This document discusses using moderate resolution MODIS satellite data at high resolutions by calculating the observation footprints in polygon format rather than relying on the predefined grid. The authors simulate MODIS data over a study area in Hungary by taking the mean of high resolution SPOT satellite data within each MODIS observation footprint. They compare simple gridding versus assigning values to crop parcels based on the proportion of pixels within each. Their quality assessment shows correlation and errors are higher when using parcel-based approach versus gridded data, indicating moderate resolution data can provide high resolution information when using observation footprints rather than the grid.

1. Pushing MODIS to the edge:
High-resolution applications of
moderate-resolution data
Dániel Kristóf – Róbert Pataki –
András Kolesár – Tamás Kovács
MultiTemp 2013 | Banff, AB, Canada, 25-27 June 2013
Institute of Geodesy, Cartography and Remote Sensing
Budapest, Hungary

2. Outline
 MODIS at a glance
 Beginning
 Expectations and observations
 Particularities of MODIS data
 New preprocessing approach & first results
 Current state
 Time series construction and quality measures
 Results achieved so far
 Conclusions and outlook
2

3. MODIS








Moderate Resolution Imaging Spectroradiometer
Onboard NASA’s Terra&Aqua satellites
36 spectral bands between 0.405 and 14.385 micrometers
Wide field of view, daily (or even more frequent) coverage
with nominal nadir resolutions of 250, 500 and 1000 meters
Sophisticated operational data processing (MODAPS)
A large number of preprocessed data & scientific products
available (for free…), timeliness
Well-published algorithms, systematic revision and
reprocessing (also backwards processing; Currently:
„Collection 5”)
An archive continuous in time: (almost) gapless data since
1999!
3

4. Beginning: initial motivation…
 A user wants to detect the date of interventions on
agricultural fields, size comparable to a 250-m MODIS pixel
Band 2
Band 2,
QC 4096
(OK)
4

5. Beginning: initial motivation…
 Terra MODIS land
surface reflectance
Day x+1
Day x+1
 Aqua MODIS land
surface reflectance
Day x
Day x
5

6. Beginning: initial motivation…
Aqua MODIS
• Terra & Aqua MODIS, the same day
Terra MODIS
6

7. Why…? MODIS particularities:
– Gridding artifacts:
• Data stored in a predefined grid, resampled
• Average overlap between observations and
their respective grid cells less than 30% [Tan et
al., RSE 105(2006):98-114]
– Problems with the raster data model itself:
• Fixed size and orientation of the cells although
the observation dimensions vary across the
scene due to the wide field of view (+/- 55
degrees), orientation definitely not N-S
• The grid cells / pixels do not represent the area
where the signal is originated from
7

8. MODIS particularities
Close to nadir
Close to swath edge
8

9. Proposed solution
• A possible solution would be the spatial and/or temporal
compositing and/or filtering, but: loss of resolution and
information…
• Our approach: Change data representation
– Calculate and store observation footprints in polygon format
– Each polygon represents the respective observation footprint (“real
pixel”) sensed during image acquisition
• Geolocation datasets (MOD/MYD03) contain all necessary
info to do this
– Ground location, dimensions and orientation of each MODIS pixel
footprint can be determined from:
Latitude, Longitude, Height, Sensor Zenith Angle, Sensor Azimuth
Angle, Slant Range
– Geolocation accuracy: 50 m at 1 sigma at nadir
– Swath images (MOD/MYD02) or „backsampled” Surface Reflectance
9

10. Proposed solution
What does it
look like?
1 km
250 m
10

11. First results: correlation with SPOT 2data (NIR band)
2
R = 0.4702
R = 0.6117
R2 = 0.4702
R2 = 0.6117
R2 = 0.7910
R2 = 0.7918
MODIS at 1 km resolution
MODIS at 250 m resolution
11

12. First results
• Possible application: one-step radiometric
normalization of high-resolution imagery
by using same-day MODIS surface
reflectance.
• What next? How to handle ever-changing
observation geometries as a time series?
12

