Early drought stress detection in cereals: simplex volume maximisation for hyperspectral image analysis

Abstract

Early water stress recognition is of great relevance in precision plant breeding and production. Hyperspectral imaging sensors can be a valuable tool for early stress detection with high spatio-temporal resolution. They gather large, high dimensional data cubes posing a significant challenge to data analysis. Classical supervised learning algorithms often fail in applied plant sciences due to their need of labelled datasets, which are difficult to obtain. Therefore, new approaches for unsupervised learning of relevant patterns are needed. We apply for the first time a recent matrix factorisation technique, simplex volume maximisation (SiVM), to hyperspectral data. It is an unsupervised classification approach, optimised for fast computation of massive datasets. It allows calculation of how similar each spectrum is to observed typical spectra. This provides the means to express how likely it is that one plant is suffering from stress. The method was tested for drought stress, applied to potted barley plants in a controlled rain-out shelter experiment and to agricultural corn plots subjected to a two factorial field setup altering water and nutrient availability. Both experiments were conducted on the canopy level. SiVM was significantly better than using a combination of established vegetation indices. In the corn plots, SiVM clearly separated the different treatments, even though the effects on leaf and canopy traits were subtle.

Proximal sensing and vegetation indices for site-specific evaluation on an irrigated crop tomato

“…It is known that the sensitivity of VIs to biophysical parameters may change substantially with vegetation density, resulting in function of the biophysical parameters [Ji and Peters, 2007], therefore monitor crop status during the crop cycle was important to detect either healthy plants or stressed plants. In this study WI do not have a significant correlation with plant yield probably because is too small the reflectance differences among minimum and maximum value in all sampling dates according to different studies such as Riggs and Running [1991] and Römer et al [2012] which confirmed the possibility to have values very close to each other. The water index divided by normalized difference vegetation index also showed no relationship with tomato yield, the R 2 was higher in WI/NDVI than WI because the NDVI is affected by structural and color changes (loss of pigments) in the overirrigated plants, and therefore it is indirectly related to the water concentration of living plants.…”

Discriminating Irrigated and Rainfed Maize with Diurnal Fluorescence and Canopy Temperature Airborne Maps

“…The heat dissipation can be detected using the Photochemical Reflectance Index (PRI), originally proposed [10] to track changes in the de-epoxidation state of the xanthophyll cycle pigments, related to heat dissipation. PRI has been used to track photosynthesis changes caused by water stress conditions with varying degrees of success [11][12][13][14]. In particular, PRI has been shown to be able to track the evolution of maize water stress from the early phase when only the plant physiology is affected [12] to the later ones when also plant structure changes [11], showing higher potentiality than the traditional greenness indices that hardly detect early stress conditions.…”

Applying hyperspectral imaging to explore natural plant diversity towards improving salt stress tolerance

“…Hyperspectral cameras compared to the NIR and VIS cameras are able to analyze big quantity of spectral bands between 350 and 2500 nm at nm-level resolution for each image pixel (Mahlein et al, 2013). Therefore, the hyperspectral imaging is an encouraging approach for the discovery of abiotic (Römer et al, 2012) and biotic (Mahlein et al, 2013) stresses.…”