Smart agricultural technology最新文献

The research results presented in this paper were obtained as part of the D2KAB project (Data to Knowledge in Agriculture and Biodiversity) which aims to develop semantic web-based tools to describe and make agronomical data actionable and accessible following the FAIR principles. We focus on constructing domain-specific Knowledge Graphs (KGs) from textual data sources, using Natural Language Processing (NLP) techniques to extract and structure relevant entities. Our approach is based on the formalization of a semantic data model using common linked open vocabularies such as the Web Annotation Ontology (OA) and the Provenance Ontology (PROV). The model was developed by formulating motivating scenarios and competency questions from domain experts. This model has been used to construct three different KGs from three distinct corpora: PubMed scientific publications on wheat and rice genetics and phenotyping, and French agricultural alert bulletins. The named entities to be recognized include genes, phenotypes, traits, genetic markers, taxa and phenological stages normalized using semantic resources such as the Wheat Trait and Phenotype Ontology (WTO), the French Crop Usage (FCU) thesaurus and the Plant Phenological Description Ontology (PPDO). Named entities were extracted using different NLP approaches and tools. The relevance of the semantic model was validated by implementing experts questions as SPARQL queries to be answered on the constructed RDF knowledge graphs. Our work demonstrates how domain-specific vocabularies and systematic querying of KGs can reveal hidden interactions and support agronomists in navigating vast amounts of data. The resources and transformation pipelines developed are publicly available in Git repositories.

Segment Anything Model (SAM) is a new “foundation model” that can be used as a zero-shot object segmentation method with the use of either guide prompts such as bounding boxes, polygons, or points. Alternatively, additional post processing steps can be used to identify objects of interest after segmenting everything in an image. Here a method is presented using segment anything together with a series of post processing steps to segment potato leaves, called Leaf Only SAM. The advantage of this proposed method is that it does not require any training data to produce its results so has many applications across the field of plant phenotyping where there is limited high quality annotated data available. The performance of Leaf Only SAM is compared to a Mask R-CNN model which has been fine-tuned on a small novel potato leaf dataset. On the evaluation dataset, Leaf Only SAM finds an average recall of 73.1 and an average precision of 73.9, compared to recall of 87.6 and precision of 84.4 for Mask R-CNN. Leaf Only SAM does not perform better than the fine-tuned Mask R-CNN model on the potato leaf dataset, but the SAM based model does not require any extra training or annotation. This shows there is potential to use SAM as a zero-shot classifier with the addition of post processing steps.

Image analysis is being developed to improve the efficiency of fishery and aquaculture technologies. Optical cameras are an easy and cost-effective method for monitoring fish and other species. In this study, a monitoring system that combines an underwater time-lapse camera and a deep learning-based image analysis was developed for utilization in integrated multi-trophic aquaculture (IMTA). The sea cucumber (Apostichopus japonicus) was used as a target species because the technology necessary for estimating growth, particularly in terms of weight, of caged sea cucumber using an underwater environment is still under study. Therefore, semantic segmentation was applied to classify the images into caged sea cucumbers and various underwater backgrounds. Multiple images of sea cucumbers were captured in a water tank that mimicked the box cage used in IMTA, and their body weights were measured simultaneously. For model development, approximately 1,300 images were prepared for the training and validation processes. The model then achieved an IoU (Intersection over Union) of approximately 94 % for the validation data. Next, the pixel numbers of sea cucumbers were converted into an area calculated using the size of the cage net as the background. The relationship between the area and weight of sea cucumbers yielded an approximate line for estimating body weight. As a result, the approximation line had a coefficient of determination of R² = 0.87 for training and validation data and RMSE (Root Mean Square Error) =1.81 and 6.78 g for sea cucumbers less than 10 and 110 g, respectively. Using the model, test images in an actual IMTA situation were applied, and the estimated body weights were close to the measured values for small sea cucumbers. If we apply this model to images obtained over an extended period, the growth of sea cucumbers in a time series can be understood.

In the field of precision agriculture, there is a significant shift towards data-driven methodologies that considerably enhance the efficiency and sustainability of agricultural operations. This research investigates the application of CAN-Bus and GNSS data to develop a comprehensive analysis of agricultural machinery's operational states during ploughing operations through advanced data analytics techniques, including machine learning. The primary tool utilized in this study is the Random Forest classifier, a robust algorithm well-suited for handling the complexity and volume of data typical in modern agricultural settings. The study evaluates Random Forest models trained on various feature subsets to accurately identify different operational states of agricultural machinery, including idle, moving, turning, and working states. By merging CAN-Bus data, which capture real-time operational parameters, with GNSS data, providing spatial and temporal context, it is possible to achieve a comprehensive understanding of machinery behaviour and its interaction with field conditions. This integration significantly enhances decision-making capabilities in farm management, leading to more effective and efficient operations.