13. And then…?
TIME SERIES CONSTRUCTION
13

14. So we selected a study area in SE Hungary…
14
SPOT4 image, 06/08/2003

15. So we selected a study area in SE Hungary…
15
SPOT4 image NIR band, 06/08/2003

16. …downloaded some MODIS geolocation files
and calculated 250-m geometries…
16
MODIS geometry, 2003216_0945

17. …downloaded some MODIS geolocation files
and calculated 250-m geometries…
17
MODIS geometry, 2003217_0850

18. …downloaded some MODIS geolocation files
and calculated 250-m geometries…
18
MODIS geometry, 2003217_1030

19. …downloaded some MODIS geolocation files
and calculated 250-m geometries…
19
MODIS geometry, 2003218_0935

20. …downloaded some MODIS geolocation files
and calculated 250-m geometries…
20
MODIS geometry, 2003219_1020

21. …downloaded some MODIS geolocation files
and calculated 250-m geometries…
21
MODIS geometry, 2003220_0925

22. …downloaded some MODIS geolocation files
and calculated 250-m geometries…
22
MODIS geometry, 2003221_1005

23. …downloaded some MODIS geolocation files
and calculated 250-m geometries…
23
MODIS geometry, 2003222_0910

24. …downloaded some MODIS geolocation files
and calculated 250-m geometries…
24
MODIS geometry, 2003223_0955

25. …downloaded some MODIS geolocation files
and calculated 250-m geometries…
25
MODIS geometry, 2003224_0900

26. …downloaded some MODIS geolocation files
and calculated 250-m geometries…
26
MODIS geometry, 2003224_1035

27. …then simulated „pure” MODIS data as the
zonal mean of SPOT NIR pixels…
27
MODIS footprints on SPOT background, 2003216_0945

28. …then simulated „pure” MODIS data as the
zonal mean of SPOT NIR pixels…
28
MODIS footprints with SPOT-derived values, 2003216_0945

29. …then simulated „pure” MODIS data as the
zonal mean of SPOT NIR pixels…
29
MODIS footprints on SPOT background, 2003217_0850

30. …then simulated „pure” MODIS data as the
zonal mean of SPOT NIR pixels…
30
MODIS footprints with SPOT-derived values, 2003217_0850

31. Setting the baseline:
THE GRIDDING PROCESS
31

32. Simulated MODIS data was then resampled
according to „simple gridding” (NN)
32
SPOT4 image NIR band, 06/08/2003

33. Simulated MODIS data was then resampled
according to „simple gridding” (NN)
33
SPOT4 image & predefined sinusoidal MODIS grid

34. Simulated MODIS data was then resampled
according to „simple gridding” (NN)
34
Simulated MODIS observations, geometry: 2003216_0945

35. Simulated MODIS data was then resampled
according to „simple gridding” (NN)
35
Simulated MODIS observations, geometry: 2003217_0850

36. Zoom-in: grid cell vs. observation geometry
36

37. Zoom-in: grid cell vs. observation geometry
37

38. Zoom-in: grid cell vs. observation geometry
MODIS observation from 2003216_0945
and the corresponding grid cell
38

39. Zoom-in: grid cell vs. observation geometry
MODIS observation from 2003217_0850
and the corresponding grid cell
39

40. Zoom-in: grid cell vs. observation geometry
MODIS observation from 2003218_0935
and the corresponding grid cell
40

41. Zoom-in: grid cell vs. observation geometry
MODIS observation from 2003219_1020
and the corresponding grid cell
41

42. Zoom-in: grid cell vs. observation geometry
MODIS observation from 2003220_0925
and the corresponding grid cell
42

43. Zoom-in: grid cell vs. observation geometry
MODIS observation from 2003221_1005
and the corresponding grid cell
43

44. Zoom-in: grid cell vs. observation geometry
MODIS observation from 2003222_0910
and the corresponding grid cell
44

45. Zoom-in: grid cell vs. observation geometry
MODIS observation from 2003223_0955
and the corresponding grid cell
45

46. Zoom-in: grid cell vs. observation geometry
MODIS observation from 2003224_0900
and the corresponding grid cell
46

47. Zoom-in: grid cell vs. observation geometry
All MODIS observation geometries
corresponding to the selected grid cell
47

