Megha-Tropiques and Sentinel2 Expertise Centers: comparison of image quality monitoring systems

Transcription

1 SpaceOps Conferences May 2016, Daejeon, Korea SpaceOps 2016 Conference / Megha-Tropiques and Sentinel2 Expertise Centers: comparison of image quality monitoring systems Jean-Louis RAYNAUD 1, Michel DEJUS 2, Thierry TREMAS 3, Julien NOSAVAN 4, Beatrice PETRUCCI 5 CNES, 18, Avenue E. Belin, Toulouse Cedex 9 and Anne-Sophie LACAMP 6 Thales Services,1-2 Avenue de l Europe, Ramonville-Saint-Agne With SPOT satellites, CNES has acquired a strong experience in tuning and monitoring Earth observation imaging systems. In this context, two satellites, Megha-Tropiques and Sentinel2A, are currently monitored through dedicated technical expertise centers. This paper will compare these two systems on several aspects of their monitoring, rising differences but, overall, highlighting the similarities between them. Nomenclature CNES = Centre National d Etudes Spatiales DRM = Data Resource Manager GIPP = Ground Image Processing Parameters IODD = Input Output Data Definition IQ = Image Quality ISRO = Indian Space Research Organization ISSDC = Indian Space Sciences Data Center MADRAS = Microwave Analysis and Detection of Rain and Atmospheric Systems MMI = Man Machine Interface MSI = Multi-Spectral Instrument MT = Megha-Tropiques NRT = Near Real Time S2A = Sentinel2A SAG = Structure d Accueil Générique SAPHIR = Sondeur Atmosphérique du Profil d'humidité Intertropicale par Radiométrie SCAPIN = ScaRaB Performances INvestigator ScaRab = Scanner for Radiation Budget SHERPA = Simulator saphir End-to-end of Radiometer for Performance Analysis TECMT = Technical Expertise Center for Megha-Tropiques TECS2 = Technical Expertise Center for Sentinel2 TOA = Top Of Atmosphere 1 Image quality engineer, DCT/ME/EI, jean-louis.raynaud@cnes.fr 2 MT head project, DCT/ME/OT, michel.dejus@cnes.fr 3 S2 head project, DCT/SI/MO, thierry.tremas@cnes.fr 4 Ground segment engineer, DCT/PS/OT, julien.nosavan@cnes.fr 5 Ground segment engineer, DCT/PS/OT, beatrice.petrucci@cnes.fr 6 Engineer, CCL LOM Aerospace, anne-sophie.lacamp@thalesgroup.com 1 Copyright 2016 by CNES. Published by the, Inc., with permission.

2 I. Introduction n CNES premises, since the beginning of the SPOT (Système Pour l Observation de la Terre Earth Observing I System) satellites operation, image quality monitoring systems have been conceived, developed and operated. In the Earth observing context, the acquired experience has leaded to a current management of a great variety of technical expertise centers. The scope of this paper is to compare two image quality technical centers that aim at monitoring and tuning the in-flight payloads of Megha-Tropiques and Sentinel2A missions. After a description of the main differences between these two missions, strong common points in the image quality monitoring systems will be illustrated and developed. A generalization will afterwards be considered through an overview of the monitoring activities of other earth observation missions. Finally, some Megha-Tropiques instrument performances will be reviewed to illustrate the abilities of an image quality system. II. Opposite missions at first sight A. Mission and Orbit MT and S2 have radically different orbit and mission purposes 1,4,5,6,10. On one hand, MT can be classified as a scientific mission, oriented to climate study and atmosphere observation. This satellite is the result of a joint collaboration between India (ISRO) and France (CNES), with a specific distribution of responsibilities. On one hand, the satellite platform and all the activities related to the orbit maintenance and the products delivery are on ISRO side. On the Figure 1. MT orbit IXION(LMD) other hand, after having developed the instruments, CNES has the responsibility of their expertise and monitoring. The goal of MT satellite is mainly to study water cycle and energetic exchanges in the inter-tropical convergence zone. The main instrument on board, SAPHIR, is dedicated to observe the atmosphere and to measure water vapor distribution. Indeed, the land is scarcely visible in only one channel of SAPHIR. ScaRab completes the global observation by retrieving radiative fluxes. Thanks to a very original circular orbit (Fig. 1), the repetitivity is excellent in many areas. With an inclination of 20, MT can observe up to 6 times a day over the latitudes between 10 and 20, and up to 4 times at other latitudes. This characteristic, associated with the fact that MT acquires in continuous mode, is crucial to follow life cycle of the mesoscale convective systems. This is also a good opportunity to deliver microwave observation in NRT (Near Real Time) to world weather agencies for assimilation in weather predictive models. On the other hand, S2 is a land observation satellite, with a strong operational purpose. It is a part of Copernicus European program, an initiative from European Commission (EC) and European Space Agency (ESA) which aim is to provide various means to gather and use observation data required to monitor land environment and security (management of the environment, effects of climate change). S2A is a part of the space component of Copernicus, and is dedicated specifically to an automatic full coverage of land surface between 83 North to 56 South, on a Figure 2. S2 orbit IXION(LMD) polar sun-synchronous orbit (Fig. 2). All the produced data will constitute a continuity of SPOT and Landsat image data acquired during the past decades. With one satellite, the revisit of observation of one satellite is 10 days (slightly better for points on earth in overlapping areas). When Sentinel-2B will be operational (scheduled in 2016, November, on the same orbit, but phased at 180 ) the global repetitivity of S2 mission will be 5 days. 2

