Precision Agriculture最新文献

Purpose

Automated plot extraction in agronomic research field trials is essential for high-throughput phenotyping and precision agriculture. Accurate delineation of plot boundaries enables reliable crop type classification, yield estimation, and crop health monitoring. However, traditional plot extraction methods rely heavily on manual digitization, which is time-consuming, labor-intensive, and prone to inconsistencies. This study aims to develop a Segment Anything Model (SAM)-based framework that automates plot extraction while maintaining high accuracy across diverse agricultural field conditions.

Methods

The proposed framework consists of mask generation, plot orientation estimation, and plot refinement. SAM is leveraged to generate plot masks, which are subsequently filtered and refined to ensure precise boundary delineation. The method is designed to function without the need for model training or fine-tuning, making it highly adaptable across different datasets.

Results

The framework was validated on five datasets, demonstrating robust performance under varying field conditions. The pixel-based evaluation yielded an average F1 score of 89.54%. For polygon-based evaluation, the framework achieved 99.71% precision at IoU=50% and an average precision of 68.51% across IoU thresholds from 50 to 95%, confirming its ability to accurately extract plot boundaries. A Canopeo-based regression analysis further demonstrated that the extracted plots provide more reliable phenotypic estimates compared to manually digitized ground reference data.

Conclusions

The proposed framework significantly reduces manual effort while ensuring high precision and scalability for large-scale phenotyping applications. By relying solely on RGB imagery and zero-shot segmentation, it enhances accessibility for real-world agricultural research. Future work will focus on extending the framework to irregular plot structures, diverse crop types, and computational optimizations for large-scale implementation.

The majority of newly developed sprayers now feature advanced capabilities, allowing herbicide application with centimeter-level precision, potentially reducing herbicide use by up to 90%. However, accurately identifying the precise locations to spray, known as the application map, remains a significant research challenge. Recently, both commercial providers and research institutions have proposed various drone-based methods for generating application maps. Despite these advancements, practical adoption is limited, primarily due to regulatory constraints and the high costs associated with the technology. A promising approach to increasing the adoption of these technologies lies in the utilization of more cost-effective hardware solutions. In this paper, we introduce and evaluate a novel detection method specifically designed for identifying Rumex obtusifolius (sorrel) and for automatically generating application maps that are compatible with most GNSS-enabled sprayers. To this end, we present a new metric for treatment success, termed the treatment F1-score, and conduct a comparative analysis of the performance of the DJI Mini 2 and the DJI Matrice 350 RTK using our proposed system, achieving treatment F1-scores of 0.61% and 0.65% , respectively. The ability of this system to deliver good performance utilizing significantly less expensive hardware than typically employed in similar applications suggests a potential for broader adoption, particularly given the unexpectedly modest performance gap of only 4 percentage points in the treatment F1-score. Under controlled experimental conditions, we observed reductions in herbicide use of up to 97% without missing any targets. In practical applications within real-world meadows, a 40% reduction in herbicide consumption was achieved with a treatment accuracy of 85% . These findings underscore the substantial potential for future technological advancements. The standalone object detector achieves a mean Average Precision (mAP) of 67.4% and an F1-score of 62%, demonstrating robust performance even on out-of-distribution drone data collected by other researchers. Still, the performance of the object detection algorithm is identified as a critical bottleneck in the system. To facilitate further research and development in this domain, we have made our training dataset available for download.

Spot spraying represents an efficient and sustainable method for reducing herbicide use in agriculture. Reliable differentiation between crops and weeds, including species-level classification, is essential for real-time application. This study compares state-of-the-art object detection models-YOLOv8, YOLOv9, YOLOv10, and RT-DETR-using 5611 images from 16 plant species. Two datasets were created, dataset 1 with training all 16 species individually and dataset 2 with grouping weeds into monocotyledonous weeds, dicotyledonous weeds, and three chosen crops. Results indicate that all models perform similarly, but YOLOv9s and YOLOv9e, exhibit strong recall (66.58 % and 72.36 %) and mAP50 (73.52 % and 79.86 %), and mAP50-95 (43.82 % and 47.00 %) in dataset 2. RT-DETR-l, excels in precision reaching 82.44 % (dataset 1) and 81.46 % (dataset 2) making it ideal for minimizing false positives. In dataset 2, YOLOv9c attains a precision of 84.76% for dicots and 78.22% recall for Zea mays L.. Inference times highlight smaller YOLO models (YOLOv8n, YOLOv9t, and YOLOv10n) as the fastest, reaching 7.64 ms (dataset 1) on an NVIDIA GeForce RTX 4090 GPU, with CPU inference times increasing significantly. These findings emphasize the trade-off between model size, accuracy, and hardware suitability for real-time agricultural applications.

