In The Netherlands a digital SLAR system (X-band) is used since many years in the radar remote sensing

Transcription

1 Abstract: A POLARMETRC ARCRAFT N NETHERLANDS P. Hoogeboom FELTNO P.O.box The Hague The Netherlands n The Netherlands a dgtal SLAR system (Xband) s used snce many years n the radar remote sensng research. Ths research programs n whch also other s used, for nstance a multband arborne scatterometer. Based on our experence n radar remote sensng a metrc Cband arcraft SAR s desgned, to be realsed n the 1990's. The system s of a novel des usng a array antenna wth sold state amplfers. The dataprocessng for ths requres the devel of software tools, that take geometrc and radometrc correctons nto account, as well as the calbraton. n the md seventes radar experments usng a short range scatterometer were started n The Netherlands, manly for the determnaton of radarsgnatures of agrcultural crops. Ths work was performed at Xband (A=3 em) and later on also at 36 GHz (X=8 mm) [1]. The measurements were taken durng a number of ng seasons to gan more nsght n the varatons from year to year. Ths led to work n the area of modelng and crop classfcaton. t was soon zed that calbraton of the data taken at dfferent tmes and under varyng ncdence s very but also te dffcult to acheve. The work descrbed above was carred out by a number of nsttutes, each one coverng a specfc dscplne. Ths multdsc nary was zed through the ROVE team (Radar Observaton on on). Nowadays the Netherlands Remote Sensng Board takes care of sponsorng ths ROVE work. The ns that was ned wth ths program led to the use of a d and calbrated Xband SLAR system. Among others the multtemporal crop classfcaton could be demonstrated wth ths system [2]. t s dscussed n more detal n 2. n the 1980's t was realsed that the work had to be extended to more and also lower frequences and to other areas of nterest. These dea's led to the desgn and the use of the DUTSCAT multband arborne scatterometer, whch s dscussed n chapter 3. The next step wll be the development of an arborne polarmetrc SAR system n the Cband, called Pharus {Phased.rray al ). The choce of the parameters for ths are based on the we ganed wth the prevous programs. The was chosen to be the same as wll be used on the ERS1 satellte. At present only a frst desgn exsts, necessary to rase the funds. We hope to realze the

2 fnal system by More detals about the Pharus project are gven n chapter 4. 2.SLAR expprments. For many years the radar remote sensng research n The Netherlands has been usng a dgtal Xband SLAR system to demonstrate varous possbltes of radar. Some propertes of ths system are gven n table 1. frequency 9.4 GHz (Xband) transmtted power 25 kw pulselength ns ( m) polarsaton HH antenna beamwdth 10 mrad (.6 ) twoway P.R.F. 200Hz recordng 4096 * 8 bt 50 MHz (3 m) nternal delay lne calbraton Table 1: Specfcaton of SLAR system parameters By usng an nternal delay lne calbraton the system s capable of delverng calbrated mages. The geometrc and radometrc dstortons n the captured data are corrected off lne wth a computer. Ths process makes use of the smultaneously recorded arcraft atttude and poston parameters, that are delvered by an nertal navgaton system. After correcton and resamplng mages of hgh qualty become avalable [3]. A hgh mage qualty s often desrable, for nstance n the case of multtemporal crop classfcaton. 07/ /07/ Fgure 1: Xband SLAR flghts over a testarea n the Flevopolder. mage dmensons: 3.7x6.2 km. Alttude 660 m.

3 Although the dscrmnaton between varous agrcultural crop types n one radar mage sometmes s dffcult, the task can be facltated by combnng several mages, taken under dfferent ncdence angles or at dfferent tmes n the growng season. To enable a prosperous dentfcaton t s necessary that the mages can be overlayed on to each other and that the reflecton coeffcents (grey values) may be compared between mages. f these condtons are met, then useful results may be expected. n fgure 1 a set of mages from the SLAR are shown, taken n the n The Netherlands at 3 dfferent tmes. Usng a multtemporal classfcaton procedure we were able to correctly classfy over 90% of ths area [2,4]. 3.Multband Arborne scatterompter expermpnts. Detaled radar sgnature studes and modelng work requre tools to measure backscatter coeffcents as a functon of frequency, ncdence angle and polarsaton. The DUTSCAT multband arborne scatterometer, desgned and bult by the Delft Unversty of Technology s such a tool, capable of acqurng data at 6 frequences rangng from 1 18 GHz. n the past a groundbased scatterometer program was carred out n The Netherlands. The datasets that came out can now be extended by the use of ths new tool. These datasets form the bass for the knowledge that s necessary to evaluate new applcatons n the feld of remote sensng. 198? DUTSCA!T ~TA for SUGARBEETS MLFLUO F 2: example of the DUTSCAT scatterometerdata captured durng the Agrscatt 87 campagn. 1

4 2 shows a 3 dmensonal of the data wth the DUTSCAT durng the 1987 AGRSCATT over a testarea n the The shows the backscatter for n twelve blocks. The upper 6 blocks are the HH arsed measurements, the lower ones VV. block a the Xaxs shows the data for a certan Wthn a block the ncdence s a the Xaxs and tme along the Yaxs. from studyng the behavour of t as a functon of or ncdence we can the arsaton, whch becomes of ng as the nterest ncreases. n the near future the DUTSCAT wll a possbl to measure two lke arsatons dfference smultaneous whch enables a for the metrc SAR A few three nsttutes n The Netherlands devel a and buld a armetrc Cband arcraft SAR of a novel des meant as a for our current SLAR system. These nsttutes are TNO n The Natonal Aerospace n Amsterdam and Mcrowave Laboratory of the Unversty of. Ths work was done under contract the Natonal Remote Sensng Board Prep. study 1987 D e f n t p H A Project plan L j prep. study prep. study study antenna antenna technology moton and SAR (PHARS) compensaton o n s t u d y R e a l R u s p H A ~ Software 14 De"'" Phu"" 2 3 Constructon of "Pharus" development,...,.,, "PASAR" res. SAR a t 0 n R u S J Pharus testphase 7 Use of Pharus 8 3: Block of the PHARUS project.