48. The other approach:
SURFACE OBJECTS AS
OBSERVATION UNITS
48

49. 49

50. Land surface objects of interest:
observation units defined a priori
50
Crop map: 512 crop parcels delineated on HR data

51. Land surface objects of interest:
observation units defined a priori
51
Crop map: 512 crop parcels delineated on HR data

52. Relevance („purity”) of each MODIS
observation (~pixel) to each observation unit
52
Portion of observation belonging to the given parcel

53. Relevance („purity”) of each MODIS
observation (~pixel) to each observation unit
53
Portion of observation belonging to the given parcel

54. Only „pure” observations (above threshold)
are selected
54
Green: retained, Red: rejected

55. Only „pure” observations (above threshold)
are selected
55
Green: retained, Red: rejected

56. Then, radiometric values are assigned to each
observation unit (parcel) for each MODIS overpass
56
Calculation based on area-weigthed mean of retained pixels

57. QUALITY ASSESSMENT
57

58. 70
120
60
100
50
40
30
100
90
80
70
60
50
40
30
20
10
0
Quality assessment: time-series comparison
80
mean_spot
20
mean_spot
60
mean_grid_simple
mean_grid_simple
40
mean_modis_limit
10
mean_modis
20
mean_modis
0
mean_modis_limit
0
100
90
80
70
60
50
40
30
20
10
0
mean_spot
mean_grid_simple
atlag_gridbol
mean_modis
mean_modis_limit
mean_spot
mean_grid_simple
mean_modis
mean_modis_limit
145
120
140
100
135
80
130
mean_spot
0
mean_grid_simple
115
mean_modis
mean_modis_limit
20
mean_spot
120
mean_modis
40
125
mean_grid_simple
60
110
mean_modis_limit
105
64
62
60
140
58
56
120
mean_spot
54
100
mean_grid_simple
52
50
48
80
mean_modis
mean_modis_limit
60
46
40
20
mean_spot
mean_grid_simple
mean_modis
mean_modis_limit
0
160
120
140
100
120
80
100
80
mean_spot
60
mean_grid_simple
40
mean_spot
mean_modis
20
0
mean_modis_limit
60
mean_grid_simple
40
atlag_gridbol
20
mean_modis
0
58
mean_modis_limit

59. Quality assessment: statistics
Observation
time
2003216_0945
2003217_0850
2003217_1030
2003218_0935
2003219_1020
2003220_0925
2003221_1005
2003222_0910
2003223_0955
2003224_0900
2003224_1035
Number Parcels Correlation Correlation Correlation Mean diff Mean diff Mean diff Max diff Max diff Max diff
of
with
gridded vs. all MODIS filtered
(~RMSE) (~RMSE) (~RMSE) (~RMSE) (~RMSE) (~RMSE)
parcels valid
reference vs.
MODIS vs. gridded all MODIS filtered gridded all
filtered
obs.
reference reference
MODIS
MODIS MODIS
512
466
0,85
0,88
0,94
6,83
7,60
5,14
75,66
37,13
29,85
512
333
0,73
0,74
0,89
10,60
10,71
7,72
51,57
40,87
40,18
512
207
0,70
0,69
0,94
10,86
11,28
6,26
58,88
44,75
29,97
512
463
0,86
0,89
0,95
6,57
7,42
4,74
63,79
41,14
36,52
512
310
0,79
0,78
0,94
9,26
9,74
5,45
41,75
43,33
33,02
512
460
0,85
0,87
0,93
7,39
8,01
5,59
41,89
38,60
35,21
512
381
0,80
0,84
0,95
8,45
8,83
5,11
58,21
41,08
26,44
512
419
0,82
0,84
0,92
8,19
8,87
6,22
58,66
43,31
31,57
512
444
0,88
0,88
0,95
6,97
7,84
5,09
41,22
35,14
30,65
512
363
0,74
0,77
0,91
9,83
10,05
6,86
69,21
49,45
29,01
512
142
0,64
0,63
0,93
11,89
12,01
6,09
55,05
43,72
28,71
59