Furthermore, the findings from this study contribute to the broader agricultural community by illustrating how data-driven approaches can harness the vast amounts of data generated by modern agricultural machinery. This research underscores the potential of machine learning modelsnot only to interpret complex data sets but also to transform these insights into actionable knowledge, which can lead to more precise and sustainable agricultural practices. Overall, this study offers a systematic approach for analysing agricultural data and lays the groundwork for future advancements in incorporating machine learning and IoT technologies into the agricultural sector. This aims to enhance productivity and sustainability in farming practices.

Remote sensing and machine learning are widely used to estimate crop yield. The use of these technologies for yield estimation of bulbous vegetables is challenging because the yield is underground and can't be directly monitored by remote sensing images. Among the bulbous vegetables, garlic (Allium sativum L.) is one of the most widely cultivated in the world. The aim of this study was to develop an accurate and transferable machine learning model to monitor and to estimate garlic yield using unmanned aerial vehicle (UAV) multispectral images. Data were collected over three growing seasons (2021, 2022, and 2023) at four different garlic phenological phases (202, 405, 407, and 409 of BBCH). The random forest (RF) algorithm was used to estimate the garlic yield by comparing two different training feature sets: the vegetation indices (VIs) and the VIs with the addition of the texture features extracted from the UAV images. The most important VIs were selected using the recursive feature elimination algorithm. Two estimation methods were compared: a direct bulb estimation and an indirect bulb estimation using the aboveground biomass as a proxy. To evaluate the transferability of the RF models, two cross-validation strategies were compared: a nested leave-one-fold-out cross-validation (LOFO_CV) and a leave-one-year-out cross-validation (LOYO_CV). The best performance was achieved by the direct bulb estimation using the LOFO_CV strategy. Regarding the transferability of the RF models between years (i.e. LOYO_CV), the indirect estimation method showed a higher transferability than the direct estimation method. Finally, the addition of texture features improved the accuracy of the RF models, but in general, their contribution was poor. This study demonstrated that the yield of bulbous vegetables can be accurately estimated by remote sensing, and that UAVs are a suitable tool to provide rapid and reliable support for garlic yield monitoring.

Despite having made much improvement in sensing, automation, and control, the current broadcast spraying system has several drawbacks, such as uneven coverage, excessive chemical use, and deviation from recommended dosage. Typically, farmers use large self-propelled sprayers to spray the entire field without knowledge of spatial pest severity, potentially resulting in an unintentional application. However, application errors and the extent of chemical use can be optimized by utilizing an intelligent site-specific decision-based sprayer to control pests more efficiently. Hence, the initial project goal was to design a robotic liquid application system for row crops (e.g., sorghum and corn) and validate sprayer system performance. The critical design considerations for the spray application system were modularity; the ability to be mounted on an autonomous platform to go within 76.2-cm spaced crops; spray on either side of the crop row using spray booms; onboard hardware and software for control and data acquisition; and record as-applied data. A system with desired design requirements was built and individual sub-systems were tested under simulated lab scenarios to quantify the response time and accuracy of the spray system. The results showed that the sprayer could maintain an average system pressure within ±5% of the target under different duty cycles for each of the six nozzles. At 40% duty cycle, the nozzle pressure settling time at an error margin of ±5% from the mean was 13 ms, 20 ms, and 19 ms, for one, three, and six nozzles, respectively. Also, no substantial pressure difference was observed between nozzles installed at different heights in two different booms. Therefore, this application system could be a viable solution for autonomous platforms to site-specifically apply pesticides only on critically infested plants, has the potential to decrease the overall input costs on chemicals and reduce the negative environmental impacts.

A sensor-based automatic watering unit was developed to irrigate the paddy seedling trays. The watering unit comprised six flat spray nozzles with a direct current (DC) water pump and flow sensor, and these were kept at the top of the conveyor frame, which was actuated by a DC water pump with a solenoid valve to drop the water in trays. The actuation of the DC solenoid valve was controlled by a limit switch fitted to the conveyor frame, which sensed the moving trays on the conveyor. The performance of the developed unit was evaluated, and process optimization was performed using the Inscribed Central Composite Design in Response Surface Methodology. The effect of main operating parameters- chain conveyor speed (0.123, 0.196, and 0.27 m/s), spray operating pressure (1.5, 3.0, and 4.5 kg/cm²), and spray height (10,30 and 50 cm) on the performance of the sensor-based automatic watering unit was studied in terms of the amount of water spray requirement and coefficient of water spray uniformity in trays. The responses of the unit at its optimal parameter operation (spray operating pressure 4.5 kg/cm², spray height 10 cm, and chain conveyor speed 0.14 m/s) were found to be 1.31 L/tray as the amount of water spray requirement and 92.41 % as the coefficient of water spray uniformity of the unit. Hence, it can be an efficient substitute with significantly less power consumption 41 W with a higher working capacity of 780 trays/h for watering into the paddy tray nursery. The developed sensor-based automatic smart watering unit saves irrigation water per square area of 68.90 %, 76.13 %, and 82.31 % compared to existing drip, sprinkler, and flood irrigation methods in mat-type paddy nurseries. The developed sensor-based automatic smart watering unit saves the watering cost INR 11,567.16 (140.02 USD) per hectare, reducing cost by 63.7 % compared to manual watering in the mat-type nursery preparation. Thus, an automatic smart watering unit in a paddy tray nursery offers consistent and uniform watering from an Indian perspective, promoting uniform germination, crop growth, and establishment. It conserves water by ensuring the right amount of water is applied at the right time, avoiding overwatering or under-watering.

