This document discusses an automated system for plant disease diagnosis using image processing. It begins with listing the internal and external members on the research committee. It then presents an outline of the topics to be covered, including where imaging technology can be used, how the system works, the types of imaging technologies available, and the role of imaging in plant pathology. The document discusses several types of imaging technologies including digital imaging, chlorophyll fluorescence imaging, hyperspectral imaging, thermal imaging, and magnetic resonance imaging. It provides examples of each technology's application in plant disease detection and assessment. Overall limitations and future directions are also presented.
Science 7 - LAND and SEA BREEZE and its Characteristics
AUTOMATED SYSTEM FOR PLANT DISEASE DIAGNOSIS BY USING IMAGE PROCESSING
1.
2. AUTOMATED SYSTEM FOR PLANT DISEASE DIAGNOSIS BY
USING IMAGE PROCESSING
k. Vignesh,
Ph. D – I Year,
Dept.of plantpathology
Research supervisor :
Dr. l. vengadeshkumar ,
Assistant professor,
dept. of plant pathology.
INTERNAL mEMBER:
Dr. S. Sanjaygandhi,
Assistant professor,
dept. of plant pathology,
EXTERNAL MEMBER :
Dr.T.sabesan ,
Associate professor,
dept. of GENETICS AND PLANT BREEDING.
Chair person :
Dr. d. john christopher ,
professor & head,
dept. of plant pathology.
3. Where we can use…?
3
How it works…?
2
Automated Image Analysis…?
1
Types of imaging
technology…?
4
Role in plant pathology… ?
5
Conclusion
6
Outline
4. Managing new pests, pathogens
Increasing the efficiency of water
use
Reducing the Environmental
footprint of Agriculture
Growing food in a changing
climate
Managing the production of
bioenergy
Introduction
2018 PCAST Report
5. Diseases - cause major production and economic losses
Losses ranging between 20 and 40 %
Detection and Diagnosis
6. Powerful alternative to visual assessments
Automation provides - calibrated image analysis
Automation allows high throughput phenotyping
Automatedimage analysis
13. Automatic recognition of powdery
mildew lesions on lettuce leaf
Disease development
by Collectotrichum destructivum
on Tobacco leaf
(Zhaou et al. 2015 ; Casadesus 2017)
15. Imaging of wheat leaves infected with strains of
the fungal pathogen Zymoseptoria tritici was compared to traditional
visual assessment methods .
This fungus causes Septoria wheat blotch, which is characterized by
chlorosis, necrotic lesions, and fungal fruiting bodies called Pycnidia.
Typical visual disease assessments rely on estimates of percent of
leaf area covered by Pycnidia or lesions.
But here the automated image analysis of infected wheat leaves
allowed them to quantify Pycnidia size and density, along with other
traits.
(Stewart et al. 2016)
21. Specific wavelength information may be lost
Changes in uniformity of illumination
Difficult to achieve under constantly changing field
conditions.
Heterogeneous conditions
Limitations
22. CFI is a well-established method for
investigating effects of pathogens on
photosynthetic metabolism of host
plants at the whole - plant, detached
leaf and leaf disks levels.
Chlorophyll Fluorescence Imaging
(CFI)
23. Fv / Fm = Maximum quantum efficiency
Where,
Fv- Maximum variable fluorescence (Fm-Fo)
Fm- Maximum chlorophyll fluorescence
Fo- Minimum chlorophyll fluorescence
Chlorophyll fluorescence imaging of Arabidopsis thaliana inoculated with
Pseudomonas syringe
Calculation:
(Berger 2007)
24. Screening of Lettuce cultivars
against downy mildew pathogen
Wheat leaves inoculated
with leaf rust and
powdery mildew
(Bauriegel et al. 2012)
(Kuckenberg et al. 2009)
25. Presymptomatic detection of Powdery mildew on winter wheat
(Burling et al. 2012) and Citrus greening (Pereira et al. 2011)
A system using laser - induced
fluorescence at 473 nm of
wavelength
Red, Blue, Far red, Green
Laser InducedFluorescence Transient
(LIFT)
26. Objective
Quantifying the leaf area of Bean infected by
Xanthomonas fuscans sub.sp. fuscans - CFI analysis
(Rousseau et al. 2013)
Supportive article
27. Analysis of symptom development in Bean (Flavert)
caused by Xanthomonas fuscans sub.sp. fuscans
(Rousseau et al. 2013)
28. Evolution of the proportions of
necrotic, wilted and impacted
tissues on bean leaflets
Analysis of disease severity in
different cultivars
(Rousseau et al. 2013)
29. Measurement in the dark - adapted state is not possible or feasible to
perform from a practical point of view
Therefore, research has been directed at extracting fluorescence
parameters from sun-induced reflectance in the field, which would
have potential for plant disease assessment at the canopy or field
level
Limitations
30. Hyperspectral imaging is a relatively new technology that
involves the acquisition of electromagnetic spectra at every pixel
in an image, thus combining spatial and spectral information .
Hyper Spectral imaging
31. It measures reflected light from plants in hundreds of narrow
bands across the electromagnetic spectrum as a hypercube
HS sensors collect information as set of images which then
combines and forms into 3 dimensional data cube for processing
and analysis
Measurement
33. Small objects such as leaf lesions or seeds
- HS cameras can be mounted on a microscope
Monitoring plants in the laboratory
- HS camera mounted to the side or above a conveyor
belt or a translation stage
Field trials
- Vehicle-mounted HS cameras are used.