Current approaches for monitoring soil organic matter (SOM) exhibit limitations in long-term predictive accuracy and data efficiency. This study aims to develop a remote sensing framework that integrating Landsat imagery and three modeling algorithms (PLSR, RF, Cubist) to address these challenges, reduce sampling workload, and enable large scale soil fertility assessments. Feature selection via Boruta and recursive feature elimination (RFE) was applied to optimize model performance, with PLSR identified astheoptimal algorithm. The framework utilized long-term Landsat imagery (2007–2021) and an inter-annual migration learning approach to map SOM dynamics. PLSR achieved cross-year SOM prediction (R² = 0.51, RMSE = 3.97 g/kg), enabling accurate mapping of non-sample years with minimal field data and long-term imagery. Analysis of SOM trends revealed a decade-long decline in the study area, strongly correlated with land-use intensity. The proposed inter-annual migration learning method demonstrates that SOM dynamics can be efficiently tracked using sparse sampling and time-series remote sensing, offering a scalable tool for soil fertility management and precision agriculture.

Timely and accurate monitoring of plant nitrogen concentration (PNC) is vital for optimizing field N management. Hyperspectral indices are commonly used as a predictor for monitoring the PNC of crops, but individual spectral indices are often susceptible to cultivars and growth stages. Machine learning (ML) is a promising method for mining more spectral variables to assess the PNC of crops. To monitor the PNC of potatoes, therefore, this study extended previous work to further use hyperspectral optimized spectral indices (OSI) as input variables of ML, while, comparing with the ML models that used full-spectrum (FS), existing spectral indices (ESI) and sensitive spectral bands (SSB) as input variables, as well as simple regression model based on OSI alone. The partial least squares regression (PLSR), random forest (RF), support vector regression (SVR), and artificial neural network (ANN) models were calibrated using a dataset encompassing three cultivars and critical fertigation growth stages under three to six N levels. The calibrated ML models were evaluated using the datasets from independent experiments and two farmers´ fields. The OSI as an input variable in ML models showed superiority for predicting the potato PNC compared to FS, SSB, and ESI. The OSI-based RF model with an R² of 0.79, RMSE of 0.27%, and RPD of 2.18 had higher accuracy for predicting potato PNC than other ML models. Comparing the simple optimized spectral indices regression model alone, the OSI-based RF model reduced RMSE by mitigating the effects of cultivars and growth stages on PNC prediction. The OSI-based RF model significantly contributes to optimum fertilization management based on actual potato N status during critical growth periods.

The Małopolskie and Podkarpackie provinces in Poland are characterized by many small farms with many small, scattered fields. This farm structure is labeled “agrarian fragmentation”. Using digital technologies in such small farm areas is usually a challenge. However, there are several digital technologies that, with minimal financial investment, can yield results in the form of improved resource management and agricultural production processes, as well as data-driven decision-making. The overall objective of this analysis is to determine the limitations of using digital technologies in farms operating in areas with high agrarian fragmentation. In addition, the aim was also to identify the differences in the potential for implementing individual digital solutions depending on farm size and activity type conducted in the surveyed area. A survey was conducted by the Paper and Pen Personal Interview (PAPI) method, in which 389 farmers took part. Research showed that the technologies most commonly used in the study area include applications recognizing plant diseases and applications supporting decision-making. The use of advanced digital tools related to precision agriculture and the automation of crop production was very rare. Farm size, the age of the farmer managing the farm, and the number of farm activities were significant factors that increased the probability of implementing digital technologies. The main barriers to their implementation were a lack of sufficient knowledge and trust. The implementation of digital technologies in small farms requires actions aimed at increasing farmer knowledge. Meanwhile, designing new digital solutions must take the specific regional conditions into account, such as geographical factors or the limited investment capacity of farms.

Efficient water management is critical for sustainable agriculture in Mediterranean climates, particularly in super-high-density (SHD) olive orchards where water scarcity poses significant challenges. This study assessed the potential of UAV-based thermal and multispectral imagery to monitor crop water status and predict yield under different regulated deficit irrigation (RDI) strategies. Conducted over two seasons (2018–2019) in a commercial SHD olive orchard (Olea europaea L., cv. ‘Arbequina’) in Villena, Spain, the experiment involved four irrigation treatments ranging from full irrigation (FI) to progressively restricted RDIs. UAV flights captured thermal infrared and multispectral imagery at key phenological stages, to calculate Crop Water Stress Index (CWSI) and Normalized Difference Vegetation Index (NDVI), which were validated against plant-based measurements of stem water potential (Ψ_stem). The results demonstrated that thermal parameters, including canopy temperature and CWSI, effectively identified water stress levels, although their sensitivity was influenced by environmental conditions and sensor limitations. NDVI proved to be a reliable indicator of vegetative growth and yield, with values closely linked to irrigation levels and fruit load. The approach incorporating both canopy and soil reflectance (NDVI_crop+ground) provided the most accurate assessment of crop performance. These findings highlight the value of UAV-based remote sensing technologies for optimizing irrigation management in SHD olive orchards, particularly under deficit irrigation regimes. However, further advancements in sensor accuracy and index normalization are recommended to enhance their applicability and precision in agricultural practices.

