The scanning pencil-beam Scatterometer configuration is pretty effective in covering a large ground-swath by rotating a moderately sized paraboloid dish at a moderate speed. For example, Oscat (Oceansat-II Scatterometer) did cover a ground-swath of 1550km using a 1m diameter reflector that was rotated at 20.5 rpm. The decade-long service (1999-2009) provided by the Seawinds instrument onboard the Quikscat mission followed by an almost half-a-decade (2009-2014) service of Oscat has made this configuration tremendously popular with the global user community. A major drawback of conventional pencil-beam systems like Seawinds and Oscat is the relatively poor spatial resolution. The ground-resolution is beamwidth-limited azimuthally while, in elevation, the resolution is improved by engaging pulse-compression and range-binning. Oscat’s Instantaneous Field of View (IFOV) was 25km wide in azimuth (az) and 50km in elevation (el) at 49° incidence angle. The range-compressed resolution bins had dimensions of 6km (el) x 25km (az). Therefore, qualified wind products could be generated upon square grids no finer than 25km x 25km resolution. According to recommendations of International Ocean Vector Wind Science Team (IOVWST) and Oscat user community, high-resolution scatterometry is the requirement of the day with wind-vector cell-size dimension of 5km or better. One way to improve the resolution is to adopt the SAR principle of Range-Doppler discrimination in the scanning pencil-beam configuration. The footprint can be resolved simultaneously in range as well as in azimuth, thus significantly improving the size of the combined Range-Doppler resolution bin (~ 1km). However, the addition of Doppler filtering to conically scanning radar brings with it its own disadvantages e.g. the limitations of dwell time and the constant change in orientation of isodop lines. This paper presents the constraints in system design of high-resolution scanning systems, the design trade-offs, the methods of handling high PRF, the radar pulsing scheme and the achievable resolution.
This paper reports the development of a millimeter-wave space-borne atmospheric Temperature Sounding Unit (TSU) in Indian Space Research Organization (ISRO). This is ISRO’s first leap towards millimeter-wave technology. The sensor has several new accomplishments to its credit which include among others, the philosophy of sounding channel selection, the new assortment of temperature sounding channels, simultaneous observation of both polarizations of all channels, compact dual-band scanning Gregorian reflector antenna, indigenously developed black-body target for in-orbit calibration, in-house developed millimeter-wave RF front-end and pre-detection automatic gain control method. The prime feature of this instrument is its unique set of channels which can profile the earth’s atmosphere from surface to 40 km altitude with vertical resolution ranging from less than a km near surface to ±2.5 km at 30km altitude. The channels are predominantly off-resonant frequencies in the 50―60 GHz O<sub>2</sub> absorption spectrum which offer near-uniform attenuation and hence more channel-bandwidth and better temperature sensitivity and yet have adequate overlap of their weighting functions to achieve the desired vertical resolution. These channels are different and have fewer bands from what has been flown in all earlier sounding missions worldwide e.g. AMSU-A, SSMIS, ATMS etc. The TSU radiometer has been characterized thoroughly using ingenious methods such as low-power active RF energizing along with frequency sweep. This is a compact, low-mass, low-power instrument and has been configured for the ISRO mini-satellite (IMS-2) bus. The flight model with improved hardware performance is being built and a suitable opportunity of flying it is being explored.
NASA-ISRO Synthetic Aperture Radar (NISAR), a novel SAR concept will be utilized to image wide swath at high resolution of stripmap SAR. It will have observations in L- and S-bands to understand highly spatial and temporally complex processes such as ecosystem disturbances, ice sheet changes, and natural hazards including earthquakes, tsunamis, volcanoes, and landslides. NISAR with several advanced features such as 12 days interferometric orbit, achievement of high resolution and wide swath images through SweepSAR technology and simultaneous data acquisition in dual frequency would support a host of applications. The primary objectives of NISAR are to monitor ecosystems including monitoring changes in ecosystem structure and biomass estimation, carbon flux monitoring; mangroves and wetlands characterization; alpine forest characterization and delineation of tree-line ecotone, land surface deformation including measurement of deformation due to co-seismic and inter-seismic activities; landslides; land subsidence and volcanic deformation, cryosphere studies including measurements of dynamics of polar ice sheet, ice discharge to the ocean, Himalayan snow and glacier dynamics, deep and coastal ocean studies including retrieval of ocean parameters, mapping of coastal erosion and shore-line change; demarcation of high tide line (HTL) and low tide line (LTL) for coastal regulation zones (CRZ) mapping, geological studies including mapping of structural and lithological features; lineaments and paleo-channels; geo-morphological mapping, natural disaster response including mapping and monitoring of floods, forest fires, oil spills, earthquake damage and monitoring of extreme weather events such as cyclones. In addition to the above, NISAR would support various other applications such as enhanced crop monitoring, soil moisture estimation, urban area development, weather and hydrological forecasting.
