The reduction of chemical control of leek rust, a simulation study

Jong, Piet de; Daamen, R. A.; Rabbinge, R.

doi:10.1007/bf01874873

Cited by 9 publications

(5 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…If only a certain percentage of rust spots in a certain area are detected, it is enough indication to have to apply fungicides in that area of the field because it is likely the wider area is also infected. Moreover, the economic threshold for leek rust disease is zero, meaning fungicides are to be applied whether one or fifty rust pustules are present [12]. In the context of reducing fungicide usage, it is therefore more important that rust pustules detected are effectively rust, and not a false positive, than to detect every rust pustule present, with the risk of increasing the amount of false positives.…”

Section: Preprocessing and Model Selectionmentioning

confidence: 99%

“…Moreover, constant full-field application of a small range of fungicides can have detrimental effects due to the build-up of fungicide resistance in the pathogen population [10]. It has further been shown that rust disease inoculum mainly stems from existing leek infection sites, with only minor significance of alternate hosts and survival spores [12]. This means that despite regular fungicide treatments, there is still a self-propagating presence of rust disease in leek growing areas.…”

Section: Introductionmentioning

confidence: 99%

“…Even though precision agriculture has risen in popularity with the development of new sensing technologies, its core principles have been in practice for a much longer time. The first records of detection and variable rate fungicide treatment of leek rust disease can be traced back as far as 1995 [12]. Farmers at that time are reported to already have experimented with waiting to apply crop protection measures until first symptoms were visible and with spraying only the infected sites in the field.…”

Section: Introductionmentioning

confidence: 99%

“…These authors reported that the visible region is interesting regarding changes in pigmentation, with an increase in the orange and red spectral region (600-680), while the NIR-region (near-infrared) reflectance reduces due to cell structural damage. However, despite an extensive literature search, no published work was found regarding the detection of leek rust disease using hyperspectral sensing, even though leek rust disease has been described in the literature for the past 40 years [12,[19][20][21][22][23][24][25][26][27][28]. Instead, the rust disease control strategies cited here focus on searching for resistant cultivars and effective control strategies using fungicides.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Detection of Leek Rust Disease under Field Conditions Using Hyperspectral Proximal Sensing and Machine Learning

2021

View full text Add to dashboard Cite

Rust disease is an important problem for leek cultivation worldwide. It reduces market value and in extreme cases destroys the entire harvest. Farmers have to resort to periodical full-field fungicide applications to prevent the spread of disease, once every 1 to 5 weeks, depending on the cultivar and weather conditions. This implies an economic cost for the farmer and an environmental cost for society. Hyperspectral sensors have been extensively used to address this issue in research, but their application in the field has been limited to a relatively low number of crops, excluding leek, due to the high investment costs and complex data gathering and analysis associated with these sensors. To fill this gap, a methodology was developed for detecting leek rust disease using hyperspectral proximal sensing data combined with supervised machine learning. First, a hyperspectral library was constructed containing 43,416 spectra with a waveband range of 400–1000 nm, measured under field conditions. Then, an extensive evaluation of 11 common classifiers was performed using the scikit-learn machine learning library in Python, combined with a variety of wavelength selection techniques and preprocessing strategies. The best performing model was a (linear) logistic regression model that was able to correctly classify rust disease with an accuracy of 98.14 %, using reflectance values at 556 and 661 nm, combined with the value of the first derivative at 511 nm. This model was used to classify unlabelled hyperspectral images, confirming that the model was able to accurately classify leek rust disease symptoms. It can be concluded that the results in this work are an important step towards the mapping of leek rust disease, and that future research is needed to overcome certain challenges before variable rate fungicide applications can be adopted against leek rust disease.

show abstract

Section: Preprocessing and Model Selectionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Detection of Leek Rust Disease under Field Conditions Using Hyperspectral Proximal Sensing and Machine Learning

2021

View full text Add to dashboard Cite

show abstract

“…When this is not the case and a large epidemiological research investment is needed first, the increased efficiency of our approach is in the application of the first step, assessment of response of decision rules to weather. While our methodology was tailored to a weather-based warning system, similar approaches have been developed and applied for warning systems relying on field monitoring of diseases (de Jong, 1995;de Jong et al, 1995) or animal pests (van der Werfetal., 1996).…”

Section: Bowas Versionmentioning

confidence: 99%

Computer‐aided design of improved warning systems – a case study for Botrytis control in flower bulbs*

Kraker

Ende²,

Rossing

et al. 2000

EPPO Bulletin

View full text Add to dashboard Cite

Development and improvement of warning systems are often done empirically, relying on extensive field testing. As this approach is both costly and time-consuming, there is a need for a more rational and efficient alternative. In a study to explore the options for improvement of a Botrytis warning system in flower bulbs, we applied a computer-based approach to systems design. The approach consisted of the construction and evaluation of modified versions of the warning system using epidemiological knowledge, data sets of recorded and forecast weather and a simulation model of epidemic development and fungicide spray impact. Performance of modified versions was evaluated with regard to fungicide input, efficacy of disease control and sensitivity to the prediction error in weather forecasts. This approach can be more efficient than a purely empirical one, as it enables the designer to limit the number of alternative versions to be field-tested on the basis of explicit performance criteria. It also has the advantage that it provides insight into the potentials for improvement of the warning system. IntroductionDisease warning or forecasting systems are used to improve the efficiency and efficacy of disease control, by timing fungicide application according to identified need. Often, an empirical approach is taken to the development and improvement of warning systems. Development involves the definition of predictive equations and decision rules, and choosing values for parameters and decision thresholds. These may be based on epidemiological relationships established in controlled experiments or correlations derived from field observations, but they usually also include arbitrary elements, e.g. the choice of an action threshold. Subsequently, the warning system is evaluated in field experiments, commonly over several seasons, and several cycles of modification and field testing may follow before acceptable or improved performance is achieved. In consequence, this empirical approach requires a large research investment and is very time-consuming. Design of warning systems could become more efficient if alternative versions of the warning system could be screened with regard to their performance before field evaluation. Field testing could then be restricted to the most promising alternatives. In a study to explore the options for the improvement of a Botrytis warning system in flower bulbs, we applied such a screening technique in the form of computer-based evaluation. In this paper, we will describe the current weather-based Botrytis warning system and its *Paper presented at the EPPO Conference on Warning Services for Plant Protection, Piacenza (IT), 1999-05-12/14, background, explain the methodology we used, present first results and, finally, discuss various aspects of our approach.

show abstract

Selected Bibliography and Further Reading

Alford¹

2000

Pest and Disease Management Handbook

View full text Add to dashboard Cite

The reduction of chemical control of leek rust, a simulation study

Cited by 9 publications

References 12 publications

Detection of Leek Rust Disease under Field Conditions Using Hyperspectral Proximal Sensing and Machine Learning

Detection of Leek Rust Disease under Field Conditions Using Hyperspectral Proximal Sensing and Machine Learning

Computer‐aided design of improved warning systems – a case study for Botrytis control in flower bulbs*

Selected Bibliography and Further Reading

Contact Info

Product

Resources

About