The deep integration of artificial intelligence technology and agriculture has significantly propelled the rapid development of smart agriculture. However, the field still faces numerous challenges, including high algorithm complexity and limited detection speed in farmland environments. To address the challenges encountered by corn spraying robots in navigating and identifying lines, we have proposed a corn crop row navigation line recognition algorithm based on the LT-YOLOv10s model. By introducing lightweight network models (GhosNet), efficient feature pyramid models (SPPFA), and efficient feature attention modules (PSCA) into the YOLOv10s network, we have reduced the complexity of the model and significantly enhanced the detection efficiency of corn plants. Then, the algorithm precisely locates corn plants using the center points of detection boxes and accurately fits crop rows using the least squares method. Finally, the navigation lines centered on the corn crop rows are determined through the adjacent centerline method. Experimental data significantly demonstrates that the comprehensive performance of the LT-YOLOv10s model surpasses industry benchmark models such as YOLOv5s, YOLOv7, YOLOv8s, YOLOv9s, and the traditional YOLOv10s. The proposed algorithm for extracting the center navigation line of corn crop rows boasts an average fitting time of just 26ms with an accuracy rate of up to 93.8%, ensuring precision and reliability in navigation line extraction. This provides robust technical support for precise navigation of corn-spraying robots.

Purpose

In contemporary agriculture, achieving sustainable food production while preserving the environment is crucial. Traditional uniform rate nitrogen fertilization (URNF) often leads to over- or under-applications of N in fields with negative economic, agronomic and environmental issues. Variable rate nitrogen fertilization (VRNF) has shown promise in optimizing N application by accounting for soil and crop variability, thus improving nitrogen use efficiency and reducing environmental impact. This study evaluates and compares two VRNF solutions in two wheat fields in Belgium and France.

Methods

The first, VRNF1 relied on onsite measurement of soil nitrate using ion-selective electrode (ISE) sensors, whereas the second, VRNF2, utilizes the fusion of on-line measured key soil properties using a visible and near-infrared spectrometer (vis-NIRS) and crop normalized difference vegetation index (NDVI). In VRNF1, soil nitrate values were used to rank the fertility level of management zones (MZs), delineated by the clustering analysis of vis-NIRS-NDVI data (like for VRNF2), with N fertilization rates adjusted by 30–50%, applying lower rates to high-fertility zones and higher rates to low-fertility zones. In VRNF2, after the fertility level of MZ was ranked by examining the on-line measurements of pH, organic carbon (OC), moisture content (MC), potassium (K), phosphorus (P), and calcium (Ca), and crop NDVI, N fertilizer rates were adjusted similarly to VRNF1.

Results

A cost-benefit analysis revealed that the gross margin of both VRNF solutions was larger than that of the URNF, with VRNF1 providing up to 289 EUR ha⁻¹ and VRNF2 up to 358 EUR ha⁻¹ more gross margin than URNF. VRNF1 increased crop yield by up to 8%, while VRNF2 resulted in a 9.2% yield increase compared to URNF. However, VRNF1 achieved a slight economic advantage (14 EUR ha⁻¹) in one field, while VRNF2 was more profitable in the other field by 69 EUR ha⁻¹. Additionally, VRNF2 demonstrated superior environmental benefits, using 14% less fertilizer than URNF and 12% less than VRNF1.

Conclusion

Overall, VRNF2 offered better economic and environmental outcomes than VRNF1 and URNF. However, the subjectivity of ranking MZs into different fertility levels in the absence of historical yield data for the VRNF2 raises concerns, calling in such a situation for VRNF1 to be adopted in future VRNF schemes.

Assessing spatial and temporal variations in crop water stress is vital for precision irrigation. This study utilized Unmanned Aerial Vehicles (UAVs) equipped with multispectral (MSS) and thermal band (TB) sensors to map the crop water stress index (CWSI) in wheat. A water deficit experiment was conducted on winter wheat under varying irrigation levels during late vegetative, reproductive, and maturation stages. CWSI was calculated using canopy temperature, ambient air temperature, and vapor pressure deficit (VPD). Six machine learning (ML) models—linear model (LM), random forest (RF), decision tree (DT), support vector machine (SVM), extreme gradient boosting (XGB), and artificial neural network (ANN)—were developed for pre-heading, post-heading, and seasonal datasets. The top five vegetation indices (VIs), selected using Recursive Feature Elimination (RFE), along with thermal data, were used as inputs to the ML models. Results showed that seasonal ML models outperformed those based only on pre-heading or post-heading data. Particularly, the RF model performed well, with respective R² and RMSE values of 0.87 and 0.09 for seasonal, 0.82 and 0.05 for pre-heading, and 0.93 and 0.06 for post-heading datasets. SHapley Additive exPlanations (SHAP) analysis identified Red Normalized Value (RNV), TB, and Green Red Vegetation Index (GRVI) as key predictors of CWSI in the RF model. CWSI maps effectively captured spatial variations in water stress, aligning with irrigation management practices. This study demonstrates the effectiveness of combining UAV remote sensing and ML for precision irrigation management.