https://lib.dr.iastate.edu/cgi/viewcontent.cgi?article=1240&context=me_pubs

Machine Learning for High-Throughput Stress Phenotyping in Plants

Color science is a multidisciplinary field with broad applications in industries such as digital imaging, coatings and textiles, food, lighting, archiving, art, and fashion. Accurate definition and measurement of color appearance is a challenging task that directly affects color reproduction in such applications. Color Appearance Models addresses those challenges and offers insight into the preferred solutions. Extensive research on the human visual system (HVS) and color vision has been performed in the last century, and this book contains a good overview of the most important and relevant literature regarding color appearance models.

https://www.spiedigitallibrary.org/journals/Journal-of-Electronic-Imaging/volume-23/issue-2/029901/Color-Appearance-Models/10.1117/1.JEI.23.2.029901.pdf

Color Appearance Models

The success of the supervised classification of remotely sensed images acquired over large geographical areas or at short time intervals strongly depends on the representativity of the samples used to train the classification algorithm and to define the model. When training samples are collected from an image or a spatial region that is different from the one used for mapping, spectral shifts between the two distributions are likely to make the model fail. Such shifts are generally due to differences in acquisition and atmospheric conditions or to changes in the nature of the object observed. To design classification methods that are robust to data set shifts, recent remote sensing literature has considered solutions based on domain adaptation (DA) approaches. Inspired by machine-learning literature, several DA methods have been proposed to solve specific problems in remote sensing data classification. This article provides a critical review of the recent advances in DA approaches for remote sensing and presents an overview of DA methods divided into four categories: 1) invariant feature selection, 2) representation matching, 3) adaptation of classifiers, and 4) selective sampling. We provide an overview of recent methodologies, examples of applications of the considered techniques to real remote sensing images characterized by very high spatial and spectral resolution as well as possible guidelines for the selection of the method to use in real application scenarios.

Domain Adaptation for the Classification of Remote Sensing Data: An Overview of Recent Advances

This review explores how imaging techniques are being developed with a focus on deployment for crop monitoring methods. Imaging applications are discussed in relation to both field and glasshouse-based plants, and techniques are sectioned into ‘healthy and diseased plant classification’ with an emphasis on classification accuracy, early detection of stress, and disease severity. A central focus of the review is the use of hyperspectral imaging and how this is being utilised to find additional information about plant health, and the ability to predict onset of disease. A summary of techniques used to detect biotic and abiotic stress in plants is presented, including the level of accuracy associated with each method.

Hyperspectral image analysis techniques for the detection and classification of the early onset of plant disease and stress.

Plant phenotyping refers to a quantitative description of the plant’s anatomical, ontogenetical, physiological and biochemical properties. Today, rapid developments are taking place in the field of non-destructive, image-analysis -based phenotyping that allow for a characterization of plant traits in high-throughput. During the last decade, ‘the field of image-based phenotyping has broadened its focus from the initial characterization of single-plant traits in controlled conditions towards ‘real-life’ applications of robust field techniques in plant plots and canopies. An important component of successful phenotyping approaches is the holistic characterization of plant performance that can be achieved with several methodologies, ranging from multispectral image analyses via thermographical analyses to growth measurements, also taking root phenotypes into account.

/pdf/plant-phenotyping-from-bean-weighing-to-image-analysis-3tmd3wun7o.pdf

Plant phenotyping: from bean weighing to image analysis

Early stress detection in crop plants is highly relevant, but hard to achieve. We hypothesize that close range hyperspectral imaging is able to uncover stress related processes non-destructively in the early stages which are invisible to the human eye. We propose an approach which combines unsupervised and supervised methods in order to identify several stages of progressive stress development from series of hyperspectral images. Stress of an entire plant is detected by stress response levels at pixel scale. The focus is on drought stress in barley ( Hordeum vulgare ). Unsupervised learning is used to separate hyperspectral signatures into clusters related to different stages of stress response and progressive senescence. Whereas all such signatures may be found in both, well watered and drought stressed plants, their respective distributions differ. Ordinal classification with Support Vector Machines (SVM) is used to quantify and visualize the distribution of progressive stages of senescence and to separate well watered from drought stressed plants. For each senescence stage a distinctive set of most relevant Vegetation Indices (VIs) is identified. The method has been applied on two experiments involving potted barley plants under well watered and drought stress conditions in a greenhouse. Drought stress is detected up to ten days earlier than using NDVI. Furthermore, it is shown that some VIs have overall relevance, while others are specific to particular senescence stages. The transferability of the method to the field is illustrated by an experiment on maize ( Zea mays ).

Detection of early plant stress responses in hyperspectral images

Hyperspectral images can be used to uncover physiological processes in plants if interpreted properly. Machine Learning methods such as Support Vector Machines (SVM) and Random Forests have been applied to estimate development of biomass and detect and predict plant diseases and drought stress. One basic requirement of machine learning implies, that training and testing is done in the same domain and the same distribution. Different genotypes, environmental conditions, illumination and sensors violate this requirement in most practical circumstances. Here, we present an approach, which enables the detection of physiological processes by transferring the prior knowledge within an existing model into a related target domain, where no label information is available. We propose a two-step transformation of the target features, which enables a direct application of an existing model. The transformation is evaluated by an objective function including additional prior knowledge about classification and physiological processes in plants. We have applied the approach to three sets of hyperspectral images, which were acquired with different plant species in different environments observed with different sensors. It is shown, that a classification model, derived on one of the sets, delivers satisfying classification results on the transformed features of the other data sets. Furthermore, in all cases early non-invasive detection of drought stress was possible.

Unsupervised domain adaptation for early detection of drought stress in hyperspectral images

Maps which are created on demand by combining geospatial data from different Web Map Services integrate conflicting portrayals and do not satisfy the requirements of effective cartographic communication. A significant improvement is achieved by selecting clearly distinguishable colours which are determined by solving an optimisation problem. Cartographic guidelines and user characteristics (e.g. colour vision impairment) can be incorporated into the optimisation model as constraints which guide the selection of colours.

Identification of Optimal Colours for Maps from the Web

. Detection of crop stress from hyperspectral images is of high importance for breeding and precision crop protection. However, the continuous monitoring of stress in phenotyping facilities by hyperspectral imagers produces huge amounts of uninterpreted data. In order to derive a stress description from the images, interpreting algorithms with high prediction performance are required. Based on a static model, the local stress state of each pixel has to be predicted. Due to the low computational complexity, linear models are preferable. In this paper, we focus on drought-induced stress which is represented by discrete stages of ordinal order. We present and compare five methods which are able to derive stress levels from hyperspectral images: One-vs.-one Support Vector Machine (SVM), one-vs.-all SVM, Support Vector Regression (SVR), Support Vector Ordinal Regression (SVORIM) and Linear Ordinal SVM classification. The methods are applied on two data sets - a real world set of drought stress in single barley plants and a simulated data set. It is shown, that Linear Ordinal SVM is a powerful tool for applications which require high prediction performance under limited resources. It is significantly more efficient than the one-vs.-one SVM and even more efficient than the less accurate one-vs.-all SVM. Compared to the very compact SVORIM model, it represents the senescence process much more accurate.

/pdf/ordinal-classification-for-efficient-plant-stress-prediction-3tykmvr3g1.pdf

Jörg Steinrücken

Papers

Detection of early plant stress responses in hyperspectral images

Unsupervised domain adaptation for early detection of drought stress in hyperspectral images

Identification of Optimal Colours for Maps from the Web

Ordinal classification for efficient plant stress prediction in hyperspectral data