Precision Agriculture最新文献

Diversifying high-input, monocropped landscapes like the US Corn Belt would provide both economic and ecosystem service benefits to the agricultural landscape. Decision support systems (DSS) and digital agriculture could help farmers decide if diversification is suitable for their operation. However, adoption of DSS by farmers remains low, likely due to lack of farmer engagement before and during the DSS development process. This study aimed to better understand the tasks, tools, and people involved in implementing farmland diversification with the goal to inform design of agricultural DSS. Semi-structured interviews were conducted with 11 farmers who had diversified their corn/soybean cropland with government-supported conservation programs (e.g., CRP, wetlands) and alternative crops (e.g., small grains, pasture) in the past four years. Interview data was transcribed and then analyzed using affinity diagramming. Results show farmers needed DSS to layer multiple sources of data and observations over several years to identify field productivity trends and drivers; spatial orientation of practices to fit management and field constraints; matching operation goals to alternative practices; financial planning and market exploration; and information on promising emerging practices like subsidized pollinator habitat. However, the interviews also highlighted structural barriers to diversification that DSS cannot or can only partially address. These included social pressures; market access; crop insurance policy; and quality of relationships with governmental agencies. Results indicate better DSS design can empower individual farmers to diversify cropland, but structural interventions will be needed to successfully diversify the agricultural landscape and support economic and ecosystem health.

The COVID-19 pandemic has highlighted the vulnerabilities of the global food system, underscoring the need for a sustainable transformation of the food system. With the advent of new digital technologies emerging as critical tools for achieving the agricultural shift, it is important to understand farmers’ adoption decisions better. This study aims to systematically uncover and delineate the varied forms of experiences farmers have with new digital technologies and investigate how these experiences impact the organizational adoption decisions on the farm. In this study, twenty interviews with apple growers, wine makers, and intermediaries from a German region encompassing Saxony, Thuringia, and Saxony–Anhalt were conducted and analyzed. Through the lens of the modified adaptive capacity wheel and alongside the interview data, five relevant types of experiences were identified. These types of experiences are closely related to farmers’ adaptation motivation (AM) and adaptation belief (AB), potentially influencing their future decisions about the adoption of digital technologies. This study highlights the importance of creating meaningful experiences with technologies to strengthen farmers’ AM and AB.

Leaf area index (LAI) is a vital input variable for crop growth and yield prediction models. Therefore, rapid and accurate crop LAI estimates can offer important information for monitoring and managing the quantity and quality of food production. Here, LAI values of maize and soybean were predicted applying height metrics and intensity metrics calculated through unmanned aerial vehicle (UAV) LiDAR data. Moreover, we compared the prediction performance of physical model with that of empirical model for estimating crop LAI. The physical model based on Beer–Lambert law yielded reliable estimation results using LiDAR height data (maize: R² = 0.815, RMSE = 0.385; soybean: R² = 0.627, RMSE = 0.515) and LiDAR intensity data (maize: R² = 0.719, RMSE = 0.474; soybean: R² = 0.548, RMSE = 0.567). However, the linear regression model obtained a higher estimation accuracy. The single linear regression model derived from LiDAR height data had an R² value of 0.837 (RMSE = 0.361) for maize and 0.658 (RMSE = 0.493) for soybean, and derived from LiDAR intensity data had an R² value of 0.749 (RMSE = 0.448) for maize and 0.460 (RMSE = 0.619) for soybean, respectively. We found that the random forest (RF) regression model yielded the lowest estimation accuracy in this study. Moreover, the RF regression model in our study was not able to reliably estimate soybean LAI whether using LiDAR height metrics (R² = 0.294) or intensity metrics (R² = 0.180). Our results show that both LiDAR intensity and height metrics are capable of reliably predicting maize and soybean LAIs, although LiDAR intensity data yielded lower estimation accuracy than LiDAR height data. In conclusion, the results presented in this study demonstrate that using UAV-LiDAR technology to predict crop LAI is a flexible, practical, and cost-effective method.

The success of weed control is critical for our food security. Non-chemical weed control is a promising technique in sustainable agriculture to ensure the food security. In this review, multiple directed energy weed control methods are reviewed with a specific focus on laser and optical radiation weed control. The mechanisms of the weed control in terms of adverse ablation, radiation thermal effects, and molecular-level damages are systematically reviewed. In particular, the underlying mathematical models determining the dose and response relationship of the weed control are also analyzed for a rigorous study of the physical law of the control process. Challenges of applying the techniques into practice are also illustrated to guide practical weed control applications.

