3. Supervisory Committee:
Prof. Dr. Heba Mohamed Abdel-Nabi Prof. of Plant pathology – Faculty of Agriculture
Prof. Dr. Tarek Youssef Bayoumi (Late) Prof. of Breeding Crops – Faculty of Agriculture
Prof. Dr. Mohamed Osman Arnos Prof. of RS & GIS Applications - Faculty of Science
4. Presented from : Dr. Ahmed Ameen Abdullah
• PhD (2022) in Plant Pathology
• Master's degree in Plant Pathology (2015).
• Bachelor of Agricultural Sciences and Plant Protection (2001)
5. Introduction
Sugar beet is one of the most important sugar crops in the world. It covers
approximately 35% of the global needs of sugar.
In Egypt, sugar beet is the second crop after sugar cane for sugar production
(Memon et al., 2004; Eweis et al., 2006). The total production of sugar beet in Egypt
from 2010 to 2019 was about 7.84 to 10.53 million tons (Wu et al., 2013; FAOSTAT,
2020).
in Egypt sugar beet is sowing from September to November and harvest from March
to April as a result, sugar beet crop attack by several pathogenic represent in
Cercospora leaf spot disease, rust disease and Powdery mildew disease.
In Egypt Cercospora leaf spot disease (CLS) is considered the most dangerous of
them which due to Crop losses of sugar beet can reach 40% and result in complete
yield loss in the absence of fungicide treatments
6. Aimofthestudy
01 Assess and determine the ability of satellite to detect
and assessment diseased severity (DS) of CLS
disease
02 Making spectral signature for CLS disease
03
04
Assess Spectral signature, Band Ratio, Vegetation indices,
Pathogenic zonation and Change detection to detect and
assessment DS of CLS
Attempted to make a mapping for CLS.
7. • Research plan
Survey to determine diseases in
study area
Isolation and identification of the
pathogen
Relationship between DS and VI
Determine change detection of the
disease on study area
Determine the best technique to
detect infected and healthy areas.
Relationship between disease
severity of CLS disease and
spectral reflectance
8. Flowchart of
work
AI- Applied
Deep learning
Validation
Filed assessment
under control
Hyperspectral
Measurement
Geospatial
Technology
Remote
sensing data
GIS
Filed Assessment
Survey
Assessment of
CLS
GPS DATA
Laboratory
Isolation and
identification
Pathogenicity test
Efficacy of plant
extract
Thermal imaging
measurement
10. Survey
SBR
Assessment two variable :
1. Disease severity (%)
2. Disease incidence (%)
Sample :-
1. Sample collected every 10 days from area 15m*15m to calculated
DS and DI
2.GPS data:-
Collected 220 points for each field from the same place as the sampling
using GPS device(Garmin)
KSF1
FAN2
FAN1
KSF2
11. RESULT-S
0%
20.05
%
20.05%
0%
40.30
%
The highest percentage of DS
recorded in KSF1 with rate from
01.15 to 49.05%
49.05%
In KSF2 The percentage DS
recorded with rate from 0.35 to
20.05%
00.00
Location DI (%) DS (%)
FAN1
FAN2
SBR
KSF2
KSF1
0%
0% 0%
0% 0%
40.30
%
49.05
%
71.90
%
20.05%
In FAN1,FAN2 and SBR didn’t
appearance
14. With (Concentration 10%,20%,30%,35% and 40%)
A
B
C
.
Leaves of Caper Extract
Bunge flowers Extract
Chicory flowers Extract
Efficacy of plants extracts on the linear
growth of Cercospora beticola
A
A
B C
A
15. Using thermal image to measure the interaction between plant extracts and fungal growth:-
.
16. Geospatial technology(Remote sensing data)
This part depend on the link between
the data recorded from the field
assessment and recorded data
satellite by using GPS device data
Data was acquired during growing
season of sugar beet data from 05th
November 2017 to April 2017 with
free cloud days
17. Landsat 8 OLI
Sentinel 2A MSI
Satellite data
Data acquisition data obtain from to types of satellites LANDSAT 8
and SENTINEL 2A da
1. Landsat 8 OLI 2. Sentinel-2A MSI
Repeat Coverage: 16 days Repeat Coverage: 10 days
18. Steps this part was followed:-
• Classification of land cover
• Distinguishing between healthy and infected vegetation and mapping the spread of the
disease within the study area Using many new techniques represented in this flowchart :
Techniques
Color composite
Spectral
signature Band ratio Vegetation indices
Pathogenic
Zonation
19. Color composite:
This technique
was mainly used to distinguish between the land cover
classes using Optimum Index Factor(OIF) statistic to
select the optimum combination (R:G:B) which has the
highest ranking(Standard-Deviations)with and highest
information for discriminating between land cover in the
study area
20.
