Browsing by Author "Kuang, Boyan Y."
Now showing 1 - 10 of 10
Results Per Page
Sort Options
Item Open Access Assessment of soil organic carbon at local scale with spiked NIR calibrations: effects of selection and extra-weighting on the spiking subset(Wiley, 2014-01-17) Guerrero, C.; Stenberg, B.; Wetterlind, J.; Viscarra Rossel, R. A.; Maestre, F. T.; Mouazen, Abdul Mounem; Zornoza, R.; Ruiz-Sinoga, J. D.; Kuang, Boyan Y.Spiking is a useful approach to improve the accuracy of regional or national calibrations when they are used to predict at local scales. To do this, a small subset of local samples (spiking subset) is added to recalibrate the initial calibration. If the spiking subset is small in comparison with the size of the initial calibration set, then it could have little noticeable effect and a small improvement can be expected. For these reasons, we hypothesized that the accuracy of the spiked calibrations can be improved when the spiking subset is extra-weighted. We also hypothesized that the spiking subset selection and the initial calibration size could affect the accuracy of the recalibrated models. To test these hypotheses, we evaluated different strategies to select the best spiking subset, with and without extra-weighting, to spike three different-sized initial calibrations. These calibrations were used to predict the soil organic carbon (SOC) content in samples from four target sites. Our results confirmed that spiking improved the prediction accuracy of the initial calibrations, with any differences depending on the spiking subset used. The best results were obtained when the spiking subset contained local samples evenly distributed in the spectral space, regardless of the initial calibration's characteristics. The accuracy was improved significantly when the spiking subset was extra-weighted. For medium- and large-sized initial calibrations, the improvement from extra-weighting was larger than that caused by the increase in spiking subset size. Similar accuracies were obtained using small- and large-sized calibrations, suggesting that incipient spectral libraries could be useful if the spiking subset is properly selected and extra-weighted. When small-sized spiking subsets were used, the predictions were more accurate than those obtained with ‘geographically-local’ models. Overall, our results indicate that we can minimize the efforts needed to use near-infrared (NIR) spectroscopy effectively for SOC assessment at local scales.Item Open Access Effect of spiking strategy and ratio on calibration of on-line visible and near infrared soil sensor for measurement in European farms(Elsevier Science B.V., Amsterdam., 2013-04-02T00:00:00Z) Kuang, Boyan Y.; Mouazen, Abdul MounemA previously developed on-line visible and near infrared (vis-NIR) spectroscopy-based soil measurement system was implemented for the measurement of soil organic carbon (OC), total nitrogen (TN) and moisture content (MC) in three fields at three European farms. The on-line sensor platform was coupled with a mobile, fibre type, vis-NIR spectrophotometer (AgroSpec from tec5 Technology for Spectroscopy, Germany), with a measurement range of 305-2200 nm, to acquire soil spectra in diffuse reflectance mode. A general calibration set of 425 soil samples, spiked with different number of spectra from the three validation fields were used to establish calibration models for the studied soil properties using partial least squares (PLS) regression analysis. Different spiking strategies and spiking ratios were investigated and results revealed that the best prediction accuracy was obtained after 20% spiking ratio with samples whose spectra were measured in the laboratory. Evaluated by the values of residual prediction deviation (RPD), which is the ratio of standard deviation to root mean square error of prediction (RMSEP), the accuracy of the on-line measurement was classified as excellent for MC (RPD = 2.76-3.96), good to very good for OC (RPD = 1.88-2.38) and good to excellent for TN (RPD = 1.96- 2.52). Reducing the number of samples used for spiking resulted in deteriorating the prediction accuracy, although 1-2 samples per ha were found to provide good predictions. There was a distinguishable spatial similarity between the on-line and laboratory measured maps for all studied properties, although the full-data point maps provided more detailed information about the spatial variation. This confirms that the on-line vis-NIR soil sensor provides correct and detailed information about soil OC, TN and MC at high sampling resolutions.Item Open Access Influence of the number of samples on prediction error of visible and near infrared spectroscopy of selected soil properties at the farm scale.