Eg: UAVs – Unnamed Aerial Vehicles
Based on need…HSI
( Uwe Knauer et al. 2017)
35. Damage on red leaf Lettuce caused by Pseudomonas sp.
Hyperspectral imaging of Lupine field (Deery et al. 2014)
36. Spectral image of barley
leaves infected with
different pathogens
Disease detection of fungal
pathogens by hyperspectral
imaging
37. Detection of powdery mildew disease progress in field
Detection of disease incidence in wheat
(Franke and Menz 2007)
(Lori 2015)
38. S. NO. CROP DISEASE REFERENCE
1. Wheat Fusarium head blight E. Bauriegel et al. 2011
2. Tomato EB and LB Chuanqi Xie et al. 2015
3. Grape Powdery mildew Uwe Knauer et al. 2017
4. Rape seed Stem rot Wenwen Kong et al. 2018
5. Barley Blast Rui-Qing Zhou et al. 2019
HSi based research
39. Fusarium - causes head blight disease can be detected by spectral
analysis (400 – 1000 nm) before harvest.
(Bauriegel et al. 2011)
Supportive article
40. Objectives:
1. To determine the wavelength ranges most appropriate for
discrimination of head blight
2. To find out the stage of grain development optimal for disease
detection
3. To elaborate a new straight forward classification method
applicable under semi-practical conditions.
Materialsand methods:
Growing wheat plants in controlled condition
Artificial inoculation of spores for 3 consecutive days after flowering
(Bauriegel et al. 2011)
41. Meanspectraof
Healthy - light grey line,
Diseased - black line
Tissues and Tissues covered with mycelia - dark grey line
Here, C area is chosen for classifying Head Blight Index
(Bauriegel et al. 2011)
44. It create statistical and computational challenges due
to multidimensionality of datasets.
Highly sensitive sensors, combined with large data
storage capacity and fast computers are needed, cost is
the limiting factor for working with hyper spectral
imaging for many laboratories.
Limitations
48. I. To evaluate the relationship between disease severity and
MTD of cucumber leaves infected by P. cubensis
II. To assess the impact of environmental conditions during
measurement of MTD in order to describe the potential of
MTD for the assessment and quantification of downy
mildew in the field.
objectives
(Oorke et al. 2006)
49. Infection with P. cubensis initially causes localized
decreases in surface temperature, which are thought to
be due to suppression of stomata closure early during
infection.
At later stages when the pathogen has caused areas of
necrosis, the temperature of the infected leaf tissue
increases to levels that are higher than that of
uninfected tissue.
(Oorke et al. 2006)
50. Cucumber leaf inoculated with
Pseudoperonospora cubensis
Seven days after inoculation,
the necrotic host tissue in the
centre of infection sites
showing a tissue temperature
0.6℃ higher than non-infected
leaf areas was surrounded by a
distinct small ring of cool
chlorotic tissue
(Oorke et al. 2006)
51. A high sensitivity of IRT to environmental factor
Limitations
52. 3D- image of growing plant in soil
Field of view, 76.8x25.6x25.6 mm; resolution, 200x200x400µm
Magnetic Resonance Imaging
(MRI)
MRI uses strong magnetic
fields and radio waves to
form three - dimensional
(3D) images of the objects.
MRI can be used to
evaluate internal tissue
structure and water
distribution in plants.
53. OBJECTIVE:
i. Investigation of changes of root geometry due to H. schachtii
presence
ii. Visualization of rotting symptoms caused by R. solani
iii. Detection of cysts and syncytia of H. schachtii on or in the
roots
iv. Examination of the inter-relationships between R. solani and H.
schachtii in a soil environment
Supportive article
55. NMR image of Sugarbeet
(Hillnhutter et al. 2011)
a) Healthy b) Heterodera inoculated
c) Rhizoctonia inoculated d) Both
56. X-ray imaging systems with high spatial resolution that can be
used for studying plant pathogen at subcellular level in 3D. Wave
length (10-100 pm)
X-ray has been tested for detection of fungal infection with
Aspergillus niger and Penicillium spp. in harvested wheat kernels.
X- ray computed tomography
65. Characterization
The impact of mixed infections on the optical properties of
plants has to be investigated
Interaction
The interaction of biotic and abiotic stress has to be explored
The connection to other knowledge based methods
FUTURE DIRECTIONS
66.
67. Vijai Singh (2020) A review of imaging techniques for plant disease
detection. Artificial Intelligence in Agriculture 229-242.
Bauriegel E, Brabandt H and Garber U (2014) Chlorophyll
fluorescence imaging to facilitate breeding of Bremia lettucae
resistant lettuce cultivars. Comput. Electron Agric 105:72-84.
Hilnhutter C, Sikora RA and Oerke VC (2012) Nuclear Magnetic
Resonance - A tool for imaging belowground damage caused
by Heterodera schachtii and Rhizoctonia solani in sugar beet. J
Exp Bot 63:319-327.
Mahlein AK (2016) Plant disease detection by image sensors-
parallels and specific demands for precision agriculture and
plant phenotyping. Plant Dis 100: 241-251
References