21. Spectral signature
This part is based on the spectral reflection properties of the land cover for Discriminating them as shown in the
following :
Soil
Water
Healthy Vegetation
Visible spectrum Near-infrared Shortwave-infrared
Red-edge Visible spectrum Near-infrared Shortwave-infrared
Soil
Vegetation
Water
22. Extract spectral signature for infecting sugar beet canopy
It is noted in Increasing in DS% lead to increase in spectral reflectance in the Visible light and SWIR
spectrum
While in case of the NIR and Red-edge increase in DS% lead to decrease spectral reflectance in this
wavelength
23. Spatial distribution of CLS diseased with development of disease severity using bands of red-egde
H: represent healthy canopy
INF: represent infected area
24. Correlation between spectral reflectance and disease severity of CLS disease :
Correlation between DS and spectral reflectance in NIR and Red-edge is negative
with value 0.893 and respectively
While correlation between DS and spectral reflectance in visible bands
(blue ,green and red ) and SWIR in positive with value 0.883,0.942
,0.946 ,0,903 and 0.937 respectively
25. Band ratio
This technique using Landsat 8 and Sentinel 2A bands to find the spectral ratios who is the most unique
spectral characteristics data to discriminate between healthy and infected area
The flow technique of (Vincent 1997) with mine modification to determine and assess the degree of band
ratio which ability to discriminate between healthy and infected area. The modification assumes a scale
ranging 0 to 3 instead of 0 to 9
26. Result
Landsat 8 and Sentinel 2A band ration the best band of ratios can determine all level of
disease severity (low , moderate and high)
27. Vegetation indices
This technique using VIs was computed using bands of Sentinel 2A and Landsat 8.
The 25 indices were selected for this study which includes many categories as were
showed as following:-
1.The broadband greenness
2.The Narrowband greenness
3.Water content
4.Soil indices
5.Dry or senescent indices
6.Leaf pigment indices
7.Light use efficiency
28. Determine VI that discriminate
between healthy and infected area
using Landsat 8 and Sentinel 2A
bands.
The best VI
29. Correlation between VI and Disease severity of CLS
The best final vegetation indices can detection CLS DS all level of
31. Change Detection Analysis
The aim of this technique to determine the ability and sensitive of satellite data to
assessment development of disease (DS) on field of sugar beet
Three vegetation indices Brightness index (BI) ,Chlorophyll- Red edge and Chl-Reg index Stress
Moisture Index (MSI) which have high correlation coefficient . three level of disease severity of
CLS (low, moderate, and high) were using in this technique
Can we use data satellite to measure disease severity in field ?
32. Chl-Reg index:-
Healthy area 6.8 14.20 45.05
Convert
Decreasing Increasing = 9.2%
Convert Convert
Change during development of CLS disease in field Healthy area = 89 feddan
Increasing= 19.30%
BI index:-
Healthy area 6.8 14.20 45.05
Convert
Decreasing
Increasing = 17.88%
Convert Convert
Increasing= 20.18% Increasing= 27.77%
Healthy area 6.8 14.2 45.05
Convert
Decreasing
Increasing = 18.88%
Convert Convert
Increasing= 22.77% Increasing= 28.88%
MSI index:-
Note from the result the more sensitive to change in severity is MSI index
DS DS DS
33. Pathogenic-Zonation
Technique was used for extracting of alteration zones associated with the status of sugar beet
vegetation (healthy, infected with diseases of CLS
technique is based on the analysis of the reaction between wavelength and status of vegetation
using Crosta analysis
Five matrix Entering these matrix in a principal component analysis process produced Thresholding
34. • Image showing the distribution of level disease with it in field using PZ technique
• Using NDVI as reference to PZ technique
PZ NDVI
35. Validation
The objective of validation data
is to evaluate the quality of the data and measurements obtained from the analysis of
satellite imagery and then to select the best techniques that can be applied in reality and
benefit from them.
Filed assessment under control
inoculation plant of sugar beet with Cercospora beticola fungal
Spectra collection was started after the inoculation plant
In each treatment spectra from 12 plants inoculated and healthy plants were taken
Calculate vegetation indices and spectral signature
39. Conclusion
Ability to use thermal imaging for determine interaction between fungus C. beticola growth and
plant extract
Wild plant extract had very significant impact in CLS growth
There are high correlation between spectral reflectance and status of canopy where any change in
status of canopy Reflects on their spectral signature
The best vegetation indices can use for discriminating plant infected are and healthy are Chl
Red-edge
From results we can using geospatial technology to determine change and disease severity of
CLS disease by comparing the results obtain from field through survey
40. Deep learning is a branch of artificial intelligence (AI). It focuses on developing
algorithms and models that can learn to perform tasks and make predictions directly
from data
Using deep learning model to identify sugar beet fields and detecting
infected area using MSI index
Steps :-
1. Data preparation
1. label object
2. export training sample
2. Training deep learning
3. Detect sugar beet fields
4. Determine infected canopy using MSI index