From visual estimates to fully automated sensor-based measurements of plant disease severity: status and challenges for improving accuracy

Plant disease severity is commonly estimated visually without or with the aid of standard area diagram sets (SADs). It is generally believed that the use of SADs leads to less biased (more accurate) and thus more precise estimates, but the degree of improvement has not been characterized in a systematic manner. We built on a previous review and screened 153 SAD studies published from 1990 to 2021. A systematic review resulted in a selection of 72 studies that reported three linear regression statistics for individual raters, which are indicative of the two components of bias (intercept = constant bias; slope = systematic bias) and precision (Pearson's correlation coefficient, r), to perform a meta-analysis of these accuracy components. The meta-analytic model determined an overall gain of 0.07 (r increased from 0.88 to 0.95) in precision. Globally, there was a reduction of 2.65 units in the intercept, from 3.41 to 0.76, indicating a reduction in the constant bias. Slope was least affected and was reduced slightly from 1.09 to 0.966, indicating marginally less systematic bias when using SADs. A multiple correspondence analysis suggested an association of less accurate, unaided estimates with diseases that produce numerous lesions and for which maximum severities of 50% are rarely attained. On the other hand, more accurate estimates were observed with diseases that cause only a few lesions and those diseases where the lesions coalesce and occupy more than 50% of the specimen surface. This was most pronounced for specimen types other than leaves. By quantitatively exploring how characteristics of the pathosystem and how SADs affect precision and constant and systematic biases, we affirm the value of SADs for reducing bias and imprecision of visual assessments. We have also identified situations where SADs have greater or lesser effects as an assessment aid.

Download Full-text

Improving Accuracy of Tomato Plant Disease Diagnosis Based on Deep Learning With Explicit Control of Hidden Classes

Frontiers in Plant Science ◽

10.3389/fpls.2021.682230 ◽

2021 ◽

Vol 12 ◽

Author(s):

Alvaro Fuentes ◽

Sook Yoon ◽

Mun Haeng Lee ◽

Dong Sun Park

Keyword(s):

Deep Learning ◽

Tomato Plant ◽

Plant Disease ◽

High Efficiency ◽

Recognition Rate ◽

Disease Diagnosis ◽

Plant Diseases ◽

Improve Model ◽

Improving Accuracy ◽

And Control

Recognizing plant diseases is a major challenge in agriculture, and recent works based on deep learning have shown high efficiency in addressing problems directly related to this area. Nonetheless, weak performance has been observed when a model trained on a particular dataset is evaluated in new greenhouse environments. Therefore, in this work, we take a step towards these issues and present a strategy to improve model accuracy by applying techniques that can help refine the model’s generalization capability to deal with complex changes in new greenhouse environments. We propose a paradigm called “control to target classes.” The core of our approach is to train and validate a deep learning-based detector using target and control classes on images collected in various greenhouses. Then, we apply the generated features for testing the inference of the system on data from new greenhouse conditions where the goal is to detect target classes exclusively. Therefore, by having explicit control over inter- and intra-class variations, our model can distinguish data variations that make the system more robust when applied to new scenarios. Experiments demonstrate the effectiveness and efficiency of the proposed approach on our extended tomato plant diseases dataset with 14 classes, from which 5 are target classes and the rest are control classes. Our detector achieves a recognition rate of target classes of 93.37% mean average precision on the inference dataset. Finally, we believe that our study offers valuable guidelines for researchers working in plant disease recognition with complex input data.