3 Table 1 gives an overview of mission and orbit main differences between MT and S2 satellites. Table 1. MT and S2 mission/orbit differences MT S2A Mission Atmosphere Land observation Orbit Equatorial Polar Inclination 20 98,62 Repetitivity 4 to 6 /day 1 to 2 / 10 days Acquisitions Continuous Only land B. Satellite and sensors Megha-Tropiques and Sentinel2A can be considered as two very different satellites. Even if the global weights of these satellites are similar (1 ton for MT and 1,2 tons for S2A), the on board instruments are very different. Their specificities, radically distinct, constitute another gap between the image quality systems under analysis here. They are described hereafter. On board of MT satellite, the mission core payload is composed of three different instruments 1 : - MADRAS (Microwave Analysis and Detection of Rain and Atmospheric Systems) is a 9 channels self-calibrating microwave imager with conical scanning. Developed jointly by ISRO and CNES, MADRAS has operating frequencies in the range 18.7GHz-157GHz for studying ice and cloud liquid water and precipitation. This instrument is no more operational since the end of January Due to this, it will not be taken into account in the frame of this paper which is mostly focused on the current comparison between MT and S2A. - SAPHIR (Sondeur Atmosphérique du Profil d'humidité Intertropicale par Radiométrie / Sounder for Probing Vertical Profiles of Humidity) is a multi-channel passive microwave humidity sounder. This CNES instrument measures brightness temperatures, which can be used afterwards to retrieve water vapor profiles in the troposphere. This instrument has 6 channels near the absorption band of the water vapor at 183 Ghz for this purpose. The spatial resolution of SAPHIR pixels ranges from 10km (nadir) to 22km (50 incidence), with a total swath of around 1700km. MADRAS being declared non-operational, SAPHIR has become the main mission instrument on board. Figure 3. Megha-Tropiques and Sentinel2A. artist views CNES,ESA Figure 4. MARFEQ/MADRAS CNES Figure 5. SAPHIR CNES 3

4 - ScaRaB (Scanner for Radiation Budget) is an optical scanning radiometer devoted to the measurement of radiative fluxes and earth radiative budget at the top-of-atmosphere (TOA) in the shortwave and longwave domain 3. In terms of wavelengths, the observed domain ranges from 0.2 to 100µm. The ScaRaB field of view (FOV) corresponds to a geographical footprint of approximately 40 km squared at nadir to 200 km on the edge, for a total swath of 2200km. Figure 6. ScaRab CNES Figure 7. S2A and MSI artist view ESA Figure 8 shows an overview of the spectral distribution of the different sensors channels previously described. Even if S2/MSI and MT/ScaRab global bandwidth have a very little part in common, ScaRab, SAPHIR and MSI are clearly separated in the frame of wavelengths. Figure 9. MT and S2 examples projection on Earth of S2/MSI product and false-color MT/Saphir acquisition Google Sentinel2A, on its side, has only one push-broom instrument named MSI (Multi Spectral Instrument). It acquires 2 images in 13 spectral bands, ranging from 490nm to 2190nm, for a total swath of 290km and a resolution depending on the spectral band (10m, 20m and 60m). It is the first satellite of a constellation. Figure 8. SAPHIR, ScaRab and MSI spectrum comparison Finally, Fig.9 and Table 2 illustrate the main differences of the sensors characteristics on board MT and S2A. Figure 9 presents an earth projection of an example of S2A/MSI acquisition and a false-color MT/SAPHIR product to show the gap existing between these two sensors. The chart in Table 2 completes this illustration by comparing some explicit figures to highlight how opposite are these two systems. Table 2. sensors differences synthesis MT (SAP / SCA) Swath 1700km / 2200km 290km Resolution 10-22km / km m Spectral Bands 183Ghz(1640µm) / 0,2-100µm S2 0,4-2,2µm 4