60. Quality measure: number of retained pixels
2003216_0945
60
Green: high, Yellow: medium, Orange: low, Red: zero

61. Quality measure: average pixel proportion
2003216_0945
61
Green: high, Yellow: medium, Red: low, White: N/A

62. Quality measure: number of retained pixels
2003224_1035
62
Green: high, Yellow: medium, Orange: low, Red: zero

63. Quality measure: average pixel proportion
2003224_1035
63
Green: high, Yellow: medium, Red: low, White: N/A

64. Area vs. accuracy
2003226_0945
30
10
20
diff_spot_modis
20
10
0
0
diff_spot_modis
30
40
2003224_1035
0
500000
1000000
1500000
area
2000000
2500000
3000000
0
500000
1000000
1500000
2000000
2500000
area
64
3000000

65. Correlation and RMS errors
N = 512
r = 0,638
RMSE = 11,89
80
60
80
mean_modis_limit
100
100
120
120
2003224_1035
40
60
mean_grid_simple
N = 142
r = 0,934
RMSE = 6,09
40
60
80
100
mean_spot
120
140
40
60
80
100
120
140
mean_spot
65

66. Correlation and RMS errors
N = 512
r = 0,855
RMSE = 6,82
100
60
60
80
80
mean_modis_limit
100
120
120
140
140
2003216_0945
40
mean_grid_simple
N = 466
r = 0,943
RMSE = 5,13
40
60
80
100
mean_spot
120
140
40
60
80
100
120
140
mean_spot
66

67. Pixel proportion, number of pixels, and RMSE
10
15
diff_spot_modis_limit
15
10
5
5
0
0
diff_spot_modis_limit
20
20
25
25
2003224_1035
0.5
0.6
0.7
mean_pix_prop_limit
0.8
0.9
5
10
15
20
pixel_count_limit
67

68. Pixel proportion, number of pixels and RMSE
20
10
15
diff_spot_modis_limit
15
10
5
5
0
0
diff_spot_modis_limit
20
25
25
30
30
2003216_0945
0.5
0.6
0.7
0.8
mean_pix_prop_limit
0.9
1.0
0
10
20
30
40
50
pixel_count_limit
68
60

69. CONCLUSIONS AND
OUTLOOK
69

70. Conclusions
• Gridding is a major source of inaccuracy and noise
• New polygon representation of MODIS observations has
significantly increased correlation with same-day highresolution data: better representation of observation
geometry
• Landscape objects delineated a priori can be efficiently
used to construct time series
• Data processing requires more computing power, but the
user has full control over the process (thresholds, quality
measures, etc.) and can obtain more accurate data with
maximal spatial and temporal coverage
70

71. Outlook
• Currently: „proof of concept” level - further work
on real case studies needed
• Point Spread Function is another source of
inaccuracy – could be integrated
• Other data representations (Spline
surface, SensorML, etc.) are also to be
assessed
• Data processing speed is already good
(Python/PostgreSQL/PostGIS), but can be
further improved by e.g. parallel computing in
cloud architecture: implementation of IQmulus
project service
71

72. Thank you!
Questions?
A High-volume Fusion and Analysis
Platform for Geospatial Point
Clouds, Coverages and Volumetric
Data Sets
Dániel Kristóf
kristof.daniel@fomi.hu
www.iqmulus.eu
IQmulus (FP7-ICT-2011-318787) is a 4-year Integrating Project (IP) in the
area of Intelligent Information Management within ICT 2011.4.4 Challenge
4: Technologies for Digital Content and Languages. IQmulus started on
November 1, 2012, and will finish October 31, 2016.
72