Search

The composition of the soil is essential for higher yields in agriculture. The assessment of agrochemical soil indicators is carried out in laboratories using standardized chemical methods, which are time-consuming and relatively expensive. In the present study, an approach for predicting agrochemical indicators (nitrogen, phosphorus, potassium, humus and acidity) using spectral characteristics of soil samples was investigated. 110 soil samples from fields in the area of the city of Ruse, Bulgaria were analysed. The classification with the SIMKA method is carried out in the Pirouette software platform, and three procedures for pre-processing the data - first derivative, second derivative and multiplicative scatter correction (MSC) - have been compared. The results obtained for the standard error of validation show that for the nitrogen the minimum SEV is for MSC and 10 factors. For phosphorus - 2-nd derivative and 6 factors. For potassium - MSC and 9 factors. For humus - MSC and 10 factors. For acidity - 2-nd derivative and 10 factors. The future work will be focus on models for prediction of the indicators and assessment of the accuracy. [ABSTRACT FROM AUTHOR]

Hyper-spectral imaging provides an opportunity for high-precision and detailed spectral measurements of the object of study with a wide range of applicability - from laboratory imaging of samples through field to remote satellite measurements. The paper presents a study in which the average spectral characteristics for each sample were obtained from the hyperspectral images of soil samples from three different pre-treatment fields. Spectral data were processed in the Pirouette software environment and principal component analysis was applied. Using the first three principal components, the soil fields are clearly separable. Interclass distances were also calculated. From the resulting discriminating power, wavelengths can be derived that are informative about the content of the soil fields. The sections that were determined from the modelling power from 1200 to 1250 nm; 1300 to 1350 and 1400 to 1450 nm, the wavelengths can be defined that would be used as inputs to the classifiers. The future work is developing of models for predicting micro- and macro-elements in soil using hyperspectral images. [ABSTRACT FROM AUTHOR]

In this report, alternative tools for determining agrochemical parameters of soil - content of organic matter (humus), pH, phosphorus, potassium and nitrogen, obtained from colour characteristics of digital images from optical devices by statistical methods are proposed. Regression models were developed using two methods: single factor linear regression analysis LR and multifactor linear regression analysis MLR Stepwise. The models were compared and evaluated using accuracy assessment criteria - coefficient of determination (R2), root mean square error (RMSE) and residual prediction deviation (RPD). With the resulting models from data from the camera device, pH can be determined with an accuracy of 99% and phosphorus P with an accuracy of 90%. With the camera of a mobile phone, the content of organic matter can be determined with the accuracy of 96%. The proposed tools could be used in modern smart agriculture as a proposed method for express monitoring of soil agrochemical indicators, as well as implemented in programming mobile, webbased applications for determining the content of organic matter (humus), phosphorus, pH and potassium. Future work involves using other linkage modelling methods to increase the accuracy rate for nitrogen determination. [ABSTRACT FROM AUTHOR]

Educators in higher education must reinvent their learning environments so that digitalization augments and enhances instruction and learning. The widespread use of digital devices and applications makes this possible. We argue that educators must therefore develop their digital skills and adopt more innovative instructional strategies. Using available technological instruments for pedagogical purposes, instructional technology permits the development of engaging, effective learning experiences. However, the implementation of instructional technology in practice is difficult. It is complex, multifaceted and fraught with challenges, and consequently, educators struggle to realize the potential benefits. This exploratory study aims to uncover key challenges to implementing instructional technology faced by educators in order to better understand their needs and requirements. To do this, we conducted a comprehensive qualitative study involving 41 higher education educators via focus groups in Ireland, Turkey, Bulgaria and Portugal. Our analysis uncovers many challenges and we curate and discuss a prioritized list in this paper. The findings present pertinent challenges for educators and institutions across four countries and suggest how to overcome them. This study also offers insight into future research and development needed to expand instructional technologies to accommodate and address respondents' concerns. [ABSTRACT FROM AUTHOR]