Dr. Daniel Vidal-Madjar
C.R.P.E.
U.V.S.Q.
10-12 Avenue de l'Europe
78140 Velizy
France
Co-Investigators:
M. Normand, CEMAGREF
O. Taconet, CETP/CNRS
S. Mascle, CETP/CNRS
M. Zribi, CETP/CNRS
C. Emblanch, CEMAGREF
C. Loumagne, CEMAGREF
Test of Roughness and Moisture Algorithms Using Multiparameter Spaceborne SAR and Application to Surface Hydrology


 OBJECTIVES
Evaluate the usefulness of radar-derived parameters in surface hydrology.

Demonstrate the usefulness of the squint mode in the case of bare soil observations.

Compare various roughness/moisture algorithms in a real space imaging mode.

PROGRESS
The Orgeval SIR-C/X-SAR Experiment in April 1994

The area surveyed by SIR-C/X-SAR included two monitored watersheds, Mélarchez and Orgeval, covering 5x5 km2 of intensive agricultural plains east of Paris (France). In April 94, the majority of fields were bare soils, except for wheat (20 cm high). During the 5 days of SAR passes, soil moisture remained high and constant (0.35cm3/cm3) over the watersheds and soil roughness and practices were the remaining variables.

A detailed survey was done and a crop map generated (70 fields of more of one ha) shown on Figure 1. Photography of soil surfaces was done over 50 bare fields. Ten fields (numbered one to 10) were intensively characterized and were representative of the region's cultures (four of wheat, two of peas, three of corn, one ploughed field): soil moisture and density data (gravimetric and neutronic), height profiles by pin-profilers, penetrometer. A SPOT image (from June 1994 ) was used to assess the field boundaries. The 70 fields were classified by roughness aspects (Table 1).

The SAR images' incidence angles range from 44° to 57°. Simultaneously, several flights of the copolar scatterometer ERASME of the CETP ( C- and X-bands ) were done in SAR site direction and over the tested region, keeping flights perpendicular and parallel to the row directions over test fields. The ERASME incidence angles were from 25° to 50°.

Intercomparison of SIR-C/X-SAR cross-sections was done over natural targets, on fields seen on the same incidence angle and the same view azimuthal angle. The intercalibration of the three radars was within 2 dB at maximum (Figures 2 and 3).


SIGNIFICANT RESULTS
Polarimetric signatures in multiconfiguration conditions (three frequencies, L, C, X, multi-incidence, and polarization ) were studied, following two approaches:

- a global analysis of the ability of L- and C-band to discriminate (thanks to unsupervised classification algorithms) the polarimetric signatures of different land cover types, for crop map elaboration in an early stage of plant growth.

- the test of existing models of backscattering over agricultural surfaces and the appro- priateness of surface descriptions.

Mono-Band and Multi-Band Classification

 First, unsupervised mono-band mono-incidence classifications were performed on HH and HV SIR-C images. Cluster characteristics were determined using the fuzzy c-means algorithm and then images were classified according to the Maximum A Posterior criterion. The numbers of clusters were four for C-band classification and six for L-band classification. Then identification rates of the main cover land types were computed. For each of the four land cover types: forest, wheat, ploughed land and seedling, figure 4 shows the C-band and figure 5 shows the L-band identification rates versus incidence angle, and the classification algorithm used: supervised (called "svs") or unsupervised (called "nsvs"), blind (subscript "0") or MAP (subscript "1"). The main results are:

a- regularization step (MAP classification) reduces classification errors;
b- L-band performs better in soil cover type discrimination than C-band;
c- identification rates in C-band increase with incidence angle (between 44° and 57° );
d- ploughed land identification in L-band decreases with incidence angle.

Then, complementarity and redundancy between L- and C-bands were studied in a quantitative way. Figure 6 shows the redundancy rate (mutual information normalized by the image entropy) between L- and C-band unsupervised classification results, versus incidence angle. We found that redundancy increases when incidence angle increases between 44° and 57° . However at an incidence angle of 57° , redundancy between L- and C-band is only about 30%.

Recalling that, in C-band, land cover identification rates are not as good as in L-band, but increase with incidence angle, we may assume that the increase of redundancy between L- and C-band versus incidence angle is mainly due to the increase of "useful" information (which is probably due to increase of contrast or reduction of speckle noise) in C-band images.

FInally, data fusion between L- and C-band was performed to improve identification rates. The aim of data fusion is to use redundancy between images to reduce classification errors and to use complementarity to put in evidence new classes. The unsupervised multi-source classification algorithm used is based on Dempster-Shafer evidence theory. Figure 7 shows the identification rates obtained after data fusion, and improvement in identification of land cover types clearly appears.

