Browsing by Author "Whetton, Rebecca"
Now showing 1 - 4 of 4
Results Per Page
Sort Options
Item Open Access Machine learning based prediction of soil total nitrogen, organic carbon and moisture content by using VIS-NIR spectroscopy(Elsevier, 2016-05-24) Morellos, Antonios; Pantazi, Xanthoula-Eirini; Moshou, Dimitrios; Alexandridis, Thomas; Whetton, Rebecca; Tziotzios, Georgios; Wiebensohn, Jens; Bill, Ralf; Mouazen, Abdul MounemIt is widely known that the visible and near infrared (VIS-NIR) spectroscopy has the potential of estimating soil total nitrogen (TN), organic carbon (OC) and moisture content (MC) due to the direct spectral responses these properties have in the near infrared (NIR) region. However, improving the prediction accuracy requires advanced modelling techniques, particularly when measurement is planned for fresh (wet and un-processed) soil samples. The aim of this work is to compare the predictive performance of two linear multivariate and two machine learning methods for TN, OC and MC. The two multivariate methods investigated included principal component regression (PCR) and partial least squares regression (PLSR), whereas the machine learning methods included least squares support vector machines (LS-SVM), and Cubist. A mobile, fibre type, VIS-NIR spectrophotometer was utilised to collect soil spectra (305–2200 nm) in diffuse reflectance mode from 140 wet soil samples collected from one field in Germany. The results indicate that machine learning methods are capable of tackling non-linear problems in the dataset. LS-SVMs and the Cubist method out-performed the linear multivariate methods for the prediction of all three soil properties studied. LS-SVM provided the best prediction for MC (root mean square error of prediction (RMSEP) = 0.457% and residual prediction deviation (RPD) = 2.24) and OC (RMSEP = 0.062% and RPD = 2.20), whereas the Cubist method provided the best prediction for TN (RMSEP = 0.071 and RPD = 1.96).Item Open Access Nonlinear parametric modelling to study how soil properties affect crop yields and NDVI(Elsevier, 2017-05-03) Whetton, Rebecca; Zhao, Yifan; Shaddad, Sameh; Mouazen, Abdul MounemThis paper explores the use of a novel nonlinear parametric modelling technique based on a Volterra Non-linear Regressive with eXogenous inputs (VNRX) method to quantify the individual, interaction and overall contributions of six soil properties on crop yield and normalised difference vegetation index (NDVI). The proposed technique has been applied on high sampling resolution data of soil total nitrogen (TN) in %, total carbon (TC) in %, potassium (K) in cmol kg−1, pH, phosphorous (P) in mg kg−1 and moisture content (MC) in %, collected with an on-line visible and near infrared (VIS-NIR) spectroscopy sensor from a 18 ha field in Bedfordshire, UK over 2013 (wheat) and 2015 (spring barley) cropping seasons. The on-line soil data were first subjected to a raster analysis to produce a common 5 m by 5 m grid, before they were used as inputs into the VNRX model, whereas crop yield and NDVI represented system outputs. Results revealed that the largest contributions commonly observed for both yield and NDVI were from K, P and TC. The highest sum of the error reduction ratio (SERR) of 48.59% was calculated with the VNRX model for NDVI, which was in line with the highest correlation coefficient (r) of 0.71 found between measured and predicted NDVI. However, on-line measured soil properties led to larger contributions to early measured NDVI than to a late measurement in the growing season. The performance of the VNRX model was better for NDVI than for yield, which was attributed to the exclusion of the influence of crop diseases, appearing at late growing stages. It was recommended to adopt the VNRX method for quantifying the contribution of on-line collected soil properties to crop NDVI and yield. However, it is important for future work to include additional soil properties and to account for other factors affecting crop growth and yield, to improve the performance of the VNRX model.Item Open Access Optimising configuration of a hyperspectral imager for on-line field measurement of wheat canopy(Elsevier, 2016-12-30) Whetton, Rebecca; Waine, Toby W.; Mouazen, Abdul MounemThere is a lack of information on optimal measurement configuration of hyperspectral imagers for on-line measurement of a wheat canopy. This paper aims at identifying this configuration using a passive sensor (400–750 nm). The individual and interaction effects of camera height and angle, sensor integration time and light source distance and height on the spectra's signal-to-noise ratio (SNR) were evaluated under laboratory scanning conditions, from which an optimal configuration was defined and tested under on-line field measurement conditions. The influences of soil total nitrogen (TN) and moisture content (MC) measured with an on-line visible and near infrared (vis-NIR) spectroscopy sensor on SNR were also studied. Analysis of variance and principal component analysis (PCA) were applied to understand the effects of the laboratory considered factors and to identify the most influencing components on SNR. Results showed that integration time and camera height and angle are highly influential factors affecting SNR. Among integration times of 10, 20 and 50 ms, the highest SNR was obtained with 1.2 m, 1.2 m and 10° values of light height, light distance and camera angle, respectively. The optimum integration time for on-line field measurement was 50 ms, obtained at an optimal camera height of 0.3 m. On-line measured soil TN and MC were found to have significant effects on the SNR with Kappa values of 0.56 and 0.75, respectively. In conclusion, an optimal configuration for a tractor mounted hyperspectral imager was established for the best quality of on-line spectra collected for wheat canopy.Item Open Access Quantifying individual and collective influences of soil properties on crop yield(CSIRO Publishing, 2017-07-20) Whetton, Rebecca; Zhao, Yifan; Mouazen, Abdul MounemQuantification of the agronomic influences of soil properties, collected at high sampling resolution, on crop yield is essential for site specific soil management. The objective of this study was to implement a novel Volterra Non-linear Regressive with eXogenous inputs (VNRX) model accounting for the linear and non-linear variability (VNRX-LN) to quantify causal factors affecting wheat yield in a 22-ha field with a waterlogging problem in Bedfordshire, UK. The VNRX-LN model was applied using high-resolution data of eight key soil properties (total nitrogen (TN), organic carbon, pH, available phosphorous, magnesium (Mg), calcium, moisture content and cation exchange capacity (CEC)). The data were collected with an on-line (tractor mounted) visible and near infrared spectroscopy sensor and used as multiple-input to the VNRX-LN model, whereas crop yield represented the single-output in the system. Results showed that the largest contributors to wheat yield were CEC, Mg and TN, with error reduction ratio contribution values of 14.6%, 4.69% and 1% respectively. The overall contribution of the soil properties considered in this study equalled 23.21%. This was attributed to a large area of the studied field having been waterlogged, which masked the actual effect of soil properties on crop yield. It is recommended that VNRX-LN is validated on a larger number of fields, where other crop yield affecting parameters e.g., crop disease, pests, drainage, topography and microclimate conditions should be taken into account.