News on Image Acquisition for the CwRS Campaign new sensors and changes

Transcription

1 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 1 News on Image Acquisition for the CwRS Campaign new sensors and changes Pär Johan Åstrand, Joanna Nowak, Maria Erlandsson

2 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 2 Outline General issues VHR, HR news RapidEye, GeoEye1 geometry tests Geoeye-1, Ljubljana. Telespazio GeoEye-1, 0.41-meter PAN, meter resolution MSP launched 6/9/2008

3 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 3 General Updated VHR/HR Specs Distribution rights IK, QB, GE1 changed new FC for GE1 feasibility iteration different from earlier years FC up for renewal for Campaign 2010 CID portal / LIODOTNET orthocorrected HR imagery upon request (semiautomated); SPOT series, Landsat, IRS image return / FTP harvest of data register to CID portal at:

4 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 4 Updated VHR/HR Specs draft available in LIODOTNET, and on our webpage pls. comment on them before 1/5/2009 Issues: introduction of GeoEye1, Rapid Eye (typical zone, formats etc.) slight change in EROS B acquisition modality SPOT products (2.5m COL as backup upon request) MSP component in PAN backup only upon request (2009) administrative (e.g. IDQA, and invoicing) tables/references updated

5 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 5 VHR GeoEye-1 GeoEye (USA) Launched 6 Sept Km orbit Pan + 4 MS 15.2 Km swath 0.5 m GSD nadir) 1.65 m GSD MSP CwRS Campaign programming approx km2 should be 1.5x more efficient than IK

6 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 6 VHR Kompsat-2 KARI (South Korea) Launched 28 July Km orbit Pan + 4 MS 15 Km swath 1 m GSD CwRS Campaign programming Additional to Ikonos, imagery on 2 zones to be tested during 2009 Campaign ITIN SE, ANOI ES pending SPOTImage/KARI

7 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 7 VHR EROS B Launched 25 April Km orbit Pan 7 Km swath 0.7 m GSD CwRS Campaign programming approx km2 start imaging half a second earlier and finish to image half a second later in order to allow correct telemetry in time for start/end of scene EROS B continued tests: new acquisition over LUED, DALD, DE optimizing no. of GCPs tests of block triangulation to reduce ground truth

8 Control Methods RapidEye EROS-A Workshop: 6 -Pansharpened 8 / 5m 4 / 2009 GSD [CwRS KO Meeting Campaign 2009] 8 10% 100%

9 Control Methods RapidEye EROS-B Workshop: 6 -Pansharpened 8 / 5m 4 / 2009 GSD [CwRS KO Meeting Campaign 2009] 9 100% 5%

10 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 10 HR RapidEye 5 small satellites built by SSTL (UK) RapidEye AG (Germany) Launched 29 August m nadir 5 MS bands 77 Km swath daily revisit on the whole world SPOT 4, 5 SWIR band ( µm) CwRS Campaign programming to be introduced in HR Campaign 8 MS out of 20 accept to use RapidEye: BG, CZ, DE, ES, IE, IT, NL, UK. 138 zones final meeting 9/4/2009 with RE

11 RapidEye over Mausanne: comparison of different band combinations, 11 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 1 including the newly introduced red edge R: near infrared G: red edge B: green 1 R: near infrared G: red B: green Spectral region at the limit of the red and near-infrared wavelengths characterized by a sharp rise in the plant reflectance.

12 RapidEye over Mausanne: comparison of different band combinations, 12 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 1 including the newly introduced red edge R: red edge G: red B: green 1 R: near infrared G: red B: green Spectral region at the limit of the red and near-infrared wavelengths characterized by a sharp rise in the plant reflectance.

13 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 13 Benchmarking the GeoEye-1 and RapidEye sensors for use in the CAP CwRS programme Joanna Nowak, Pär Åstrand Pedro MiguelSanz Muñoz JRC, CID Tragsatec, Spain

14 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 14 CAP CwRS Geometric Accuracy Requirements - The EU standard for the orthoimagery to be used for the purpose of the Common Agriculture Policy (CAP) Control with Remote Sensing (CwRS) requires the assessment of the final orthoimage. 1-D RMSE of < 2.5m (prime sensor requirement) - The RMS error calculated for Independent Control Points (ICPs) in each dimension (either Easting or Northing) should not exceed 2.5m. - The RMS results from the image geometric correction phase (sometimes referred as to internal quality control) are also useful because they are part of the orthoimage quality assurance.