(Blackwell Publishing Ltd, 2013-01-23) Kuang, Boyan Y.; Mouazen, Abdul MounemAlthough visible and near infrared (vis-NIR) spectroscopy has proved to be a fast, inexpensive and relatively accurate tool to measure soil properties, considerable research is required to optimise the calibration procedure and establish robust calibration models. This paper reports on the influence of the number of samples used for the development of farm-scale calibration models for moisture content (MC), total nitrogen (TN) and organic carbon (OC) on the prediction error expressed as root mean square error of prediction (RMSEP). Fresh (wet) soil samples collected from four farms in Czech Republic, Germany, Denmark and the UK were scanned with a fibre type vis-NIR, AgroSpec spectrophotometer (tec5 Technology for Spectroscopy, Germany) with a spectral range of 305 - 2200 nm. Spectra were divided into calibration (two-third) and prediction (one-third) sets and the calibration spectra were subjected to a partial least squares regression (PLSR) with leave-one-out cross validation using Unscrambler 7.8 software (Camo Inc., Oslo, Norway). The RMSEP values of models with large sample number (46 - 84 samples from each farm) were compared with those of models developed using small sample number (25 samples selected from the large sample set of each farm) for the same variation range. Both large set and small set models were validated by the same prediction set for each property. Further PLSR analysis was carried out on samples from the German farm, with different sample number of the calibration set of 25, 50, 75 and 100 samples. Results showed that the large-size dataset models resulted in lower RMSEP values than the small-size dataset models for all the soil properties studied. The results also demonstrated that with the increase in sample number used in the calibration set, RMSEP decreased in almost linear fashion, although the largest decrease was between 25 and 50 samples. Therefore, it is recommended to chose the number of samples according to accuracy required, although 50 soil samples is considered appropriate in this study to establish calibration models of TN, OC and MC with smaller expected prediction errors as compared with smaller sample numbers.Item Open Access Methods and procedures for automatic collection and management of data acquired from on-the-go sensors with application to on-the-go soil sensors.(Elsevier Science B.V., Amsterdam., 2012-02-01T00:00:00Z) Peets, Sven; Mouazen, Abdul Mounem; Blackburn, Kim; Kuang, Boyan Y.; Wiebensohn, JensSensors for on-the-go collection of data on soil and crop have become essential for successful implementation of precision agriculture. This paper analyses the potentials and develops general procedures for onthe- go data acquisition of soil sensors. The methods and procedures used to manage data with respect to a farm management information system (FMIS) are described. The current data communication standard for tractors and machinery in agriculture is ISO 11783, which is rather well established and has gained market acceptance. However, there are a significant number of non-ISO 11783 compliant sensors in practice. Thus, two concepts are proposed. The first concept is on-the-go data collection based on ISO 11783, which mostly covers data on parameters related to tractor and machine performance, e.g. speed, draught, fuel consumption, etc. Process data from sensors with Control Area Network (CAN) interfaces is converted into ISO 11783 XML and then imported into relational database at FMIS using RelaXML tool. There is also the export function from database to task controller (TC) to provide task management, as described in ISO 11783:10. The second concept is on- the-go data collection with non-ISO 11783 sensors. This data is likely to be recorded in many formats, which require an import service. An import service is based on local or public sharing or semantic mapping outputting a common format for FMIS (e.g. AgroXML). Import is best performed as close to the generation of sensor data as possible to maximise the availability of metadata. A case study of sensor based variable rate fertilisation (VRF) has been undertaken focussing on German fertilisation rules.Item Open Access Non-biased prediction of soil organic carbon and total nitrogen with vis-NIR spectroscopy, as affected by soil moisture content and texture(Elsevier Science B.V., Amsterdam, 2013-03-01T00:00:00Z) Kuang, Boyan Y.; Mouazen, Abdul MounemThis study was undertaken to evaluate the effects of moisture content (MC) and texture on the prediction of soil organic carbon (OC) and total nitrogen (TN) with visible and near infrared (vis-NIR) spectroscopy under laboratory and on-line measurement conditions. An AgroSpec spectrophotometer was used to develop calibration models of OC and TN using laboratory scanned spectra of fresh and processed soil samples collected from five fields on Silsoe Farm, UK. A previously developed on-line vis-NIR sensor was used to scan these fields. Based on residual prediction deviation (RPD), which is the standard deviation of the prediction set (S.D.) divided by the root mean square error of prediction (RMSEP), the validation of partial least squares (PLS) models of OC and TN prediction using on-line spectra was evaluated as very good (RPD = 2.01-2.24) and good to excellent (RPD = 1.86-2.58), respectively. A better accuracy was obtained with fresh soil samples for OC (RPD = 2.11-2.34) and TN (RPD = 1.91-2.64), whereas the best accuracy for OC (RPD = 2.66-3.39) and TN (RPD = 2.85-3.45) was obtained for processed soil samples. Results also showed that MC is the main factor that decreases measurement accuracy of both on-line and fresh samples, whilst the accuracy was greatest for soils of high clay content. It is recommended that measurements of TN and OC under on-line and laboratory fresh soil conditions are made when soils are dry, particularly in fields with high clay content.Item Open Access On-line measurement of soil properties without direct spectral response in near infrared spectral range(Elsevier Science B.V., Amsterdam., 2013-08-01T00:00:00Z) Marin-González, O.; Kuang, Boyan Y.; Quraishi, M. Z.; Munoz-Garcia, M. A.; Mouazen, Abdul MounemSo far, the majority of reports on on-line measurement considered soil properties with direct spectral responses in near infrared spectroscopy (NIRS). This work reports on the results of on-line measurement of soil properties with indirect spectral responses, e.g. pH, cation exchange capacity (CEC), exchangeable calcium (Ca-ex) and exchangeable magnesium (Mg-ex) in one field in Bedfordshire in the UK. The on-line sensor consisted of a subsoiler coupled with an AgroSpec mobile, fibre type, visible and near infrared (vis-NIR) spectrophotometer (tec5 Technology for Spectroscopy, Germany), with a measurement range 305-2200 nm to acquire soil spectra in diffuse reflectance mode. General calibration models for the studied soil properties were developed with a partial least squares regression (PLSR) with one-leave-out cross validation, using spectra measured under non-mobile laboratory conditions of 160 soil samples collected from different fields in four farms in Europe, namely, Czech Republic, Denmark, Netherland and UK. A group of 25 samples independent from the calibration set was used as independent validation set. Higher accuracy was obtained for laboratory scanning as compared to on-line scanning of the 25 independent samples. The prediction accuracy for the laboratory and on-line measurements was classified as excellent/very good for pH (RPD = 2.69 and 2.14 and r(2) = 0.86 and 0.78, respectively), and moderately good for CEC (RPD = 1.77 and 1.61 and r(2) = 0.68 and 0.62, respectively) and Mg-ex (RPD = 1.72 and 1.49 and r(2) = 0.66 and 0.67, respectively). For Ca-ex, very good accuracy was calculated for laboratory method (RPD = 2.19 and r(2) = 0.86), as compared to the poor accuracy reported for the on-line method (RPD = 1.30 and r(2) = 0.61). The ability of collecting large number of data points per field area (about 12,800 point per 21 ha) and the simultaneous analysis of several soil properties without direct spectral response in the NIR range at relatively high operational speed and appreciable accuracy, encourage the recommendation of the on-line measurement system for site specific fertilisationItem Open Access On-line measurement of some selected soil properties for controlled input crop management systems(Cranfield University, 2012-02) Kuang, Boyan Y.; Mouazen, A. M.The evaluation of the soil spatial variability using a fast, robust and cheap tool is one of the key steps towards the implementation of Precision Agriculture (PA) successfully. Soil organic carbon (OC), soil total nitrogen (TN) and soil moisture content (MC) are needed to be monitored for both agriculture and environmental applications. The literature has proven that visible and near infrared (vis-NIR) spectroscopy to be a quick, cheap and robust tool to acquire information about key soil properties simultaneously with relatively high accuracy. The on-line vis-NIR measurement accuracy depends largely on the quality of calibration models. In order to establish robust calibration models for OC, TN and MC valid for few selected European farms, several factors affecting model accuracy have been studied. Nonlinear calibration techniques, e.g. artificial neural network (ANN) combined with partial least squares regression (PLSR) has provided better calibration accuracy than the linear PLSR or principal component regression analysis (PCR) alone. It was also found that effects of sample concentration statistics, including the range or standard derivation and the number of samples used for model calibration are substantial, which should be taking into account carefully. Soil MC, texture and their interaction effects are other principle factors affecting the in situ and on-line vis-NIR measurement