Information Processing in Agriculture最新文献

As the new generation of artificial intelligence (AI) continues to evolve, weather big data and statistical machine learning (SML) technologies complement each other and are deeply integrated to significantly improve the processing and forecasting accuracy of fishery weather. Accurate fishery weather services play a crucial role in fishery production, serving as a great safeguard for economic benefits and personal safety, enabling fishermen to carry out fishery production better, and contributing to the sustainable development of the fishery industry. The objective of this paper is to offer an understanding of the present state of research and development in SML technology for simulating and forecasting fishery weather. Specifically, we analyze the current state of research and technical features of SML in weather and summarize the applications of SML in simulation and forecasting of fishery weather, which mainly include three aspects: fishery weather scenario generation, fishery weather forecasting, and fishery extreme weather warning. We also illustrate the main technical means and principles of SML technology. Finally, we summarize the most advanced SML fields and provide an outlook on their application value in the field of fishery weather.

Compared with manual sorting of citrus fruit, vision-based sorting solutions can help achieve higher accuracy and efficiency. In this study, we present a vision system based on CNN-LSTM, which can cooperate with robotic grippers for real-time sorting and is readily applicable to various citrus processing plants. A CNN-based detector was adopted to detect the defective oranges in view and temporarily classify them into corresponding types, and an LSTM-based predictor was used to predict the position of the oranges in a future frame based on image sequential data. The fusion of CNN and LSTM networks enabled the system to track defective ones during rotation and identify their true types, and their future path was also predicted which is vital for predictive control of visually guided robotic grasping. High detection accuracy of 94.1% was obtained based on experimental results, and the error for path prediction was within 4.33 pixels 40 frames later. The average time to process a frame was between 28 and 62 frames per second, which also satisfied real-time performance. The results proved the potential of the proposed system for automated citrus sorting with good precision and efficiency, and it can be readily extended to other fruit crops featuring high versatility.

Preconsolidation pressure (${\sigma }_P$) of soil can be considered as an indicator of the Load Bearing Capacity (LBC), which is the tolerated surface pressure before compaction, often caused by the traffic of agricultural machinery. In this pioneering study, a remote sensing approach was introduced to estimate LBC through ${\sigma }_P$ from soils of the “Rio Preto” Hydrographic Basin, Bahia State, Brazil, in a monthly time lapse from 2016 to 2019. Traditionally, ${\sigma }_P$ is measured by a laborious and time demanding laboratory analysis, making it unfeasible to map large areas. The innovative methodology of this work consists of combining active–passive satellite data on soil moisture and pedotransfer functions of clay content and water matric potential to obtain geo-located estimates of ${\sigma }_P$. Estimates were analysed under different classes of soil use, land cover and slope; 95% confidence intervals were built for the time series of mean values of LBC for each class. The overall seasonal variation in LBC estimates is similar in areas with annual crops, grasslands and natural vegetation, and flat areas are less affected by soil moisture variations over the year (between seasons). LBC decreased, in general, at about 0.5% a year in flat areas. Therefore, these areas demand attention, since they occupy 86% of the Basin and are mostly subjected to agricultural soil management and surface pressure by heavy machinery.

The droplet deposition is a key index to evaluate the quality of unmanned aerial vehicle (UAV) spraying. The detection of the droplet deposition is time-consuming and costly, therefore, it is difficult to achieve large-scale and rapid acquisition in the field. To solve the above problems, a droplet deposition acquisition system (DDAS) was developed. It was composed of the multiple sensors, processing units, remote server database and Android-based software. A droplet deposition prediction model based on field experimental data was established by using a one-dimensional convolutional neural network (1D-CNN) algorithm, and the effects of different inputs on the prediction ability of the model were analyzed. The results showed that adding temperature and humidity data to the inputs can achieve higher prediction accuracy than only using UAV spraying operation parameters and wind speed data as the inputs to the model. In addition, the prediction accuracy of the 1D-CNN model was the highest when compared with other models such as back propagation neural network, multiple correlation vector machine, and multiple linear regression. The 1D-CNN model was embedded into the DDAS, and the evaluation experiments were carried out in the field. The correlation analysis was conducted between two datasets of the droplet deposition obtained by the DDAS and water sensitive paper (WSP), respectively. The R² was 0.924, and the RMSE was 0.026 μL/cm². It is proved that the droplet deposition values obtained by the DDAS and WSP have high consistency, and the DDAS developed can provide an auxiliary solution for the intelligent evaluation of UAV spraying quality.