Download Full-text

Systemic Acquired Resistance Delays Race Shifts to Major Resistance Genes in Bell Pepper

Phytopathology ◽

10.1094/phyto.2004.94.12.1376 ◽

2004 ◽

Vol 94 (12) ◽

pp. 1376-1382 ◽

Cited By ~ 8

Author(s):

A. M. Romero ◽

D. F. Ritchie

Keyword(s):

Disease Severity ◽

Plant Disease Resistance ◽

Plant Disease ◽

Systemic Acquired Resistance ◽

Field Experiments ◽

Specific Resistance ◽

Acquired Resistance ◽

R Genes ◽

Host Pathogen Interaction ◽

Major Resistance Genes

The lack of durability of host plant disease resistance is a major problem in disease control. Genotype-specific resistance that involves major resistance (R) genes is especially prone to failure. The compatible (i.e., disease) host-pathogen interaction with systemic acquired resistance (SAR) has been studied extensively, but the incompatible (i.e., resistant) interaction less so. Using the pepper-bacterial spot (causal agent, Xanthomonas axonopodis pv. vesicatoria) pathosystem, we examined the effect of SAR in reducing the occurrence of race-change mutants that defeat R genes in laboratory, greenhouse, and field experiments. Pepper plants carrying one or more R genes were sprayed with the plant defense activator acibenzolar-S-methyl (ASM) and challenged with incompatible strains of the pathogen. In the greenhouse, disease lesions first were observed 3 weeks after inoculation. ASM-treated plants carrying a major R gene had significantly fewer lesions caused by both the incompatible (i.e., hypersensitive) and compatible (i.e., disease) responses than occurred on nonsprayed plants. Bacteria isolated from the disease lesions were confirmed to be race-change mutants. In field experiments, there was a delay in the detection of race-change mutants and a reduction in disease severity. Decreased disease severity was associated with a reduction in the number of race-change mutants and the suppression of disease caused by the race-change mutants. This suggests a possible mechanism related to a decrease in the pathogen population size, which subsequently reduces the number of race-change mutants for the selection pressure of R genes. Thus, inducers of SAR are potentially useful for increasing the durability of genotype-specific resistance conferred by major R genes.

Download Full-text

A phytopathometry glossary for the twenty-first century: towards consistency and precision in intra- and inter-disciplinary dialogues

Tropical Plant Pathology ◽

10.1007/s40858-021-00454-0 ◽

2021 ◽

Author(s):

Clive H. Bock ◽

Sarah J. Pethybridge ◽

Jayme G. A. Barbedo ◽

Paul D. Esker ◽

Anne-Katrin Mahlein ◽

...

Keyword(s):

Plant Pathology ◽

Disease Severity ◽

Plant Disease ◽

Yield Loss ◽

First Century ◽

Control Methods ◽

Visual Estimation ◽

Disease Symptoms ◽

Disease Epidemiology ◽

Measurement Science

AbstractPhytopathometry can be defined as the branch of plant pathology (phytopathology) that is concerned with estimation or measurement of the amount of plant disease expressed by symptoms of disease or signs of a pathogen on a single or group of specimens. Phytopathometry is critical for many reasons, including analyzing yield loss due to disease, breeding for disease resistance, evaluating and comparing disease control methods, understanding coevolution, and studying disease epidemiology and pathogen ecology. Phytopathometry underpins all activities in plant pathology and extends into related disciplines, such as agronomy, horticulture, and plant breeding. Considering this central role, phytopathometry warrants status as a formally recognized branch of plant pathology. The glossary defines terms and concepts used in phytopathometry based on disease symptoms or visible pathogen structures and includes those terms commonly used in the visual estimation of disease severity and sensor-based methods of disease measurement. Relevant terms from the intersecting disciplines of measurement science, statistics, psychophysics, robotics, and artificial intelligence are also included. In particular, a new, broader definition is proposed for “disease severity,” and the terms “disease measurement” and “disease estimate” are specifically defined. It is hoped that the glossary contributes to a more unified cross-discipline approach to research in, and application of the tools available to phytopathometry.

