Wheat accounts for more than 50% of Australia’s total grain production. The capability to generate accurate in-season yield predictions is important across all components of the agricultural value chain. The literature on wheat yield prediction has motivated the need for more novel works evaluating machine learning techniques such as random forests (RF) at multiple scales. This research applied a Random Forest Regression (RFR) technique to build regional and local-scale yield prediction models at the pixel level for three southeast Australian wheat-growing paddocks, each located in Victoria (VIC), New South Wales (NSW) and South Australia (SA) using 2018 yield maps from data supplied by collaborating farmers. Time-series Normalized Difference Vegetation Index (NDVI) data derived from Planet’s high spatio-temporal resolution imagery, meteorological variables and yield data were used to train, test and validate the models at pixel level using Python libraries for (a) regional-scale three-paddock composite and (b) individual paddocks. The composite region-wide RF model prediction for the three paddocks performed well (R² = 0.86, RMSE = 0.18 t ha⁻¹). RF models for individual paddocks in VIC (R² = 0.89, RMSE = 0.15 t ha⁻¹) and NSW (R² = 0.87, RMSE = 0.07 t ha⁻¹) performed well, but moderate performance was seen for SA (R² = 0.45, RMSE = 0.25 t ha⁻¹). Generally, high values were underpredicted and low values overpredicted. This study demonstrated the feasibility of applying RF modeling on satellite imagery and yielded ‘big data’ for regional as well as local-scale yield prediction.

Rice quality assessment is essential for meeting high-quality standards and consumer demands. However, challenges remain in developing cost-effective and rapid techniques to assess commercial rice grain quality traits. This paper presents the application of computer vision (CV) and machine learning (ML) to classify commercial rice samples based on dimensionless morphometric parameters and color parameters extracted using CV algorithms from digital images obtained from a smartphone camera. The artificial neural network (ANN) model was developed using nine morpho-colorimetric parameters to classify rice samples into 15 commercial rice types. Furthermore, the ANN models were deployed and evaluated on a different imaging system to simulate their practical applications under different conditions. Results showed that the best classification accuracy was obtained using the Bayesian Regularization (BR) algorithm of the ANN with ten hidden neurons at 91.6% (MSE = <0.01) and 88.5% (MSE = 0.01) for the training and testing stages, respectively, with an overall accuracy of 90.7% (Model 2). Deployment also showed high accuracy (93.9%) in the classification of the rice samples. The adoption by the industry of rapid, reliable, and accurate methods, such as those presented here, may allow the incorporation of different morpho-colorimetric traits in rice with consumer perception studies.

Grapevine smoke exposure and the subsequent development of smoke taint in wine has resulted in significant financial losses for grape growers and winemakers throughout the world. Smoke taint is characterized by objectional smoky aromas such as “ashy”, “burning rubber”, and “smoked meats”, resulting in wine that is unpalatable and hence unprofitable. Unfortunately, current climate change models predict a broadening of the window in which bushfires may occur and a rise in bushfire occurrences and severity in major wine growing regions such as Australia, Mediterranean Europe, North and South America, and South Africa. As such, grapevine smoke exposure and smoke taint in wine are increasing problems for growers and winemakers worldwide. Current recommendations for growers concerned that their grapevines have been exposed to smoke are to conduct pre-harvest mini-ferments for sensory assessment and send samples to a commercial laboratory to quantify levels of smoke-derived volatiles in the wine. Significant novel research is being conducted using spectroscopic techniques coupled with machine learning modeling to assess grapevine smoke contamination and taint in grapes and wine, offering growers and winemakers additional tools to monitor grapevine smoke exposure and taint rapidly and non-destructively in grapes and wine.

In this study, four textural classes of soil (Clay, Clay Loam, Loam, and Loamy Sand) were used to investigate the dielectric properties of soils, using a vector network analyser with an open-ended coaxial probe kit at room temperature (25 ± 2°C) in the 700–7000 MHz microwave frequency range. Four levels of soil moisture content (oven dry, 33% field capacity, 66% field capacity and 100% field capacity) were maintained to perform the experiment with three replication and three observation each. The results showed that, with increasing soil moisture, from oven dry conditions to 100% field capacity, both the real (Dielectric Constant) and imaginary (Loss Factor) components of the dielectric properties increased; however, the responses were not linear. The dielectric properties of oven dry soils were very low compared with the soils with higher moisture content. Therefore, soil moisture was the major contributor to the dielectric behaviour of soil. The dielectric properties of sandy soil were much lower than the other soils; however, the dielectric loss factor of the Dookie clay soil was higher compare with the other soils. Models were developed to explain the dielectric properties of soils as a function of frequency and moisture content. The goodness of fit (r²) for these models varies between 0.952 for the Dookie Sandy Soil to 0.997 for the Dookie Loam Soil, suggesting that these models were adequate to describe the dielectric properties of these soils over the range of frequencies and moisture contents assessed in this study. Another model was developed to estimate the expected penetration depth of electromagnetic waves in these soils, based on the model of the dielectric properties. It was clear that penetration decreases with both frequency and moisture content. Low frequencies penetrate further into the soils than higher frequencies. Similarly, dry soils allow further penetration than moist soils.

Wine aroma is an important quality trait in wine, influenced by its volatile compounds. Many factors can affect the composition and levels (concentration) of volatile aromatic compounds, including the water status of grapevines, canopy management, and the effects of climate change, such as increases in ambient temperature and drought. In this study, a low-cost and portable electronic nose (e-nose) was used to assess wines produced from grapevines exposed to different levels of smoke contamination. Readings from the e-nose were then used as inputs to develop two machine learning models based on artificial neural networks. Results showed that regression Model 1 displayed high accuracy in predicting the levels of volatile aromatic compounds in wine (R = 0.99). On the other hand, Model 2 also had high accuracy in predicting smoke aroma intensity from sensory evaluation (R = 0.97). Descriptive sensory analysis showed high levels of smoke taint aromas in the high-density smoke-exposed wine sample (HS), followed by the high-density smoke exposure with in-canopy misting treatment (HSM). Principal component analysis further showed that the HS treatment was associated with smoke aroma intensity, while results from the matrix showed significant negative correlations (p < 0.05) were observed between ammonia gas (sensor MQ137) and the volatile aromatic compounds octanoic acid, ethyl ester (r = −0.93), decanoic acid, ethyl ester (r = −0.94), and octanoic acid, 3-methylbutyl ester (r = −0.89). The two models developed in this study may offer winemakers a rapid, cost-effective, and non-destructive tool for assessing levels of volatile aromatic compounds and the aroma qualities of wine for decision making.