5 We are presently n the process of fund rasng, whch sofar took consderably more tme than was antcpated n the begnnng. The plans for a SAR are based on the experence that was bult up wth the prevously descrbed programs. The choce for Cband was based on the development of the ESA ERS satellte program [5]. Fgure 3 shows a block dagram of the PHARUS project (PHARUS stands for Phased Array ~nversal ~AR). The projectteam s currently watng for a GO on the frst three preparatory studes. These studes are consdered essental for a proper desgn of the PHARUS. The system wll have an actve array antenna, reason why a preparatory study on antenna technology s ncluded. Especally the problems of decouplng between the dfferent polarsatons and the ntegrated antenna desgn (ncludng power and low nose amplfers) wll be studed. The antenna moton and compensaton study s necessary to buld up the experence wth correctons of arcraft or rather antenna movements. n the thrd preparatory study a SAR testbed wll be realsed n the arcraft that wll also carry the PHARUS. The testbed s necessary to study general problems of arcraft SAR and to study the coherent ntegraton processes whch n the end determne the senstvty of the system. Fnally the testbed can be used to determne the antenna motons from the radarsgnal (va autofocus technques). The results wll be compared wth moton measurements taken from other sensors, lke gyro's and accelerometers. Ths forms an mportant nput for the fnal choce on a moton compensaton system. n table 2 some key parameters of the PHARS SAR testbed are gven. t should be consdered as a very smple SAR system wth a lmted range. There are separate patch antenna's for transmsson and recepton. These antenna's are fxed to the arcraft. The beam can be steered n coarse steps of 3.5 frequency antenna's 5.3 GHz (Cband) two 7element patch antenna's beamwdth 12 * 24 degr, HH pol. coarse step beamsteerng ( Watt peak by 7 transstors 3500 Hz transmtted power PRF pulsewdth 32 ns (4.8 m) after compresson 12.8 ~s before compresson dgtsaton 2048 samples and Q, 8 40 MHz range km azmuth presummng: 16 x arcraft Swaerngen Metro; used at an alttude of 6 km, and a speed of 100 m/s Table 2: Propertes of the SAR testbed to be realsed n the preparatory study on arcraft SAR PHARS. to compensate for the average drftangle. The beamwdth of the antenna s wde enough to elmnate the nfluence of arcraft yaw. Each element of the array antenna s equpped wth ts own transstor power amplfer of Watts. The use of dstrbuted power generaton wth transstors nstead of central TWT (travelng wave tube) power on wll yeld a much smaller than usual peak power n both the

6 PHARS testbed and the fnal PHARUS system. The senstvty of the system reles on the use of a hgh PRF (3500 Hz) and a large pulse compresson rato of 400. n the PHARUS system the peak power wll be ncreased to approxmately 2 kw. The dgtal data that remans after the azmuth presummng wth a factor of 16 wll be recorded on a hgh densty tape recordng system. Ths system s also used for the SLAR and the DUTSCAT. Apart from the azmuth presummng there s no on board processng for the testbed. By gettng down as much data as possble, very flexble experments are enabled wth the system at the cost of long processng tmes. Ths s not a major dsadvantage snce the amount of data that wll be gathered wth the testbed s small anyway. The arcraft that s used n the project s a Swaerngen Metro, a twn engne busness plane, owned by NLR and n use as a laboratory arcraft. t wll fly the PHARS and the PHARUS at an alttude of 6000 meter wth a speed of approxmately 100 m/s. The experence that s bult up n the preparatory studes wll be used n the desgn phase of the PHARUS. After the desgn s completed the system wll be realsed. At the same tme a flexble software package for SAR processng, called PASAR (!:reprocessng of 6_rborne SAR_ data) wll be developed. We hope to make use of the experence that was bult up wth the processng software for the present SLAR system. t s stll to early to gve detaled specfcatons for the PHARUS system. They wll be fxed durng the desgn phase. The plans are headng for a polarmetrc SAR wth user selectable values for resoluton (3 10 meter), swath (near range, far range or wde swath wth reduced resoluton), polarsatons. The frequency wll most probably be the same as used for the PHARS: Cband (5.3 GHz). f the plans can be carred out followng the current tme schedule, then the fnal system should be ready by References: [1] Loor. G.P. de, Hoogeboom, P. and Attema, E.P.W., The Dutch ROVE program, EEE Trans Geosc ELectron Vol GE19, 37. [2] Hoogeboom, P., Classfcaton of agrcultural crops n radar mages, EEE Trans GRS Vol GE21, [3] Hoogeboorn, P., Bnnenkade, P. and Veugen, L.M.M., An algorthm for radometrc and geometrc correcton of dgtal SLAR data, EEE Trans GRS Vol GE22, no. 6, [4] Hoogeboom, P., radar mages, Proc dentfyng SPRS Symp, agrcultural crops Enschede, 2529 n Aug, [5] Brooks, S.R., Joyce, H. et.al., The ERS1 synthetc aperture radar and scatterometer, GEC JournaL of Research, vol..3, no.2.