Near-infrared (NIR) and mid-IR spectroscopy applied to soil compositional analysis started to develop markedly in the 1990s, taking advantage of earlier advances in instrumentation and chemometrics for agricultural products. Today, NIR spectroscopy is envisioned as replacing laboratory analysis in certain applications (e.g., soil-carbon-credit assessment at the farm level). However, accuracy is still unsatisfactory compared with standard laboratory procedures, leading some authors to think that such a challenge will never be met. This article investigates the critical points to be aware of when accuracy of NIR-based measurements is assessed. First is the decomposition of the standard error of prediction into components of bias and variance, only the latter being reducible by averaging. This decomposition is not used routinely in the soil-science literature. Contrarily, a log-normal distribution of reference values is very often encountered with soil samples [e.g., elemental concentrations (e.g., carbon)] with numerous small or zero values. These very skewed distributions make us take precautions when using inverse regression methods (e.g., principal component regression or partial least squares), which force the predictions towards the centre of the calibration set, leading to negative effects on the standard error prediction-and therefore on prediction accuracy-especially when log-normal distributions are encountered. Such distributions, which are very common for soil components, also make the ratio of performance to deviation a useless, even hazardous, tool, leading to erroneous conclusions. We propose a new index based on the quartiles of the empirical distribution-ratio of performance to inter-quartile distance-to overcome this problem.

No abstract

AbbreviationsCP crude protein CV coeffi cient of variation DM dry matter EC European Commission FT Fourier transformation NIR near infrared N number PCR principal component regression PLS partial least squares R 2 coeffi cient of determination RP recoverable protein RPD ratio of the standard deviation of the calibration set to the SECV or SEP SD standard deviation SECV standard error of cross-validation SEL standard error of the laboratory reference method SEP standard error of prediction ST starchMost factories belonging to the Dutch company Avebe are working on products based on potato starch. More recently, there has been a strong development in the extraction and use of other potato constituents (i.e. protein, fi bres). At present, farmers are paid on the outcome of the "under water weight" measurements, a procedure prescribed by the EC to predict the starch content. However, since there is an outlet in the market, the company is also willing to pay out according to the protein content. The introduction of near infrared (NIR) spectroscopy as a fast quality control technology may be of great help in reaching this objective. NIR, placed either at-the-gate or at-line or in-line at plant level, will contribute to the setting up of a more objective payment method which will, RPD values (4.2 and 3.1, respectively). However, the equations obtained for CP and RP presented a low predictive ability (RPD ≈ 1.5). A discriminant analysis performed by using PLS2 regression correctly classifi ed 87.5% of mashed potato samples in groups of low (< 14 mg g -1 ) and high (≥ 14 mg g -1 ) protein content. A feasibility study with entire potatoes and a diode array spectrometer (Corona 45 VIS+NIR) was carried out and the preliminary results show great expectations concerning further implementation of NIR technology at the factory gate. However, further research and demonstration activities are needed before application will become possible. in turn, contribute to an improvement in the fi nancial situation of the farmers and also to better control of the complete manufacturing process. A total of 275 mashed potatoes were analysed by using a FT-NIR (ABB Bomem MB160D) spectrometer and calibration equations were developed to predict dry matter (DM), starch (ST), crude protein (CP) and recoverable protein (RP) content. The equations developed for the prediction of DM and ST presented an accuracy and precision acceptable for routine analysis according to their

No abstract

The control of refrigeration in the food chain is fundamental at all stages, with special emphasis on the retail stage. The implementation of information and communication technologies (IoT, open-source hardware and software, cloud computing, etc.) is representing a revolution in the operational paradigm of food control. This paper presents a low-cost IoT solution, based on free hardware and software, for monitoring the temperature in refrigerated retail cabinets. Specifically, the use of the ESP-8266-Wi-Fi microcontroller with DS18B20 temperature sensors is proposed. The ThingSpeak IoT platform is used to store and process data in the cloud. The solution presented is robust, affordable, and flexible, allowing to extend the scope of supervising other relevant parameters in the operating process (light control, energy efficiency, consumer presence, etc.).