5 C. Data processing As seen previously, MT and S2 are two very distinct missions. In CNES premises, there are two specific technical expertise centers dedicated to their image quality monitoring: TECMT and TECS2. An expertise center (which is also an operation tool) can be described as a pool of human resources and technical tools, specifically conceived to monitor specific payloads. Various activities can therefore be triggered: monitoring instrument status, measuring performances, tuning and delivering parameters of both on board and on ground systems, analyzing anomalies or users feedbacks All these tasks are carried on through the use of dedicated tools integrated in these technical expertise centers, and on the basis of on board and on ground data processing organized around a central database. In what concerns MT and S2 missions, CNES responsibility is not the same for these two systems. As a consequence, their respective image quality monitoring are also quite different. This explains the big gap existing between the data processed in these expertise centers. For MT payloads, CNES has two main responsibilities: health instruments monitoring and both radiometric and geometric image quality checking. These activities are based on the processing of all telemetry raw acquisitions, and also through the analysis of all the Level-1 brightness temperature products delivered by ISRO. This implies that TECMT can be described as a pool of tools which systematically process all the data produced by ISRO (more than 3000 products downloaded and monitored by month). From the point of view of the data volumetry, MT products are fortunately not very big (1 month of monitoring products represents 50Go of data). Nevertheless, the regular and automatic tasks that should be carried on every day imply a strict monitoring of the whole expertise center. For S2/MSI, there is no instrument health monitoring directly at CNES. The key role for CNES has been to accurately tune on board and on ground parameters (during In Orbit Commissioning phase) to ensure the best image quality on Level-1 products. Currently, CNES and ESA/ESRIN collaborate closely to monitor MSI image quality. For that purpose, a sampling of products (on specific and well-known geographic sites on Earth) is processed on TECS2 center. Unlike MT products, S2A image products represent a huge amount of data. Indeed, the 13 spectral bands, associated with more than pixels wide for 10 m resolution bands, generate more than 500 Go of data for 1 month of expertise support! Thereby, by comparing at this time the data used in a context of image quality monitoring, MT and S2 have also important differences. On one hand, TECMT needs a lot of products occupying tiny disk space, on the other hand, few products representing huge amount of disk space are processed in TECS2 center. Chart in Table 3 gives an overview of these practical differences. Table 3. MT and S2 data processing differences Type of Monitoring Acquisitions frequency Nb products/month Disk space /month MT Instrument supervision & Image quality Systematic S2A Image quality Samples >3000 < Go >500Go III. Similar image quality monitoring As seen in the previous paragraphs, MT and S2 are really different missions, associated with distinct types of image processing and monitoring. Moreover, data (in terms of global coverage, frequency of downloading, disks space occupation ) processed in each image quality center, are definitely opposite. Nevertheless, it can be shown that, in the end, many similarities exist between the corresponding image quality monitoring systems. Even more, a special focus will be made on how these common points have been taken into account to create a unique software environment for the expertise centers. 5