The conclusion is that multi-band L and C can be successfully used to discriminate the different land cover types. Similar results were obtained with AIRSAR data for discrimination of the different culture types in Summer 1991.

 Backscattering modeling and surface description

The main results are:

- There is a systematic azimuthal dependence of copolar radar cross sections (HH, VV) over the same field with the view angle towards the row direction in C- and X-bands, as well for smooth fields of peas eroded by rains and for rough ploughed fields. It is obtained by complementary view angles of targets with ERASME and SIR-C/X-SAR or with fields of very similar roughness seen by SIR-C/X-SAR for smooth fields. The mean difference between perpendicular and parallel direction is about 2 dB, which is large compared to dielectric constant variations occurring between dry and moist soils (Figure 8). Points are from fields with incidence angles between 25 to 57° . In L-band the restricted number of cases prevents any conclusions.

- The complementary use of the ERASME scatterometer and SIR-C/X-SAR allows study in copolarization of the variation of radar cross sections with incidence angle. Over the 10 test fields, the surface roughness parameters are deduced from pin profiler data, parallel and perpendicular to rows and using the Shin and Kong's 1984 quasi periodic description of agricultural soils. Small spatial scale (clods) and large (rows) parameters are summed in Table 2 and localized in ks/kl space (k the wave number, s the r.m.s. h eights, l the length of correlation at small scale) on Figure 9. In L-band, all test fields are within IEM validity range. In C-band, only smooth ones (peas) are within. Most other ploughed fields are out of range. Field W3 is taken for eroded smooth soil, W2 for smooth ones, and C9 for ploughed ones, remaining within IEM limits.

Results of the comparison between IEM simulations and ERASME, SIR-C/X-SAR data are shown on Figures 10 and 11:

- There is in L-band a good agreement for smooth fields and a 5-dB gap underestimation by the model for the rough one.

- The agreement in C-band remains good for smooth fields between 25 to 40° . For larger incidence angles (40° to 57° ), there is a systematic overestimation of the model. It is more pronounced in X-band where cross sections keep decreasing with incidence angle, even on the rougher C9 surface (-15 dB from 25° to 57° ), as the model gives a flat response.

To explain these differences, they must be analyzed in two directions, the physical hypothesis of backscattering model, and also the adequacy of the surface used in them. For smooth ones, the surface remains a randomly quasi-sinusoidal surface, with a short range of spatial frequencies to describe it. For very rough soils, surfaces become discontinuous and other scattering models have to be considered, and soil-description parameters to be defined.


FUTURE PLANS
Unsupervised polarimetric SAR image classification

The following activities are proposed.

- A study of the complementarity of polarimetric components and comparison of classification results using different combinations of polarimetric information has begun. The aim is to provide a quantitative measurement of the increase of information provided by each polarimetric component (HH, HV, VV, amplitude and phase), versus land cover type.

- We aim at studying the influence of speckle filtering on classification results, and to derive a new speckle filter using unsupervised classification results, in an iterative way.

Surface Scattering Process

The following activities are proposed.

- Over rough surfaces, a PHD has begun in 1995 to define a new description of discontinuous surfaces and modeling in intensity in co- and cross-polarizations. This work gathers at CETP the remote sensing team of Dr. D. Vidal-Madjar and the team of Pr Lavergnat on Theoretical Electromagnetism. Various ground surfaces are studied and analyzed by geometrical image algorithms in collaboration with Dr. P. Boissard of NRA/Grignon (Institut National de Recherche Agronomique) and with Dr. M. Chapron of ENSEA (Ecole Nationale Superieure de l'Electronique et de ses Applications).

- The quantification of azimuthal dependence of cross sections with rows. Two approaches are foreseen. The first one by analysis of surface description with a larger frequency spectrum. The second one by analysis of different databases, obtained over "Pays de Caux" in 1995 with a large multiplicity of radar configurations and completed by the SIR-C/X-SAR Orgeval campaign. A polarimetric campaign over bare soils is forecasted in 1997 with the two scatterometers of CETP in the mentioned frame (EU proposal RESEDA in the south of France).

TABLE I: Classification of the Orgeval fields by soil cover and roughnesses

CLASS FIELDS DIRECTION
TOWARDS
ROWS FOR
SIR-C/X-SAR
WHEAT 2,3,71,78
4,5,62
67,83,85,68
Perpendicular
Parallel
Between
ERODED SMOOTH
BARE SOILS 		10,12
40
23,18,31,42
Perpendicular
Parallel
Between
SMOOTH BARE
SOILS 		35
11,46
16,47,53
Perpendicular
Parallel
Between
PLOUGHED
FIELDS 		6, 8
1,9,7,49
Perpendicular
Parallel

