Remote%20Sensing%20Applications%20In%20Identifying%20Crop%20Infestation

1
Remote Sensing Applications In Identifying Crop
Infestation EES5053 Fall 2006 Meaghan
Metzler December 5, 2006
2
Outline

• Background and Overall Objectives
• Method Description
• Data Used
• Results
• General Conclusions

3
Background

• Remote sensing is the process whereby information
  can be obtained from a source without actually
  coming in contact with it.
• Popular applications include water management,
  homeland security, disaster management, and
  agriculture efficiency.

4
Agricultural Efficiency

• Crop stress associated with pest infestation is
  predominantly monitored through human visual
  inspection.
• Pest infestation is not normally evenly
  distributed within a field.
• Pesticide application is expensive and excessive
  application can cause resistance in pest species
  and damage to the environment

5
Overall Objective

• RS techniques play an important role in timely
  detection and early warning of disease and insect
  pests El-Khatib et. al.
• Damage (or stress) can be identified by changes
  in the spectral signature of crop species.

6
Case-In-Point 1 RiceInfested with Sheath Blight

• Location Hazen City, Arkansas
• Field size 762m x 467m
• Infestation naturally infected
• with sheath blight, caused by Rhizoctonia solani

7
Methodology

• Field was divided into 11 strips with 50 sampling
  sites for observing the severity of the disease
  infestation.
• 3 indices used
• SHBDI infected tillers
• SHBDH height (cm) of sheath blight symptoms
• TH height (cm) of plant canopy above soil
• Ground disease index (DI) SHBDI SHBDH / TH

8
Data Used

• Airborne RS images (7/12, 7/24/, 8/04, and 8,17)
• 4 bands 450-540nm, 530-600nm, 610-680nm, and
  780-1000nm.
• Resolution 1m
• DN 0-255

9
Results

• The change of DI was proportional to that of the
  change of DN values however, it was weak.
• It was difficult to discriminate from healthy
  plants and those with only a light infestation.
• Higher spatial resolution and spectral resolution
  is required to diagnose crop disease efficiently
  and accurately.

10
Case-In-Point 2 Rice Infested with Blast

• Location nursery field at Nattaf
• Research Station, Giza, Egypt
• Field size 70 faddans (one faddan 1.038 acres)
• Infestation infected with bast, caused by
  Pyricularia oryzae

11
Methodology

• Reading were taken from a spectroradiometer
• The reading were taken to measure simultaneous
  reflected radiance flux in four wavelength of
  LANDSAT Multispectral Scanner (MSS) band
  0.5-0.6µm, 0.6-0.7µm, 0.7-0.8µm, and 0.8-1.1µm
  El-Khatib et. al.
• Ratio Vegetation Index (RVI) and Normalized
  Difference Vegetation Index (NDVI) were
  calculated

12
Results

• The most noticeable changes in the reflection
  occurred first in the NIR region.
• In the primary stages of blast infestations there
  was a decrease in both the green and NIR bands.

13
Case-In-Point 3 Rio Red Grapefruit infested
with Greasy Spot

• Location Texas AM
• University-Kingsville Citrus
• Center in Weslaco, TX
• Field Size about 9.2 acres of grapefruit trees
• Infection Mycosphaerella citri

14
Methodology Data Used

• Airborne multispectral and hyperspectral images
  were taken in May 2004
• Multispectral
• Cameras with NIR 845-857nm filter, red 625-635nm
  filter, and yellow-green 555nm-565nm filter.
• 0.7m spatial resolution

15
Data Used cont.

• Hyperspectral
• Pushbroom camera system
• 128 spectral bands 457.2nm-921.7nm
• Spectral resolution 3.63nm
• Spatial resolution 1.3m

16
Results

• The spatial resolution using the hyperspectral
  airborne camera system (1.3m) is rougher, it made
  individual trees no longer discernable from one
  another
• With the increase in bands more detail could be
  recorded so there was more classes when
  unsupervised classification was done
• Hyperspectral system more precisely determines
  the health condition of the trees

17
Importance of RS to Agriculture

• Can be used to determine vegetation conditions
  more accurately
• Saves time
• Saves wear and tear on farming equipment
• Saves money
• Helps limit pesticide use

18
General Conclusions

• By reviewing these experiments the similarities
  of each was that because satellite technology
  does not have very high spatial resolution needed
  that other methods must be implemented such as
  airborne multispectral hyperspectral platforms,
  or handheld hyperspectral spectroradiometer
• From these experiments baseline recordings of
  spectral characteristics can be recorded to
  validate future recordings from space-borne
  platforms.

19
Questions?