6 A. Operational communication As far as organization is concerned, these systems deal with similar interfaces. First of all, let s recall that MT and S2 are both the result of international cooperation: with ISRO for MT and ESA for S2. In these similar contexts, communication and interfaces are key aspects for the success of image quality monitoring, as shown in this paragraph. 1. Image loop The organization, in technical expertise systems, is centered on what is called the operational image loop. For TECMT and TECS2 centers, such image loop is based on communication between entities. The programming centers, the ground segments, the radiometric and geometric experts, the operation team and the final users are key interfaces in the global image IQ monitoring. This organization, strongly based on human resources, is detailed hereafter to highlight many common points between MT and S2 For Megha-Tropiques image quality monitoring, the raw materials are, in equal shares, Level 0 products (telemetry) and Level-1 products. Thus, the image loop (Fig. 10) begins by an automatic download of all these products generated by the Indian ground segment (located at ISSDC, Bangalore, India). These data are processed afterwards in TECMT center. In this context, 474 different parameters are monitored on SAPHIR and ScaRab Level 0 and Level-1 products to check health instrument and image quality. All these activities, even automatically triggered, need human resources to check the operational processing, and also to analyze the results and the Figure 10. MT operational communication loop possible anomalies raised. Communication and exchanges between TECMT and Indian ground segment are also a crucial task to ensure all measurement exhaustiveness. Throughout all these tasks, monitored by an operation team, monthly syntheses are created to give an overview of the instruments behavior, and the products quality. These reports are used to finalize the image loop : the possible delivery of new image producing parameters. For Megha-Tropiques products, on ground parameters dedicated to produce optimal Level-1 products from Level 0 products are gathered in so-called IODD (Input Output Data Definition). These IODD can evolve during the instrument life, and TECMT is the main actor in this evolution. Through accurate analysis of monthly reports, experts can decide to update IODD, in collaboration with ISRO expert team. In parallel of this image loop (recuperation of products, analyzing and delivering of parameters), other important communication exist in the frame of TECMT. To ensure optimal instrument monitoring, some specific on board operation can be asked to the programming center. For example, for ScaRab instrument, monthly specific calibrations are programmed on board. Other specific calibration can be asked to the Indian programming center, if needed. Last but not least, what would be an image quality monitoring without the final users feedbacks! Indeed, in Megha-Tropiques, the opinions delivered by the final users to CNES and ISRO agencies are extremely valuable. In the frame of the expertise center, this information is used to correct products problems and increase global image quality level. Moreover, even if Megha-Tropiques can be qualified as a scientific mission, the operational aspect of SAPHIR data must be enhanced. Indeed, weather agencies assimilate SAPHIR data to improve the quality of their numerical weather predictions. In this frame, NRT (Near Real Time) SAPHIR products are delivered by ISSDC to the international community in less than 3 hours (after the end of each product acquisition). This operational use of data makes even more significant the users observations. 6

7 Now, as far as Sentinel2A image loop is concerned (Fig. 11), many similarities can be highlighted with respect to Megha-Tropiques. Indeed, the basis of S2A image quality monitoring and tuning are the Level-0 products generated by the ground segment (located in ESRIN, Frascati, Italy). Currently, the Level-1 directly generated by the ground segment are not used in TECS2, but are soon foreseen to be considered as a basis of image quality monitoring. Once Level-0 downloaded, they are used to generate Level-1 products through TECS2. Figure 11. S2 operational communication loop Afterwards, these products are regularly processed in the expertise center through specific radiometric and geometric tasks. The goal, as TECMT center, is to check if some image quality parameters should be updated or not, in order to have radiometric and geometric performances respecting the system specifications. The regularity of the S2A images processing is extremely important to be able to follow potentially fast evolutions and to insure one the most important mission objective: generating high quality time series. As an example, the absolute calibration is regularly measured to check a possible radiometric transmission decrease. As a consequence, the equivalents of IODD (for MT) are regularly generated in TECS2. These so-called GIPP (Ground Image Processing Parameters) ensure an optimal image quality for S2A products. Another common point between S2 and MT is the CNES ability to ask for some special programming requests. Indeed, after ESA acceptance, special operations can be decided to optimize the final products image quality. Decontamination or diffuser acquisition are examples of such requests. Finally, feedbacks of users are also extremely important in the communication loop. Unlike MT, the anomalies found by users are sent directly to ESA. Nevertheless, they may afterwards be accurately analyzed by ESA and CNES jointly. In the end, the comparison of operational loop in both MT and S2A image quality monitoring systems shows they have many common points. The data fine analyses (Level-0 and Level-1) delivered by operational ground segment, associated with the final users feedbacks, are used to maintain an optimal image quality through the update of specific parameters. All these activities are managed in a strong context of system specifications and operational constraints. 2. Expertise and Operation As seen in the description of the MT and S2A operational image loop, many similar entities are involved in their image quality centers. A focus is hereafter made on a specific part of these organizations: experts and operations team. In CNES way of conceiving and organizing an image quality center, experts and operation team collaborate very closely. This collaboration is illustrated in Fig. 10 and Fig. 11, and has many common points in MT and S2 contexts. First of all, in terms of communication, regular meetings between experts and operators are scheduled to achieve all needed image quality activities. For S2A, during commissioning phase and during current routine activities, many meetings involving experts and operators have been organized to plan specific productions and to share radiometric and geometric image analysis results. On MT side, even after more than four years in orbit, routine meetings allow following the Megha-Tropiques instruments evolution and products quality. Secondly, in what concerns image quality activities, a lot of links can be identified between experts and operation team. Indeed, the global monitoring of S2A and MT image quality centers are managed by the same operation team (whereas experts are different on these two systems). On the basis of procedures described by experts, operators produce many intermediate data and first level analysis. Based on these results, experts can easily finalize image quality analysis to possibly deliver new accurate parameters to on ground or on board systems. This organization clearly enhances a virtuous circle: operation team achieves interesting and motivating activities and learns on satellite image analysis; experts can rely on operation team to process routine activities and can dedicate themselves mostly to future sensors. 7

