative of different BSD stages infection. The trained model, showing an accuracy of 82 % in training and 78 % in validation, was integrated into a specifically developed mobile application for field use. The Android app allows to acquire, identify the georeferenced infection stage, sync all to a remote dedicated host from which the results can be mapped and exported to a .csv file for easy data management. The distinction between healthy and diseased leaves can be achieved using the Smart BSD device for real-time acquisition, establishing the right intervention strategy.
{"title":"Smart IoT device for in field Black Sigatoka Disease recognition and mapping","authors":"Simone Figorilli , Lavinia Moscovini , Simone Vasta , Francesco Tocci , Simona Violino , Dyan Abraham , Solomon Pascal , Kelvin Benjamin , Roberto Sandoval , Raisa Spencer , Corrado Costa , Antonio Scarfone , Luciano Ortenzi , Federico Pallottino","doi":"10.1016/j.atech.2024.100762","DOIUrl":"10.1016/j.atech.2024.100762","url":null,"abstract":"<div><div>Recently banana plantations have been affected by the Black Sigatoka Disease (BSD), producing streaks, lesions and yellow and brown spots on the leaves until the appearance of entire dead parts. The disease causes reductions in yield making it essential to assess infection by monitoring plants status and implementing agronomical measures. This work aims to develop a physical field device to identify the BSD presence. It consists in a 3D printed prototype embedding a smartphone acquiring and processing banana leaves images. An advanced Artificial Intelligence model was trained and implemented for real-time processing. The algorithm is a Convolutional Neural Network (CNN) able to classify the samples into 6 classes representative of different BSD stages infection. The trained model, showing an accuracy of 82 % in training and 78 % in validation, was integrated into a specifically developed mobile application for field use. The Android app allows to acquire, identify the georeferenced infection stage, sync all to a remote dedicated host from which the results can be mapped and exported to a .csv file for easy data management. The distinction between healthy and diseased leaves can be achieved using the Smart BSD device for real-time acquisition, establishing the right intervention strategy.</div></div>","PeriodicalId":74813,"journal":{"name":"Smart agricultural technology","volume":"10 ","pages":"Article 100762"},"PeriodicalIF":6.3,"publicationDate":"2024-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143181515","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