This paper presents an overview of the estimation of noise, techniques for noise removal in Single Look Complex (SLC) as well as hybrid polarimetry decomposed images and the effects of noise removal in SAR images, using RISAT-1 data. Thermal noise affects the signal to noise ratio as well as radiometry of the SAR images. Different approaches adopted for estimating the thermal noise using onboard noise measurements and from the noise equivalent area of the images like calm reservoirs, lakes, etc., for RISAT-1 SAR are discussed. Subsequent to noise removal, its effect on minimum detectable noise and SNR of the images is addressed. Traditional noise removal methods affect the phase of the data, which in turn affect advanced SAR applications. A brief overview of the hybrid polarimetry configuration of RISAT-1 SAR, one of the emerging trends in polarimetry domain, is given and the effect of using noise removed single look complex (SLC) images for polarimetry decomposition is brought out. Thereby, a new technique for thermal noise removal in polarimetry decomposed data is presented.
Radar Imaging Satellite (RISAT-1) payload system is configured to perform self-calibration of transmit and receive paths before and after imaging sessions through a special instrument calibration technique. Instrument calibration architecture of RISAT-1 supported ground verification and validation of payload including active array antenna. During on-ground validation of 126 beams of active array antenna which needed precise calibration of boresight pointing, a unique method called "collimation coefficient error estimation" was utilized. This method of antenna calibration was supported by special hardware and software calibration architecture of RISAT-1. This paper concentrates on RISAT-1 hardware and software architecture which supports in-orbit and on-ground instrument calibration. Efforts are also put here to highlight use of special calibration scheme of RISAT-1 instrument to evaluate system response during ground verification and validation.
provide single, dual, compact and quasi-quad polarization imaging modes. Centre frequency for S-band SAR is 3200MHz with highest bandwidth of 75MHz. S-Band SAR utilizes 24 transmit receive modules (T/R Modules) to illuminate >240kms swath during transmit event and digital beam forming (DBF) on receive to reduce data rate by combining 24 receive channels and enhance SNR of the system. This paper provides details of S-band SAR system design, configuration and realization which is a challenging task since both L-band and S-band radars need to operate at same PRF and clock reference during simultaneous imaging operation. Further to this, SweepSAR technique demands PRF dithering (changing) to avoid dead gaps in the swath due to receive echo conflicting with transmit event.
Proc. SPIE. 6412, Disaster Forewarning Diagnostic Methods and Management
KEYWORDS: Digital signal processing, Doppler effect, Synthetic aperture radar, Image processing, Image resolution, Data processing, Signal processing, Radar imaging, Commercial off the shelf technology, Floods
Since last few years, ISRO has embarked upon the development of two complex Synthetic Aperture Radar (SAR) missions, viz. Spaceborne Radar Imaging Satellite (RISAT) and Airborne SAR for Disaster Mangement (DMSAR), as a capacity building measure under country's Disaster Management Support (DMS) Program, for estimating the extent of damage over large areas (~75 Km) and also assess the effectiveness of the relief measures undertaken during natural disasters such as cyclones, epidemics, earthquakes, floods and landslides, forest fires, crop diseases etc. Synthetic Aperture Radar (SAR) has an unique role to play in mapping and monitoring of large areas affected by natural disasters especially floods, owing to its unique capability to see through clouds as well as all-weather imaging capability. The generation of SAR images with quick turn around time is very essential to meet the above DMS objectives. Thus the development of SAR Processors, for these two SAR systems poses considerable challenges and design efforts. Considering the growing user demand and inevitable necessity for a full-fledged high throughput processor, to process SAR data and generate image in real or near-real time, the design and development of a generic SAR Processor has been taken up and evolved, which will meet the SAR processing requirements for both Airborne and Spaceborne SAR systems. This hardware SAR processor is being built, to the extent possible, using only Commercial-Off-The-Shelf (COTS) DSP and other hardware plug-in modules on a Compact PCI (cPCI) platform. Thus, the major thrust has been on working out Multi-processor Digital Signal Processor (DSP) architecture and algorithm development and optimization rather than hardware design and fabrication. For DMSAR, this generic SAR Processor operates as a Quick Look SAR Processor (QLP) on-board the aircraft to produce real time full swath DMSAR images and as a ground based Near-Real Time high precision full swath Processor (NRTP). It will generate full-swath (6 to 75 Kms) DMSAR images in 1m / 3m / 5m / 10m / 30m resolution SAR operating modes.
