Low-cost hyper-spectral imaging system using a linear variable bandpass filter for agritech applications.

Hyperspectral imaging for agricultural applications provides a solution for non-destructive, large-area crop monitoring. However, current products are bulky and expensive due to complicated optics and electronics. A linear variable filter was developed for implementation into a prototype hyperspectral imaging camera that demonstrates good spectral performance between 450 and 900 nm. Equipped with a feature extraction and classification algorithm, the proposed system can be used to determine potato plant health with ∼88% accuracy. This algorithm was also capable of species identification and is demonstrated as being capable of differentiating between rocket, lettuce, and spinach. Results are promising for an entry-level, low-cost hyperspectral imaging solution for agriculture applications.

PHYLIS: A Low-Cost Portable Visible Range Spectrometer for Soil and Plants.

Monitoring soil and crop condition is vital for the sustainable management of agricultural systems. Often, land management decision-making requires rapid assessment of conditions, which is difficult if samples need to be taken and sent elsewhere for analysis. In recent years, advances in field-based spectroscopy have led to improvements in real-time monitoring; however, the cost of equipment and user training still makes it inaccessible for most land managers. At the James Hutton Institute, we have developed a low-cost visible wavelength hyperspectral device intended to provide rapid field-based assessment of soil and plant conditions. This device has been tested at the Institute's research farm at Balruddery, linking field observations with existing sample analysis and crop type information. We show that it is possible to rapidly and easily acquire spectral information that enables site characteristics to be estimated. Improvements to the sensor and its potential uses are discussed.

The potential for modelling peatland habitat condition in Scotland using long-term MODIS data.

Globally, peatlands provide an important sink of carbon in their near natural state but potentially act as a source of gaseous and dissolved carbon emission if not in good condition. There is a pressing need to remotely identify peatland sites requiring improvement and to monitor progress following restoration. A medium resolution model was developed based on a training dataset of peatland habitat condition and environmental covariates, such as morphological features, against information derived from the Moderate Resolution Imaging Spectroradiometer (MODIS), covering Scotland (UK). The initial, unrestricted, model provided the probability of a site being in favourable condition. Receiver operator characteristics (ROC) curves for restricted training data, limited to those located on a peat soil map, resulted in an accuracy of 0.915. The kappa statistic was 0.8151, suggesting good model fit. The derived map of predicted peatland condition at the suggested 0.56 threshold was corroborated by data from other sources, including known restoration sites, areas under known non-peatland land cover and previous vegetation survey data mapped onto inferred condition categories. The resulting locations of the areas of peatland modelled to be in favourable ecological condition were largely confined to the North and West of the country, which not only coincides with prior land use intensity but with published predictions of future retraction of the bioclimatic space for peatlands. The model is limited by a lack of spatially appropriate ground observations, and a lack of verification of peat depth at training site locations, hence future efforts to remotely assess peatland condition will require more appropriate ground-based monitoring. If appropriate ground-based observations could be collected, using remote sensing could be considered a cost-efficient means to provide data on changes in peatland habitat condition.

Exploring the Impact of Different Input Data Types on Soil Variable Estimation Using the ICRAF-ISRIC Global Soil Spectral Database.

Using the International Centre for Research in Agroforestry-International Soil Reference and Information Centre (ICRAF-ISRIC) global soil spectroscopy database, models were developed to estimate a number of soil variables using different input data types. These input types included: (1) site data only; (2) visible-near-infrared (Vis-NIR) diffuse reflectance spectroscopy only; (3) combined site and Vis-NIR data; (4) red-green-blue (RGB) color data only; and (5) combined site and RGB color data. The models produced variable estimation accuracy, with RGB only being generally worst and spectroscopy plus site being best. However, we showed that for certain variables, estimation accuracy levels achieved with the "site plus RGB input data" were sufficiently good to provide useful estimates (r2 > 0.7). These included major elements (Ca, Si, Al, Fe), organic carbon, and cation exchange capacity. Estimates for bulk density, contrast-to-noise (C/N), and P were moderately good, but K was not well estimated using this model type. For the "spectra plus site" model, many more variables were well estimated, including many that are important indicators for agricultural productivity and soil health. Sum of cation, electrical conductivity, Si, Ca, and Al oxides, and C/N ratio were estimated using this approach with r2 values > 0.9. This work provides a mechanism for identifying the cost-effectiveness of using different model input data, with associated costs, for estimating soil variables to required levels of accuracy.