Effective yield forecasting is a key strategy for adaptation when facing food loss to climate variability. Currently, solar-induced chlorophyll fluorescence (SIF) is an emerging remote-sensing index owing to its high relevance to plant photosynthesis, and sensitivity to drought. Despite many studies have focused on drought monitoring and production assessment by SIF, little puts it into practice for in-season yield prediction. In this study, we combined multi-source satellite and meteorological data, especially coupling with subseasonal-to-seasonal (S2S) dynamic atmospheric prediction climate model (IAP-CAS FGOALS-f2), with an addition of SIF, to predict maize yields in the U.S. Corn Belt, based on the developed machine learning dynamical hybrid model (MHCF). By comparison, we found that SIF performed well in the correlation analysis with yield, with average correlations up to 0.719 in August. Then we utilized different algorithms, different models (S2S data for MHCF, climate data for the Benchmark), and different input combinations to train and predict maize yields. All four algorithms using SIF significantly improved prediction performance. S2S + VIs + SIF combination (FGOALS-f2、NDVI、EVI、SIF) can achieve the best performance, while the XGBoost algorithm reached 0.897 of R². With the best combination, it can achieve 4 months before maize harvest (with R² value of 0.85, and RMSE < 13 bu/acre). In 2012, the year had a severe drought, although predictive capability decreased in all the predictions, the models with SIF still maintained robust and improved the prediction (improved R² by 5.92%, and RMSE decreased by 18.08% of XGBoost). According to the study, it can be expected, the combination of MHCF and SIF will play a greater role in subseasonal yield prediction. We also provide an operational proposition of hybrid yield forecasting method to fully integrating climate prediction and machine learning for early notice of crop production losses.

Smart spray technology developed at the University of Florida was designed to reduce off-target applications when applying postemergence (POST) herbicides for weed control in plasticulture systems. A trial was conducted in the fall of 2021 and spring of 2022 to evaluate smart spray technology in row middles in a banana pepper field at the Gulf Coast Research and Education Center in Balm, FL. The objective of this study was to evaluate the efficacy of targeted POST-herbicide applications in plasticulture pepper row middles in the presence or absence of a pre-emergent (PRE) herbicide. Flumioxazin reduced broadleaf and overall weed densities in both seasons and lowered grass density in the spring. Two targeted applications reduced the nutsedge density in spring compared to the two banded applications. No significant pepper damage was observed in any treatments. Applied POST herbicide volume following PRE-herbicide was reduced by 84% and 54% for fall and spring respectively. In the absence of a PRE herbicide, targeted applications reduced POST-herbicide volumes by 30% and 45% for fall and spring respectively. No reduction in weed control or pepper yield was observed when comparing targeted with banded applications. Overall, the use of smart spray technology for POST herbicides in row middles reduced applied spray volume with no reduction in weed control, significant injuries on pepper, or negative effects on yield.

Uniform plant spacing, seeding depth, and emergence are important factors heavily influenced by both machine settings and soil conditions. Understanding load distribution across the planter toolbar at varying planter settings and soil conditions provide feedback to improve planter performance and achieve desired seed placement consistency. One important soil property that affects opening disc load requirement in creating seed trench is soil texture which relates to soil strength. However, none of the existing methods (soil apparent electrical conductivity (ECa) maps, historic soil maps, and cone penetrometer) provide accurate soil strength data on a high spatial resolution which could be used to optimize planter performance. This study was conducted to (1) quantify the percentage of time row-planters need uplift during planting and (2) quantify opening disc loads using real-time machine control system recorded data across different ECa zones. Results showed that uplift events varied from 13 to 18% with wing and track sections revealed higher instances of uplift. Higher instances of uplift were observed on the non-track section for planter with wing wheels. Results revealed a modest correlation between soil ECa and opening disc load with 435 N more or 12% higher opening disc load applied on high soil ECa zones as compared in low soil ECa zones.