Download Full-text

Resistance of Hosta Cultivars to Petiole Rot Caused by Sclerotium rolfsii var. delphinii

HortTechnology ◽

10.21273/horttech.13.2.0302 ◽

2003 ◽

Vol 13 (2) ◽

pp. 302-305 ◽

Cited By ~ 2

Author(s):

Brooke A. Edmunds ◽

Mark L. Gleason ◽

Stephen N. Wegulo

Keyword(s):

Disease Severity ◽

Plant Disease ◽

Leaf Shape ◽

Sclerotium Rolfsii ◽

Disease Development ◽

Wide Range ◽

Bare Root

Eighteen cultivars of hosta (Hosta spp.), selected to represent a wide range of size, leaf shape and color, and genetics, were evaluated for reaction to Sclerotium rolfsii var. delphinii in a greenhouse in Ames, Iowa in 2000 and 2001. Bare-root, single-eye plants were planted in 15.2-cm (6-inch) pots in a soil-containing (2000) and soilless (2001) mix and grown in a greenhouse for 3 months. Plants were then inoculated by placing a carrot disk infested with mycelium of S. rolfsii at the base of the plant. Disease severity was assessed weekly for 6 weeks as percent symptomatic petioles. Disease development varied significantly (P < 0.05) among cultivars. Overall, `Lemon Lime', `Munchkin', `Nakaiana', `Platinum Tiara', and `Tardiflora' had the most severe symptoms and `Halcyon' showed the least disease.

Download Full-text

The Relationship Between Plant Disease Severity and Yield

Annual Review of Phytopathology ◽

10.1146/annurev.py.33.090195.001003 ◽

1995 ◽

Vol 33 (1) ◽

pp. 119-144 ◽

Cited By ~ 84

Author(s):

R E Gaunt

Keyword(s):

Disease Severity ◽

Plant Disease ◽

The Relationship

Download Full-text

Standard Area Diagrams for Aiding Severity Estimation: Scientometrics, Pathosystems, and Methodological Trends in the Last 25 Years

Phytopathology ◽

10.1094/phyto-02-17-0069-fi ◽

2017 ◽

Vol 107 (10) ◽

pp. 1161-1174 ◽

Cited By ~ 24

Author(s):

Emerson M. Del Ponte ◽

Sarah J. Pethybridge ◽

Clive H. Bock ◽

Sami J. Michereff ◽

Franklin J. Machado ◽

...

Keyword(s):

Disease Severity ◽

Methodological Innovation ◽

Black And White ◽

Related Data ◽

Severity Estimation ◽

Essential Variable ◽

Color Combinations ◽

Improving Accuracy ◽

Prolific Authors ◽

True Color

Standard area diagrams (SAD) have long been used as a tool to aid the estimation of plant disease severity, an essential variable in phytopathometry. Formal validation of SAD was not considered prior to the early 1990s, when considerable effort began to be invested developing SAD and assessing their value for improving accuracy of estimates of disease severity in many pathosystems. Peer-reviewed literature post-1990 was identified, selected, and cataloged in bibliographic software for further scrutiny and extraction of scientometric, pathosystem-related, and methodological-related data. In total, 105 studies (127 SAD) were found and authored by 327 researchers from 10 countries, mainly from Brazil. The six most prolific authors published at least seven studies. The scientific impact of a SAD article, based on annual citations after publication year, was affected by disease significance, the journal’s impact factor, and methodological innovation. The reviewed SAD encompassed 48 crops and 103 unique diseases across a range of plant organs. Severity was quantified largely by image analysis software such as QUANT, APS-Assess, or a LI-COR leaf area meter. The most typical SAD comprised five to eight black-and-white drawings of leaf diagrams, with severity increasing nonlinearly. However, there was a trend toward using true-color photographs or stylized representations in a range of color combinations and more linear (equally spaced) increments of severity. A two-step SAD validation approach was used in 78 of 105 studies for which linear regression was the preferred method but a trend toward using Lin’s correlation concordance analysis and hypothesis tests to detect the effect of SAD on accuracy was apparent. Reliability measures, when obtained, mainly considered variation among rather than within raters. The implications of the findings and knowledge gaps are discussed. A list of best practices for designing and implementing SAD and a website called SADBank for hosting SAD research data are proposed.

Download Full-text