A method for automatic segmentation and splitting of hyperspectral images of raspberry plants collected in field conditions.

Hyperspectral imaging is a technology that can be used to monitor plant responses to stress. Hyperspectral images have a full spectrum for each pixel in the image, 400-2500 nm in this case, giving detailed information about the spectral reflectance of the plant. Although this technology has been used in laboratory-based controlled lighting conditions for early detection of plant disease, the transfer of such technology to imaging plants in field conditions presents a number of challenges. These include problems caused by varying light levels and difficulties of separating the target plant from its background. Here we present an automated method that has been developed to segment raspberry plants from the background using a selected spectral ratio combined with edge detection. Graph theory was used to minimise a cost function to detect the continuous boundary between uninteresting plants and the area of interest. The method includes automatic detection of a known reflectance tile which was kept constantly within the field of view for all image scans. A method to split images containing rows of multiple raspberry plants into individual plants was also developed. Validation was carried out by comparison of plant height and density measurements with manually scored values. A reasonable correlation was found between these manual scores and measurements taken from the images (r2 = 0.75 for plant height). These preliminary steps are an essential requirement before detailed spectral analysis of the plants can be achieved.

Gypsophile chemistry unveiled: Fourier transform infrared (FTIR) spectroscopy provides new insight into plant adaptations to gypsum soils.

Gypsum soils are among the most restrictive and widespread substrates for plant life. Plants living on gypsum are classified as gypsophiles (exclusive to gypsum) and gypsovags (non-exclusive to gypsum). The former have been separated into wide and narrow gypsophiles, each with a putative different ecological strategy. Mechanisms displayed by gypsum plants to compete and survive on gypsum are still not fully understood. The aim of this study was to compare the main chemical groups in the leaves of plants with different specificity to gypsum soils and to explore the ability of Fourier transform infrared (FTIR) spectra analyzed with neural network (NN) modelling to discriminate groups of gypsum plants. Leaf samples of 14 species with different specificity to gypsum soils were analysed with FTIR spectroscopy coupled to neural network (NN) modelling. Spectral data were further related to the N, C, S, P, K, Na, Ca, Mg and ash concentrations of samples. The FTIR spectra of the three groups analyzed showed distinct features that enabled their discrimination through NN models. Wide gypsophiles stood out for the strong presence of inorganic compounds in their leaves, particularly gypsum and, in some species, also calcium oxalate crystals. The spectra of gypsovags had less inorganic chemical species, while those of narrow gypsum endemisms had low inorganics but shared with wide gypsophiles the presence of oxalate. Gypsum and calcium oxalate crystals seem to be widespread amongst gypsum specialist plants, possibly as a way to tolerate excess Ca and sulphate. However, other mechanisms such as the accumulation of sulphates in organic molecules are also compatible with plant specialization to gypsum. While gypsovags seem to be stress tolerant plants that tightly regulate the uptake of S and Ca, the ability of narrow gypsum endemisms to accumulate excess Ca as oxalate may indicate their incipient specialization to gypsum.

The physical environment and health-enhancing activity during the school commute: global positioning system, geographical information systems and accelerometry.

Active school travel is in decline. An understanding of the potential determinants of health-enhancing physical activity during the school commute may help to inform interventions aimed at reversing these trends. The purpose of this study was to identify the physical environmental factors associated with health-enhancing physical activity during the school commute. Data were collected in 2009 on 166 children commuting home from school in Scotland. Data on location and physical activity were measured using global positioning systems (GPS) and accelerometers, and mapped using geographical information systems (GIS). Multi-level logistic regression models accounting for repeated observations within participants were used to test for associations between each land-use category (road/track/path, other man-made, greenspace, other natural) and moderate-to-vigorous physical activity (MVPA). Thirty-nine children provided 2,782 matched data points. Over one third (37.1%) of children's school commute time was spent in MVPA. Children commuted approximately equal amounts of time via natural and man-made land-uses (50.2% and 49.8% respectively). Commuting via road/track/path was associated with increased likelihood of MVPA (Exp(B)=1.23, P <0.05), but this association was not seen for commuting via other manmade land-uses. No association was noted between greenspace use and MVPA, but travelling via other natural land-uses was associated with lower odds of MVPA (Exp(B)=0.32, P <0.05). Children spend equal amounts of time commuting to school via man-made and natural land-uses, yet man-made transportation route infrastructure appears to provide greater opportunities for achieving health-enhancing physical activity levels.