15 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 15 GeoEye-1 Test Data and Software Two Geo (2A Standard Ortho-Ready) level images: geometrically processed and mapped to a cartographic projection using spacecraft parameters and constant DEM (average height measured on the GTOPO30, 30 arcseconds DEM) Image ID acquisition date Nominal collection azimuth Nominal collection elevation approx. off-nadir angle Scanning directio n site name, country code po_ _pan_ tif :39 GMT 182 deg 59 deg 29 deg East to West (i.e. reverse) Mausanne, FR po_ _pan_ tif :39 GMT 57 deg 81 deg 8 deg East to West (i.e. reverse) Mausanne, FR Auxiliary data: GCP/CP/ICP of RMSEx,y=0.05m, RMSEz=0.1m (DGPS measurements); DEM generated from ADS40 (Leica Geosystems) digital airborne image with 2m resolution and RMSEz=0.6m The geo-imaging software packages used : PCI Geomatica OrthoEngine (RPC IKONOS model, degree 1 polynomial) ERDAS LPS 9.2. (RPC IKONOS model, degree 1 polynomial)

16 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 16 GeoEye-1 Geometric Image Quality Analysis (1) RMSE [m] of ICPs as a function of the number of the GCPs (0, 1, 3, 4, 5, 6, 9, 11 respectively) RMSE_ICP_X Tragsatec, PCI RMSE_ICP_X JRC CID, ERDAS LPS RMSE_ICP_Y Tragsatec, PCI RMSE_ICP_Y JRC CID, ERDAS LPS V_0 V_1 V_3 V_4 V_5 V_6 V_9 V_11 RMSE based on the measurements of the ICPs on the final orthoimages The number and distribution of the ICP is constant (13) The figure illustrates the worst case of all the tested project configurations

17 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 17 GeoEye-1 Geometric Image Quality Analysis (2) RMSE [m] of ICPs as a function of the number of the GCPs (3, 4, 5, 6, 9, 11 respectively) RMSE_ICP_X Tragsatec, PCI RMSE_ICP_X JRC CID, ERDAS LPS RMSE_ICP_Y Tragsatec, PCI RMSE_ICP_Y JRC CID, ERDAS LPS V_3 V_4 V_5 V_6 V_9 V_11 RMSE based on the measurements of the ICPs on the final orthoimages The number and distribution of the ICP is constant (13) The figure illustrates the worst case of all the tested project configurations

18 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 18 Conclusions Based on the limited GeoEye-1 sample images that were made available to us: - the orthoimage accuracy of the Geo (2A Standard Ortho-Ready) level imagery FULFILL our geometric accuracy requirements (1-D RMSE of < 2.5m) provided a minimum of 3* GCPs (evenly distributed). (*) it is recommended that GCPs cover full AOI

19 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 19 RapidEye Test Data and Software Level 3A product: orthorectifed, 25x25 km scene size, 6bit GeoTiff format (UTM, WGS 84), 5m resolution (resampled from the original 6.5m); processing options: yes/no atmospheric correction Image ID Acquisition date Azimuth Angle Across Track Incidence Angle Orbit Direction Test Site Name T111140_RE5_3A- NAC_ : deg 9.6 deg (East) descending Maussane, FR T111140_RE5_3A- NAC_ : deg 7.5 deg (East) descending Maussane, FR T111906_RE3_3A- NAC_ : deg 9.9 deg (East) descending Maussane, FR T111905_RE3_3A- NAC_ : deg 2.7 deg (East) descending Maussane, FR RapidEye's automated ortho-rectification auxiliary data: DEM: proprietary SRTM-3 (90m) dataset, and GTOPO30 for areas above 60 degrees North; GCPs: derived from the Global Landsat Mosaic (2-7 GCPs per scene); Off-the-shelf geo-imaging software do not support the RapidEye image orthorectification yet.

20 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 20 RapidEye Geometric Image Quality Analysis RMSE [m] of ICPs as a function of the across track incidence angle RMSE_X RMSE_Y deg 7.5 deg 9.6 deg 9.9 deg RMSE based on the measurements of the ICPs on the final orthoimages RMSE remain almost unchanged with increasing across track incidence angle

21 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 21 Conclusions Based on the limited RapidEye sample images that were made available to us: - the RMS error calculated for ICPs in each dimension varies from 19.4 to 24.0m RapidEye's Product Specs say: RapidEye Ortho Product can meet an accuracy of 6m 1-sigma (12.7 m CE90). The highest accuracy achieved by these products will meet 1:25,000 NMAS standards. RapidEye is now proceeding to re-orthorectify data with better DEMs and GCPs. We will test this option.

22 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 22 Thank you for your attention!

23 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 23 Online archive of satellite imagery Contains JRC satellite imagery Currently: 15 TB of data - planned by end 2009: 30+ TB Currently: HR scenes (5 50 m) km² VHR (< 5 m) Image Coverage > 20 No Data Main users: - DG AGRI for Agricultural Controls - EC Services for Scientific Research Programmes - available for GMES, and for GMES (Africa)

24 Control Methods Workshop: 6-8 / 4 / 2009 [CwRS KO Meeting Campaign 2009] 24 how to get access? Are you eligible? National Administration in the agricultural field - YES Contractor for CwRS - YES Apply online to get access to the CID Portal Activate your account We hope you will find CID Portal useful!!!