Consumer-grade cameras have emerged as a cost-effective alternative to conventional scientific cameras in precision agriculture applications. However, there is a lack of information on their appropriate use and calibration. This study focused on developing practical methodologies for determining optimal camera settings and converting image digital numbers (DNs) to reflectance. Two Nikon D7100 and two Nikon D850 cameras with visible and near-infrared (NIR) sensitivity were deployed on both manned and unmanned aircraft for image acquisition. To optimize camera settings, including exposure time and aperture, an approach that considered flight parameters and image histograms was employed. Linear and nonlinear regression analyses based on multiple nonlinear models were performed to accurately characterize the reflectance-DN relationship across all four bands (blue, green, red and NIR) based on seven calibration tarps. The results revealed that the exponential model with vertical translation was the optimal model for reflectance conversion for both camera types. Based on the optimized camera parameters and the optimal model type, this study provided an extensive analysis of the models and their root mean square errors (RMSE) derived from all 952 possible 2- to 6-tarp combinations for all bands in both camera types. This analysis led to the selection of optimal tarp combinations based on the desired level of accuracy for each of the five multi-tarp configurations. As the number of tarps increased to 4, 5, or 6, the RMSE values stabilized for all bands, indicating 4-tarp combinations were the optimal choice. These findings hold significant practical implications for practitioners in precision agriculture seeking guidance for configuring consumer-grade cameras effectively while ensuring accurate reflectance conversion.

Soil fertility is one of the most critical bases for high productivity and sustainability in crop production. Within-field heterogeneity is often problematic in both crop management practices and crop productivity. Besides, appropriate soil management practices leads to the effective carbon sequestration. Since the soil carbon content (SCC) is the most simple and effective indicator of soil fertility, accurate and high-resolution mapping of SCC is an essential basis for addressing these issues. Here, we developed a tractor-based hyperspectral sensing system for speedy and accurate mapping of SCC. A new hybrid spectral algorithm linking normalized difference spectral index (h-NDSI) and machine learning proved superior. Appropriate algorithms were implemented to generate diagnostic map and prescription map from SCC map for the variable-rate application of pellet manure. The field performance of the sensing/mapping system was tested in the farmers' fields in the Fukushima region of Japan where the within-field heterogeneity of soil fertility was disastrous due to the decontamination after the nuclear power-plant disaster. The structure and functioning of the system proved promising. Moreover, the spatial simulation by linking the SCC data and a dynamic simulation model clearly showed the significant impact of variable-rate application of pellet manure on the chronosequential change of SCC, within-field heterogeneity, and carbon stock. The systematic linkage of the sensing/mapping system with the variable-rate spreader and dynamic simulation model would be effective for improving soil fertility and soil carbon stock. Applicability of the system will be extended through an extensive validation of the predictive models.

The precision monitoring of film-mulched winter wheat growth facilitates field management optimization and further improves yield. Unmanned aerial vehicle (UAV) is an effective tool for crop monitoring at the field scale. However, due to the interference of background effects caused by soil and mulch, achieving accurate monitoring of crop growth in complex backgrounds for UAV remains a challenge. Additionally, the simultaneous inversion of multiple growth parameters helped us to comprehensively monitor the overall crop growth status. This study conducted field experiments including three winter wheat mulching treatments: ridge mulching, ridge–furrow full-mulching, and flat cropping full-mulching. Three machine learning algorithms (partial least squares, ridge regression, and support vector machines) and deep neural network were employed to process the vegetation indices (VIs) feature data, and the residual neural network 50 (ResNet 50) was used to process the image data. Then the two modalities (VI feature data and image data) were fused to obtain a multi-modal fusion (MMF) model. Meanwhile, a film-mulched winter wheat growth monitoring model that simultaneously predicted leaf area index (LAI), aboveground biomass (AGB), plant height (PH), and leaf chlorophyll content (LCC) was constructed by coupling multi-task learning techniques. The results showed that the image-based ResNet 50 outperformed the VI feature-based model. The MMF improved prediction accuracy for LAI, AGB, PH, and LCC with coefficients of determination of 0.73–0.92, mean absolute errors of 0.29–3.89 and relative root mean square errors of 9.48–12.99%. A multi-task MMF model with the same loss weight distribution ([1/4, 1/4, 1/4, 1/4]) achieved comparable accuracy to the single-task MMF model, improving training efficiency and providing excellent generalization to different film-mulched sample areas. The novel technique of the multi-task MMF model proposed in this study provides an accurate and comprehensive method for monitoring the growth status of film-mulched winter wheat.