REMOTE%20SENSING%20INDICATORS%20FOR%20CROP%20GROWTH%20MONITORING%20AT%20DIFFERENT%20SCALES

1
REMOTE SENSING INDICATORS FOR CROP GROWTH
MONITORING AT DIFFERENT SCALES

• Zongnan Li1, 2 and Zhongxin Chen1, 2
•  
• 1Key Laboratory of Resources Remote Sensing and
  Digital Agriculture, MOA, Beijing 100081
• 2Institute of Agricultural Resources and Regional
  Planning, CAAS, Beijing 100081

IGARSS 2011, Vancouver, 24-29 July, 2011
2
Outline

• ?. INTRODUCTION
• ?. DATA AND PROCESS
• ?. RESULT DISCUSSION
• ?. CONCLUSION

3
?. INTRODUCTION

• Crop growth is critical agricultural information.
  It can be used in the scientific management of
  crop and agricultural practice. It is also
  important in yield estimation and prediction
• There are several methods for crop growth
  monitoring, including in-situ field agronomic
  method, crop growth diagnostic model, and remote
  sensing method
• Remote sensing indicators are widely useed in
  vegetation monitoring
• Vegetation indices (VIs)are still the important
  indicators for regional crop growth monitoring

4

• Problem with VIs application
• Some VIs are sensitive to the soil background and
  non-vegetation fraction
• The scale effect
• different spatial resolutions
• spatial heterogeneity of land surface

5
Research Objectives

• Through testing the relationship between VIs and
  crop growth parameters, to investigate
• if there is/are optimal crop growth monitoring
  indicators at canopy scale and regional scale for
  different crop phenological stages
• if there are any trends for the relationship
  between VIs and crop growth parameters at
  different spatial scales

6
field experiment
canopy spectra
VIs
Relationsip between VIs and crop growth
crop parameters
Correlation analysis
crop yield
HJ-1 Imagery
Geom. Correction
Relationsip between VIs and LAI at different
scales
Atmos. Correction
scaling up
VIs at different scales
VIs
Correlation analysis
LAI in-situ
LAI regional
7
?. DATA AND PROCESS
research region
8
Field experiment plots in Langfang (11636'E,
3936'N).
Regional study in Hebei province
9
?. DATA AND PROCESS

• Field experiment and observation

5 levels for N fertilizer treatments 4 times
repeat
N application treatments N1- 0 N2- 15kg/ha
N3- 45 kg/ha N4- 105 kg/ha N5- 225kg/ha
10
?. DATA AND PROCESS

• Field experiment and observation

canopy spectra, LAI, foliar chlorophyll, plant
hight, coverage and biomass were measured at 5
phenological stages on 3/30, 4/14, 4/24, 5/5 and
5/17, 2009.
Canopy LAI
Chlorophyll SPAD
Canopy spectra
11
?. DATA AND PROCESS

• Field experiment and observation

early elongation stage
jointing stage
milk stage
heading stage
12
LAI evolution for various N applications
13
HJ-1A CCD Image 3/25/2009
Specification
Bands (µm) Blue0.43-0.52 Green0.52-0.60 Red0.63-0.69 infrared 0.76-0.90
Swath 360360km
Resolution 30m
HJ-1A CCD Image 4/21/2009
14
?. DATA AND PROCESS

• Caculation of VIs Correlation analysis

15
?. DATA AND PROCESS

• Processing of HJ-1 multi-spectral images

16
?. DATA AND PROCESS

• LAI Inversion (Beers law)

where KNDVI0.29 NDVI80.97 NDVIs0.11
17
LAI in study region
March 25 (elongation)
April 21 (heading)
18
High crop cover
Low crop cover
Canopy
19
?. RESULT DISCUSSION

• Remote sensing indicators for crop growth at
  canopy scale
• (sample sizes 20)

Date and Crop Stages 2009-3-30 2009-4-14 2009-5-5 2009-5-17
Date and Crop Stages early elongation stage jointing stage heading stage milk stage
NDVI 0.5173 0.8462 0.8778 0.9068
PVI 0.5484 0.6612 0.7033 0.8165
SAVI(L0.1) 0.5060 0.8447 0.8146 0.8993
SAVI(L0.2) 0.5494 0.8507 0.7815 0.8857
SAVI(L0.3) 0.5680 0.8229 0.7544 0.8857
SAVI(L0.5) 0.5504 0.8191 0.7416 0.8737
MSAVI 0.5504 0.8191 0.7484 0.8677
EVI 0.5504 0.8236 0.7379 0.8361
20
?. RESULT DISCUSSION

• Remote sensing indicators for crop growth at
  regional scales
• Low crop cover/the sample sizes n30.

Date 2009-3-25 early elongation stage 2009-3-25 early elongation stage 2009-3-25 early elongation stage 2009-4-21 heading stage 2009-4-21 heading stage 2009-4-21 heading stage
Resolution 240m 480m 960m 240m 480m 960m
PVI 0.9288 0.9362 0.9440 0.9592 0.9357 0.9536
SAVI(L0.1) 0.9431 0.9504 0.9723 0.9697 0.9643 0.9665
SAVI(L0.3) 0.9514 0.9486 0.9746 0.9689 0.9654 0.9686
SAVI(L0.5) 0.9472 0.9474 0.9722 0.9689 0.9638 0.9700
MSAVI 0.9440 0.9446 0.9714 0.9685 0.9621 0.9674
EVI 0.9262 0.9582 0.9472 0.9400 0.9361 0.9499
good but no obvious trend
21
?. RESULT DISCUSSION

• Remote sensing indicators for crop growth at
  regional scales
• High crop cover/the sample sizes n30.

Date 2009-3-25 early elongation stage 2009-3-25 early elongation stage 2009-3-25 early elongation stage 2009-4-21 heading stage 2009-4-21 heading stage 2009-4-21 heading stage
Resolution 240m 480m 960m 240m 480m 960m
PVI 0.9261 0.9450 0.9799 0.5750 0.6512 0.7261
SAVI(L0.1) 0.9536 0.9816 0.9943 0.9437 0.9512 0.9519
SAVI(L0.3) 0.9456 0.9726 0.9898 0.8247 0.8349 0.8936
SAVI(L0.5) 0.9394 0.9671 0.9888 0.7209 0.8006 0.8284
MSAVI 0.9408 0.9651 0.9877 0.7784 0.8260 0.8770
EVI 0.9125 0.9463 0.9639 0.7932 0.8072 0.8598
22
?. CONCLUSION

• At canopy scale, SAVI with different L values are
  suitable for winter wheat growth monitoring.
• At regional scale, soil adjusted vegetation
  indices have limitations in dense crop coverage.
• For dense crop coverage, the relationship between
  VIs improve with the increased pixel size, But
  this trend is not obvious for low crop coverage.

23
Acknowledgements

• The research was supported by the MOA 948 program
  project with contract no. 2010-S2 and 2009-Z31,
  and international corporation project from
  MOST(Ministry of Science and Technology of China
  ) with contract no. 2010DFB10030.

24
Thanks for your attention!