The advent of remote livestock monitoring systems provides numerous possibilities for improving on-farm productivity, efficiency, and welfare. One potential application for these systems is for the detection of calving events. This study describes the integration of data from multiple sensor sources, including accelerometers, global navigation satellite systems (GNSS), an accelerometer-derived rumination algorithm, a walk-over-weigh unit, and a weather station for parturition detection using a support vector machine approach. The best performing model utilised data from GNSS, the ruminating algorithm, and weather stations to achieve 98.6% accuracy, with 88.5% sensitivity and 100% specificity. The top-ranking features of this model were primarily GNSS derived. This study provides an overview as to how various sensor systems could be integrated on-farm to maximise calving detection for improved production and welfare outcomes.

Due to the worldwide population growth and the increasing needs for sugar-based products, accurate estimation of sugarcane biomass is critical to the precise monitoring of sugarcane growth. This research aims to find the imperative predictors correspond to the random and fixed effects to improve the accuracy of wet and dry sugarcane biomass estimations by integrating ground data and multi-temporal images from Unmanned Aerial Vehicles (UAVs). The multispectral images and biomass measurements were obtained at different sugarcane growth stages from 12 plots with three nitrogen fertilizer treatments. Individual spectral bands and different combinations of the plots, growth stages, and nitrogen fertilizer treatments were investigated to address the issue of selecting the correct fixed and random effects for the modelling. A model selection strategy was applied to obtain the optimum fixed effects and their proportional contribution. The results showed that utilizing Green, Blue, and Near Infrared spectral bands on models rather than all bands improved model performance for wet and dry biomass estimates. Additionally, the combination of plots and growth stages outperformed all the candidates of random effects. The proposed model outperformed the Multiple Linear Regression (MLR), Generalized Linear Model (GLM), and Generalized Additive Model (GAM) for wet and dry sugarcane biomass, with coefficients of determination (R²) of 0.93 and 0.97, and Root Mean Square Error (RMSE) of 12.78 and 2.57 t/ha, respectively. This study indicates that the proposed model can accurately estimate sugarcane biomasses without relying on nitrogen fertilizers or the saturation/senescence problem of Vegetation Indices (VIs) in mature growth stages.

As a complement to traditional estimates of stem dimensions from numerical models, terrestrial light detection and ranging (Lidar) provides direct stem diameter and volume values using cylindrical models constructed from point clouds. This study used two approaches to estimate total stem volume using Lidar and compared them with two empirical equations, one used by the Forest Inventory Analysis in the Pacific Northwest (FIA-PNW) and one based on a taper equation. We fitted point clouds of 10 Douglas-fir with three sets of cylinder models that are distinguished by their segment length (i.e. 0.5 m, 1 m, and 2 m), then developed three taper equations based on the point-cloud-based diameter estimated previously. We estimated the total stem volume of the tree with eight models: six-point cloud-based (i.e. three taper and three cylinders) and two empirical. Finally, we used simulations to extrapolate the volume estimations of various methods for different diameters at breast height (DBH) classes. We found that all the point-cloud-based taper equations were similar in their performance ($R^2=0.94$, RMSE = 4.6 cm) and produced mean volume estimates greater than mean estimates of the existing models. The cylinder models produced 11–16% greater mean volume estimates than the FIA-PNW estimate, with the 0.5 m segment length producing the greatest values, followed by the 1 m and 2 m segment length. The simulated data suggested that the mean volume estimates of a given DBH class are different when using different computation methods. ANOVA revealed a combined effect of the computation methods and the DBH class on the mean volume estimates. We conclude that the point-cloud-based taper equations, after being symmetrically calibrated, would be consistent with the regional stem volume estimates, whereas the cylinder models would be better in estimating individual stem volume. When constructing Lidar-based cylinder models in future applications, cylinder segment length would need to be adjusted to the length and DBH of the stem, as well as to the objectives of the research.