8 Basically, proximity and interactions between experts and operators make the success of CNES image quality processing. The close collaboration between these teams generates an operational experience in earth observation systems, which the next generation of expertise center takes advantage of. B. Operational Monitoring System Following previous parallelisms between MT and S2, a focus can now be made on the software environment of the expertise centers. Indeed, the image quality of MT and S2 payloads is monitored via almost the same framework named SAG 7 (Structure d Accueil Générique generic host structure). This generic and reusable software structure has been developed by CNES with the feedbacks of some previous image quality monitoring systems. In this way, SPOT (Système Pour l Observation de la Terre Earth Observing System), VGT (VeGeTation Instrument on board SPOT satellite), IASI (Interféromètre Atmosphérique de Sondage dans l Infra-rouge Infra-red atmospherical sounding interferometer) expertise centers (among others) have been considered as models for the conception of SAG framework. Indeed, they have been initially developed by CNES to fit with the specificities of each image quality monitoring. But, at the same time, their structure contain common tasks which have been extracted and analyzed to develop the SAG schema. The result is that MT and S2 8 expertise/exploitation centers, both built around SAG, have a similar software environment. Figure 12 and Fig. 13 introduce the graphic user interface TECMT and TECS2 expertise centers. These pictures clearly illustrate the strong visual likeness between these two systems. Figure 12. MT technical expertise center Figure 13. S2 technical expertise center Each system using a SAG-type software environment can be divided into specific parts: - a centralized database structure (using POSTGRES), built around a data modeling specific to each payload (SAPHIR and ScaRab for MT, MSI for S2) - various MMI (JAVA language) allowing users to connect themselves to the system (identification LDAP - Lightweight Directory Access Protocol - process) and to interact with it - a Data Resource Manager (DRM) to allocate processing on remote computing servers in an optimized and controlled way - an overlay of processing code (expertise or operation tools) specific to each mission monitoring Thanks to such a system, the users can carry on every activity necessary to image quality monitoring (from the phase of in-flight calibration/validation to the routine operations): - to search data in the database through complex requests - to visualize image products or auxiliary data with specific external software (that are integrated to the center) - to process data to calculate performances and to update correction parameters - to integrate personalized process to monitor the system or to add new topics to global monitoring - to follow the progress of all these activities - 8