For RISAT mission, this generic Quick Look SAR Processor will be mainly used for browse product generation at NRSA-Shadnagar (SAN) ground receive station. RISAT QLP/NRTP is also proposed to provide an alternative emergency SAR product generation chain. For this, the S/C aux data appended in Onboard SAR Frame Format (x, y, z, x', y', z', roll, pitch, yaw) and predicted orbit from previous days Orbit Determination data will be used. The QLP / NRTP will produce ground range images in real / near real time. For emergency data product generation, additional Off-line tasks like geo-tagging, masking, QC etc needs to be performed on the processed image. The QLP / NRTP would generate geo-tagged images from the annotation data available from the SAR P/L data itself. Since the orbit & attitude information are taken as it is, the location accuracy will be poorer compared to the product generated using ADIF, where smoothened attitude and orbit are made available. Additional tasks like masking, output formatting and Quality checking of the data product will be carried out at Balanagar, NRSA after the image annotated data from QLP / NRTP is sent to Balanagar. The necessary interfaces to the QLP/NRTP for Emergency product generation are also being worked out.
As is widely acknowledged, QLP/NRTP for RISAT and DMSAR is an ambitious effort and the technology of future. It is expected that by the middle of next decade, the next generation SAR missions worldwide will have onboard SAR Processors of varying capabilities and generate SAR Data products and Information products onboard instead of SAR raw data. Thus, it is also envisaged that these activities related to QLP/NRTP implementation for RISAT ground segment and DMSAR will be a significant step which will directly feed into the development of onboard real time processing systems for ISRO's future space borne SAR missions. This paper describes the design requirements, configuration details and salient features, apart from highlighting the utility of these Quick Look SAR processors for RISAT and DMSAR, for generation of emergency products for Disaster management.
ScanSAR mode of SAR operation is a very important mode with which range swath can be increased multifold. This
operation is based on burst mode of operation, in which raw data is collected for a fraction of the SAR aperture time, and
different bursts are considered corresponding to different contiguous range-swaths, so that we can get an equivalent
swath width, which is much larger than what we can get with normal strip map-mode SAR imaging. This is achieved by
re-orienting the antenna beam in the elevation direction to different sub-swaths for each burst-duration. Hence, this mode
is called ScanSAR mode. Of course the penalty paid in the process, is degraded resolution.
This paper presents the ScanSAR operation philosophy and the development of its processing algorithm. The algorithm
is tested using Radarsat-1 data in the ScanSAR narrow mode. Different parameters like variable Doppler rate and
variable PRF across the sub-swaths are estimated from the data set. In the case of ScanSAR processing, estimation of
Doppler Centroid is very crucial, as burst times are very limited and hence it needs a different technique to estimate this
parameter from the small burst of data. The estimation algorithm is elaborated. Another important phenomenon called
the "scalloping effect" is observed in the case of ScanSAR images. The genesis of this effect and its compensation are
also discussed. Also due to the unique mode of imaging in bursts, mosaicing of the sub swaths in both the range as well
as the azimuth direction is required in the processed image. The overall design strategy followed for processing the
Radarsat-1 data is explained, and results are presented.
SAR Payload of RISAT (Radar Imaging Satellite), the first SAR satellite from ISRO, is currently under development.
This payload is based on active antenna technology, and it supports variety of resolution and swath requirements in C-band.
Both conventional stripmap and scanSAR modes are supported with dual polarization operation. Additionally a
quad polarization stripmap mode is provided for availing additional resource classification. In all these modes resolutions
from 3m-50 m can be achieved with swath ranging 30 km -240 km. On experimental basis, a sliding spotlight mode is
also available. The payload hardware is organized in such a way that that co and cross polarization images are available
for any operating modes. Additionally, a quad polarization mode is also supported. Active array configuration of this
payload called for development of many new technologies ranging from MMICS, TR modules, miniaturised power
supplies, high speed digitisers, dual polarized printed antenna and distributed control systems. A completely new bus is
being designed for aiding the payload operation. The RISAT spacecraft is configured around the payload to minimize the
spacecraft weight, suitable for launching by ISRO's PSLV launcher. RISAT will be placed on dawn to dusk sunsynchronous
polar orbit to ensure maximum solar power availability. All the basic building blocks have already crossed
design stage and have undergone rigorous space qualification program. Presently a complete SAR with one tile has been
integrated as design verification model and is under rigorous testing. This development ensured demonstration of end to
end hardware, on-board control software and beam control behavior.