Grasslands represent a key element of agroecosystems for sustainable food systems. A better understanding of the grazing behaviour of domestic herbivores is essential to support innovations for grassland management and define grazing practices that support rather than enter into conflict with biodiversity. A key component of the grazing process is the grass-severing bite by which the herbivore collects forage from a pasture. How often, where, and when such bites are performed are relevant indicators of the grazing behaviour of cattle and could be used as indicators to guide farmers in pasture management. In this work, we developed a methodology to create a Machine Learning (ML) model for identifying grass-severing bite events from the Inertial Measurement Unit (IMU) signals of a sensor placed on the neck of cows. The two-phase process consisted of classifying every period of behaviour of cattle into two mutually exclusive behaviours: “ingestion” and “other” (phase 1), and then counting the number of bites taken during each period classified as “ingestion” (phase 2). Seven dry red-pied Holstein cattle and two Blonde d'Aquitaine x Belgian White and Blue cross-breds were observed. A total of 39 h and 25 min of video were recorded and tagged for the different behaviours to train several ML algorithms. During phase 1, four different window segmentations and two different splits of the data were used to train and test four ML classification algorithms: Bagged Tree, Medium k-NN, Fine tree and linear SVM. The results show that Bagged Tree algorithms with 30 s windows and 90 % overlap gave the best results during the first phase, with an accuracy of 97.83 % for split 1 and 98.07 % for split 2. During phase 2, the same four window segmentations as for phase 1 were used, to test regression algorithms to quantify the number of bites taken during each time-window. Two machine learning algorithms were tested: Bagged Tree and Medium NN, on 5 sessions of 30 min. The sessions ranged between 0 % and 94 % of ingestion time. Phase 2 results showed that Bagged Tree regression algorithms with 10 s windows and 90 % overlap performed the best, with an average RMSE of 1.83 for the tested value and an error percentage of -1.93 % and 0 % for the session with 94 % or 0 % of ingestion time, and between +15.06 % and +26.97 % of error for sessions where the animal alternates frequently between both behaviours. The data and code used in this study are openly available on a public depository

The purpose of agriculture is to support humankind. There are currently 7.7 billion people on the planet and this figure will increase to nine billion by 2050. As the population grows, even greater amounts of food will be needed, creating a significant challenge for farmers. Emerging digital technologies such as hyper-automation have the potential to revolutionize conventional agricultural methods. This study assessed the current use of hyper-automation systems in agriculture and examined whether new uses of this technology could benefit agricultural industries. One example could be to use an automated variable-seed control system, which has reported seeding accuracy of 98 %, indicating a cost-effective solution. Overall, our analysis revealed that to sustain future agricultural production and ensure food security, countries throughout the world need to focus on hyper-automation in the agriculture sector.

The global food supply heavily depends on utilizing fertilizers to meet production goals. The adverse impacts of traditional fertilization practices on the environment have necessitated the exploration of new alternatives in the form of smart fertilizer technologies (SFTs). This review seeks to categorize SFTs, which are slow and controlled-release Fertilizers (SCRFs), nano fertilizers, and biological fertilizers, and describes their operational principles. It examines the environmental implications of conventional fertilizers and outlines the attributes of SFTs that effectively address these concerns. The findings demonstrate a pronounced environmental advantage of SFTs, including enhanced crop yields, minimized nutrient loss, improved nutrient use efficiency, and reduced greenhouse gas (GHG) emissions. Nevertheless, amidst these benefits, the challenges and constraints associated with these technologies, such as production expenses and potential environmental impacts of specific components, are also discussed. A comparative assessment of these SFTs emphasizes the importance of a balanced approach, considering three crucial factors: efficiency, environmental safety, and cost-effectiveness. While no single SFT achieves optimal balance across these dimensions, integrating multiple fertilizer technologies may help mitigate individual drawbacks. Also, financial and cost-to-benefit analyses are essential to gauge their applicability across diverse cropping environments. Future perspectives shed light on emerging SFTs and innovative approaches to overcome prevailing challenges and cultivate a more impactful role in fostering sustainable agriculture.