9 Moreover, MT and S2 operational architectures are connected to CNES generic and transversal services: backup and restoration service, antivirus support, long term archiving facility and system supervision. The consequences of the reuse of this common framework between TECMT and TECS2 are mostly very convenient: - many common points in operation procedures of the systems, enabling people to switch from one center to another (pooling human resources) - ease of global system maintenance and problems identification thanks to similar architecture in different centers The only drawback in the use of such a framework is the need of yearly basis update of SAG system (for maintenance reasons) on every technical center based on SAG. This consideration, associated with inherent genericity arbitrages and limitations, are nevertheless minor with respect to all the advantages provided by such a system. C. Level-1 Production Speaking about geophysical first level data production ( Level-1 ), CNES has played a similar role on both Megha-Tropiques and Sentinel2. Indeed, it was responsible of both algorithms specifications and Level-1 products generator prototypes. On Megha-Tropiques side, two main types of Level-1 products can be considered: - Level-1A are brightness temperature (SAPHIR) or radiance (ScaRab) products constituted by the samples acquired on board in the instrument geometry, with geo-location, time of acquisition, and other relevant parameters and flags - Level-1A2 are also brightness temperature (SAPHIR) or radiance (ScaRab) products, with pixels defined as non-overlapping synthetic footprints covering the scan swath For Sentinel2A, the different Level-1 productions 8,9 processed in CNES premises are slightly different (only ScaRab and MSI share common points, like for example the final radiance measurement): - Level-1A products correspond to uncompressed data with few radiometric processing (defective or saturated pixel masking ) - Level-1B products have more radiometric corrections (defective pixel interpolation, dark signal and pixels response non uniformity correction ). Moreover, an enhanced physical geometric model is estimated - Level-1C products provide ortho-rectified Top Of Atmosphere (TOA) reflectance with a sub-pixel multispectral and multi-date registration, associated with specific masks. Even if the Level-1 products are different for S2 and MT, their origin is eventually the same. Indeed, the specificities (algorithms, specifications) of the whole products generation processing have been elaborated by CNES. Furthermore, CNES has actively participated to the industrial development of MT and S2 Level-1 products generator prototypes. For Megha-Tropiques, two different tools have been conceived to deal with Level-1, both based on CNES specifications. - SCAPIN (ScaRab Performances Investigator illustrated in Fig. 14) has been developed by Magellium. Its goal is to produce Level-1 products for ScaRab instrument, but also to easily investigate on all performances parameters available for this payload. - SHERPA (Simulator saphir End-to-end of Radiometer for Performance Analysis) has been implemented by CLS and is used to generate Level-1 SAPHIR validation products in the CNES premises. These tools are prototypes, the official Level-1 software being integrated in ISSDC ground segment. Nevertheless, they are regularly and operationally used to generate Level-1 products samples, and to compare themselves to the equivalent products generated by ISSDC. Thanks to these comparisons, numerous since MT launch, some anomalies have been raised on the operational ISSDC products. After accurate analyses, jointly driven by CNES and ISRO, the usage of SCAPIN and SHERPA has eventually leads to the global amelioration of operational Level-1 products. Figure 14. MT : ScaRab Level-1 prototype 9

10 On Sentinel2A side, during In Orbit commissioning phase, the generator of the Level-1 products has been the GPP (Ground Prototype Processor, MMI in Fig. 15). GPP 8 has been developed by an industrial consortium composed of Advanced Computer Systems (ACS), Magellium and DLR, under the joint management of ESA-ESTEC and CNES. During commissioning phases, the operational generator of Level-1 is the GPP, while the official ESA ground segment is on its validation period. Figure 15. S2 : GPP Level-1 prototype IV. Generalization Finally, however different the two missions MT and S2 may be, many similarities can be detected in their respective image quality monitoring systems. The global expertise centers overview which comes out of MT-S2 comparison can afford to build a generalization towards other systems monitored by CNES. Indeed, a pool of expertise centers is currently in operation in the same CNES department, most of them for a long duration. The chart Fig. 16 gives an overview of how these systems can be distinguished. Figure 16 Overview of current CNES pool of image quality systems In this chart, each circle corresponds to an image quality expertise center. This chart can be analyzed taking into consideration that - the size of each circle can be correlated with the global size of the monitoring center (taking into account various parameters : number of lines of code, activity around anomalies and evolutions on the system ) - the color of each system is associated with the type of monitoring (image and/or instrument) - the reuse of SAG framework is enhanced for each center (red circle) - the position of each expertise center (top or bottom) corresponds to the level of responsibility in IQ monitoring (only during commissioning phase at bottom, during the whole life of satellite at top) 10

