Proceedings International Conference of Agricultural Engineering, Zurich, 06-10.07.2014 – www.eurageng.eu 1/9 Ref: C0608 On-line visible and near infrared (vis-NIR) measurement of key soil fertility parameters in vegetable crop fields Virginia Jimenez-Donaire, Boyan Kuang, Toby Waine and Abdul Mouazen, Cranfield Soil and AgriFood Institute CSAFI, Cranfield University, Cranfield, MK42 0AL United Kingdom Abstract Fertiliser applications in vegetable crops are one of the main input costs in vegetable produc- tion. Thematic soil maps have been widely used for decades to characterise soil nutrients and, therefore, apply variable rate fertilisers. However, traditional variable rate methods are time-consuming, costly and not accurate. Thus, they fail in providing a true estimate of the nutrients soil needs. To obtain better crop response to inputs, a rapid, non-destructive, timely and cost-effective soil analysis are needed to enable site-specific fertiliser applications. Prox- imal soil sensing with visible and near infrared (vis-NIR) spectroscopy is a promising tool to assist in variable rate applications This paper aims to develop reliable calibration models for a previously developed on-line visible (vis) and near infrared (NIR) spectroscopy sensor, for the prediction of soil properties in vegetable crop fields for a better N fertiliser management. We hypothesized that for a better prediction performance, it is not always true to assume that field scale calibration models will lead to the best prediction accuracy and that heterogene- ous data set may improve prediction accuracy. To test this hypothesis, experiments were established in crops of cauliflower (Brassica oleracea) and savoy cabbage (Brassica oleracea var. sabauda L.) over the 2013 season in two fields (namely F1 and F2), in UK. A mobile, fibre-type, vis–NIR spectrophotometer (AgroSpec, Tec5 Technology for Spectrosco- py, Germany) with a measurement range of 305-2200 nm was used to measure soil spectra in diffuse reflectance mode, measuring up to ~1500 points per ha. Four different calibration sets were tested to establish the most accurate calibration model for moisture content (MC), soil organic carbon (OC), pH and total nitrogen (TN), using partial least squares (PLS) re- gression analysis along with full cross validation. The scenarios were selected according to different spectral library size and geographical scale: Scenario 1 (SC1 (local); n=136), Sce- nario 2 (SC2 (regional); n=286), Scenario 3 (SC3 (national); n=472), Scenario 4 (SC4 (conti- nental); n=556). The best results in cross-validation were obtained for MC with SC2 (R 2 = 0.89; RPD = 3.09), followed by SC4 (R 2 = 0.88; RPD = 2.91); and SC1 and SC4 worked very well for MC on-line prediction (R 2 > 0.90 and RPD > 2.5) in both fields. SC3 and SC4 both provided the best per- formance for OC and TN in cross-validation, whereas no clear trend was observed for on-line prediction. Poor model performance was obtained for pH in on-line predictions (R 2 < 0.15 and RPD < 0.9 in F1; and R 2 < 0.30 and RPD < 0.9 in F2). Although the calibration models using the on-line vis-NIR sensor provided good and detailed information of the soil nutrients analysed, future research will be needed to estimate these properties more accurately, with the aim to develop reliable vis-NIR calibration models for the on-line measurement in vege- table crop fields. Keywords: Vis-NIR spectroscopy, horticulture, on-line measurement, nitrogen fertiliz- er, variable-rate