accuracy. This study confirmed that MC is the main negative effect, whereas soil clay content plays a positive role. The general calibration models developed for soil OC, TN and MC for farms in European were validated using a previously developed vis-NIR on-line measurement system equipped with a wider vis-NIR spectrophotometer (305 – 2200 nm) than the previous version. The validation results showed this wider range on-line vis-NIR system can acquire larger than 1500 data point per ha with a very good measurement accuracy for TN and OC and excellent accuracy for MC. The validation also showed that spiking few target field samples into the general calibration models is an effective and efficient approach for upgrading the implementation of the on-line vis-NIR sensor for measurement in new fields in the selected European farms.Item Open Access Potential of on-line visible and near infrared spectroscopy for measurement of pH for deriving variable rate lime recommendations(MDPI AG, 2013-08-08T00:00:00Z) Tekin, Yücel; Kuang, Boyan Y.; Mouazen, Abdul MounemThis paper aims at exploring the potential of visible and near infrared (vis-NIR) spectroscopy for on-line measurement of soil pH, with the intention to produce variable rate lime recommendation maps. An on-line vis-NIR soil sensor set up to a frame was used in this study. Lime application maps, based on pH predicted by vis-NIR techniques, were compared with maps based on traditional lab-measured pH. The validation of the calibration model using off-line spectra provided excellent prediction accuracy of pH (R2 = 0.85, RMSEP = 0.18 and RPD = 2.52), as compared to very good accuracy obtained with the on-line measured spectra (R2 = 0.81, RMSEP = 0.20 and RPD = 2.14). On-line predicted pH of all points (e.g., 2,160) resulted in the largest overall field virtual lime requirement (1.404 t), as compared to those obtained with 16 validation points off-line prediction (0.28 t), on-line prediction (0.14 t) and laboratory reference measurement (0.48 t). The conclusion is that the vis-NIR spectroscopy can be successfully used for the prediction of soil pH and for deriving lime recommendations. The advantage of the on-line sensor over sampling with limited number of samples is that more detailed information about pH can be obtained, which is the reason for a higher but precise calculated lime recommendation rate.Item Open Access Prediction of soil TN and TC at a farm-scale using VIS-NIR spectroscopy(2011-12-31T00:00:00Z) Yang, Haiqing Q.; Kuang, Boyan Y.; Mouazen, Abdul MounemBuilding cost-effective models is of academic and practical value for fast measurement of soil properties, especially at a farm-scale. The aim of this study is to build quantitative models for soil total nitrogen (TN) and total carbon (TC) using visible and near infrared (VIS-NIR) spectroscopy. Dried samples (n=122) collected from an experimental farm, at Silsoe, Bedfordshire, United Kingdom, were scanned from 350 to 2500 nm at 1-nm intervals. Samples were divided into a calibration set (75%) and an independent validation set (25%). A partial least squares regression (PLSR) with leave-one-out cross validation was carried out based on different spectral ranges. Result shows that the best predictions (R2>0.90 and RPD>3.3) are achieved for TN using the VIS range (400- 700nm) and for TC using the VIS-NIR range (400-2500nm). It is concluded that VIS-NIR spectroscopy coupled with PLSR can be adopted for the prediction of soil TN and TC at a farm-scale.Item Open Access Size estimation of tomato fruits based on spectroscopic analysis(2011-12-31T00:00:00Z) Yang, Haiqing Q.; Kuang, Boyan Y.; Mouazen, Abdul MounemThis study used visible and near-infrared (VIS-NIR) spectroscopy for size estimation of tomato fruits of three cultivars. A mobile, fibre-type, VIS-NIR spectrophotometer (AgroSpec, Tec 5, Germany) with spectral range of 350-2200 nm, was used to measure reflectance spectra of on-vine tomatoes growing from July to September 2010. Spectra were divided into a calibration set (75%) and an independent validation set (25%). A partial least squares regression (PLSR) with leave-one-out cross validation was adopted to establish calibration models between fruit diameter and spectra. Furthermore, the latent variables (LVs) obtained from PLS regression was used as input to back-propagation artificial neural network (BPANN) analysis. Result shows that the prediction of PLSR model can produce good performance with coefficient of determination (R2) of 0.82, root-mean-square error of prediction (RMSEP) of 4.87 mm and residual prediction deviation (RPD) of 2.35. Compared to the PLSR model, the PLS-BPANN model provides considerably higher prediction performance with R2 of 0.88, RMSEP of 3.98 mm and RPD of 2.89. It is concluded that VIS-NIR spectroscopy coupled with PLS-BPANN can be adopted successfully for size estimation of tomato fruits.