11 Clearly, the reuse of the same framework SAG is the rule since Pléiades satellite monitoring system (PHR - Pléiades High Resolution). The goal is, of course, to mutualize the experience acquired with older payload monitoring systems. In this frame, some examples can be mentioned: IASI, CALIPSO (Cloud Aerosol Lidar and Infrared Pathfinder Satellite Observations) or H (Hélios). All the activities related to these expertise centers development and exploitation have been gathered to help the conception and integration of new systems. Today, thanks to the Pléiades, Megha-Tropiques and Sentinel2A image quality expertise centers, all based on the SAG framework, CNES has the appropriate tools to easily conceive and set up any earth observation expertise system. The global organization from development to exploitation is extremely well-established on CNES side, from a technical and a human point of view. Indeed, it is important to notice that the SAG framework is flexible enough to allow various profiles of actors to work together simultaneously. Thus, during commissioning phase (intense period of satellite payload validation) experts and operators, together with industrial maintenance, all access and work around the expertise center. From this graphic, it can also be noticed that S2 represents a new type of collaboration in the frame of image quality. CNES is responsible during commissioning phase, but another agency (ESA) drives the image quality activities during routine period. Finally, this overview also shows a relative homogeneous distribution of the 2 main types of monitoring systems between the different sensors: - instrument and image quality monitoring for IASI, CALIPSO and Megha-Tropiques - only image quality monitoring for Hélios, Pléiades and Sentinel2A. Beyond this distribution, other expertise centers are currently in preparation for future Earth observation systems. VIQ (Venµs Image Quality) monitoring system is, for example, currently in a qualification phase to be operational for Venµs launch (foreseen in 2017). V. Conclusion Megha-Tropiques and Sentinel2A are certainly distinct missions. Many differences can be found in terms of orbit, image characteristics, types of monitoring or data processing. Nevertheless, these systems can be put into perspective and significant similarities are eventually highlighted in their respective image quality monitoring centers. The operational communication loop, the teams working around these projects, the role of CNES in Level-1 generation, and, above all, the common software environment articulated around the same framework, make eventually these systems very comparable. The result of this comparison demonstrates that, although opposite some imaging systems can be, it is possible to conceive a global and generic monitoring system. In terms of human organization or software environment, a generalized model of expertise center can allow maintaining an optimal image quality in the final users products. Thanks to almost 30 years of experience (since SPOT satellites), CNES has the appropriate tools to easily set up such complete expertise centers. Beyond these means, the experienced experts and operation teams make the CNES a reference in image quality monitoring domain. 11

12 Appendix Finally, in order to practically illustrate examples produced by an image quality monitoring system, hereafter are reviewed some Megha-Tropiques instrument performances or measurement charts. a b Figure 17. exemples of typical charts produced by TECMT a Tendency since launch of SAPHIR sensibility of each channel b ScaRab temperature measurement of specific equipment on a specific orbit 12

13 References 1 Nadia Karouche,Garudachar Raju, Megha-Tropiques Satellite Mission : Sensors performances, Proc. SPIE 7826, Sensors, Systems, and Next-Generation Satellites XIV, 78260Q (October 13, 2010); doi: / Thierry Trémas, Cécile Déchoz, Sophie Lacherade, Julien Nosavan, Beatrice Petrucci, Sentinel 2: Presentation of the CAL/VAL commissioning phase,proc. SPIE 9643, Image and Signal Processing for Remote Sensing XXI, (October 23, 2015); doi: / A. Rosak, T. Tremas, N. Karouche, L. Gillot, O. Simonella, 2012, Radiometric and geometric Scarab-Megha-tropiques ground calibration comparison with first in-orbit calibration, IGARSS 2012, Munich Germany 4 Michel Capderou, MTTM, Megha-Tropiques Technical Memorandum, Sampling Comparison with other Meteorological Satellites, March Michel Capderou, The MT Orbit and its Angular Sampling, Proceedings of ERB (Earth Radiation Budget), Workshop 2010, Paris France, Sept , Michel Capderou, Satellites : de Kepler au GPS, Springer, Peña-Luque S, Nosavan J., Selle A., SAG Data hosting structure for generic purposes in space operations AIAA SpaceOps Conference, paper , doi: / J. Nosavan, B. Petrucci, J-L. Raynaud, T. Tremas, CNES Cal/Val expertise Centre for Sentinel-2 in orbit tests (TEC-S2): architecture and data processing, Proc. SPIE 9639, Sensors, Systems, and Next-Generation Satellites XIX, (October 12, 2015); doi: / B. Petrucci, C. Dechoz, S. Lacherade, C. L Helguen, J-L Raynaud, T. Trémas, C. Picard, A. Rolland, P.Martimort, C.Isola, F.Spoto, The ground prototype processor: Level-1 production during Sentinel-2 in-orbit acceptance, Proc. SPIE 9643, Image and Signal Processing for Remote Sensing XXI, (October 15, 2015); doi: / Rémy Roca, Hélène Brogniez, Philippe Chambon, Olivier Chomette, Sophie Cloché, Marielle E.Gosset, Jean-Francois Mahfouf, Patrick Raberanto and Nicolas Viltard The Megha-Tropiques mission: a review after three years in orbit Original research 18 May 2015, doi: /feart