TABLE 2: Roughness parameters

Field s l S L P ks
(L)		 kl
(L)		 ks
(C)		 kl
(C)		 ks
(X)		 kl
(X)
C1 2.66 6.1 0 0 0 0.70 1.6 2.95 6.77 5.28 12.1
W2 0.70 5.97 0.97 34.69 99.58 0.18 1.57 0.78 6.63 1.40 11.9
W3 0.55 12.68 0.73 143.38 197.31 0.14 3.33 0.61 14.08 1.09 25.2
W4 0.92 2.51 0.83 42.37 104.56 0.24 0.66 1.02 2.79 1.82 4.8
L6 3.77 7.35 0.94 128.87 94.35 0.99 1.93 4.19 8.16 7.5 14.6
C7 2.49 8.1 1.73 147 45.2 0.65 2.12 2.76 8.99 4.94 16.9
P8 1.31 3.52 1.44 93.67 43.0 0.34 0.92 1.45 3.9 2.59 6.9
C9 0.97 8.07 1.63 115.26 40.22 0.25 2.12 1.07 8.96 1.91 16.2
P10 0.62 5.66 0.3 172.08 120.68 0.16 1.48 0.68 6.28 1.21 11.2

W:Wheat, C: Corn, P: Peas, L: Ploughed field





FIGURE 2-a

FIGURE 2-b


FIGURE 3



FIGURE 4
Identification rates of (a) forest, (b) wheat, (c) ploughed land, (d) seedling versus incidence angle, provided by C-band classification.



FIGURE 5
Identification rates of (a) forest, (b) wheat, (c) ploughed land, (d) seedling versus incidence angle, provided by L-band classification.


FIGURE 6
Redundancy rate between L- and C-band classification results vs incidence angle, case of unsupervised classification.





FIGURE 7
Identification rates of (a) forest, (b) wheat, (c) ploughed land, (d) seedling versus incidence angle, provided by multi-band L-C-band classification.


ORGEVAL'SIR-C/X-SAR APRIL '94 EXPERIMENT

(a)

ORGEVAL'SIR-C/X-SAR APRIL '94 EXPERIMENT


(b)
FIGURE 8



(a)

(b)

(c)
FIGURE 9

(a) very smooth soil

(b) smooth soils with clods

(c) ploughed field

FIGURE 10

The first line figures are for very smooth soil, the second line ones for smooth surface with clods and the third line ones for ploughed field
FIGURE 11
PUBLICATIONS
1 TECHNICAL REPORT

[TR.1] EMBLANCH, C., Stage de DEA en Hydrologie, Premiers résultats de la campagne de Télédétection Radar Orgeval 94 pour la détermination de l'humidité de surface des sols, University Orsay-Paris XI, September 1994.

[TR.2] ZRIBI, M., Interpretation de l'imagerie radar polarimetrique pour l'hydrologie (Campagnes de la Nasa: Mac Europe'1991 et SIR-C/X-SAR'1994), Stage d'Ecole d'Ingénieurs, ENSICA-Toulouse, 22 June 1995.

2 CONGRESS COMMUNICATIONS

[AC.1] Mascle, S., D. Vidal-Madjar, M. Zribi, and O. Taconet, Comparison between L- and C- bands SIR-C polarimetric data versus incidence angle, in Proceedings of International Symposium: "Retrieval of bio- and geophysical parameters from SAR data for land applications" , in Toulouse, France, on October 10-13 1995.

[AC.2] Zribi, M., O. Taconet, D. Vidal-Madjar, S. Mascle, C. Loumagne and M. Normand, Backscattering response of bare soils with roughness from combination of SIR-C/X-SAR imagery and ERASME airborne scatterometer data (Orgeval'94 site), PIERS96 , University of Innsbrück, Institute for Meteorology and Geophysics, Innsbrück, Austria, 8-12 July 1996.

[AC.3] Taconet, O., D. Vidal-Madjar, C. King, Y. Le Bissonnais, M. Zribi, S. Mascle, C. Loumagne and M. Normand, Soil backscattering behaviour with roughness from combination of SIR-C/X-SAR imagery and airborne scatterometer data (ERASME and RENE), IGARSS'96 , Burnham Yates Conference Center, Lincoln, Nebraska, USA, 27-31 May 1996.

3 ARTICLE

 [AR.1]
Zribi, M., O. Taconet, D. Vidal-Madjar, S. Mascle, C.Loumagne and M.Normand, Backscattering response with direction angle relative with soil practices from combination of SIR-C/X-SAR imagery and ERASME airborne scatterometer data (Orgeval'94), in preparation for Remote Sens. of Environment .

Table of Contents

Converted to HTML by Alvin Wong, al.wong@jpl.nasa.gov

The Jet Propulsion Laboratory
4800 Oak Grove Drive
Pasadena, Cailfornia 91109