26 9 2011 9 ADVANCES IN EARTH SCIENCE Vol. 26 No. 9 Sep. 2011. CCD J. 2011 26 9 971-979. Liu Rui Sun Jiulin Wang Juanle et al. Data quality evaluation of chinese HJ CCD sensor J. Advances in Earth Science 2011 26 9 971-979. * CCD 1 2 2 2 2 1. 400047 2. 100101 4 CCD TM 2 CCD 1 CCD 0. 05% 2CCD TM 3CCD 4 3 2 CCD / 4CCD 4 TM TP732 A 1001-8166 2011 09-0971-09 1 3 2008 9 6 A B 4 CCD Charge Coupled Device 2 Thomson 4 Arkon-2 Enhanced Thematic Mapper Plus 2 6 IRS-P6 LISS3 CCD ETM + 1 2 * 2010-12-11 2011-07-31. * ISO 40801180 SB2007FY442 201109075. 1983- GIS. E-mail liur@ lreis. ac. cn
972 26 CCD 80% 2 100% LandSat TM Thematic Mapper 2. 2 CCD 1 TM 2 2. 1 CCD 4 CCD 4 3 2 600 ~ 800 m 5 CCD 452 ~ 518 nm 528 ~ 609 nm 626 ~ 693 nm 776 ~ 904 nm 4 5 4 CCD 1a 8 1 2009 8 12 TM 1b TM 1 ~ 4 1 25 30 m CCD 6 360 kmtm 180 km 2 CCD 2 TM LPGS 1 ~ 4 700 km 1 1 Fig. 1 1 Study area and its vegetation cover a. b. 1 25 a. Study area b. Vegetation cover of study area 1 250 000 Land cover dataset Table l 1 Basic parameters of experimental images HJ1A-CCD2 HJ1A-CCD1 HJ1B-CCD2 HJ1B-CCD1 TM ID L20000154803 154803 L20000155266 155266 L20000156526 156526 L20000157070 157070 LT51240262009224 124026 2009-08-11 2009-08-12 2009-08-13 2009-08-14 2009-08-12 UTC 03 07 52 03 08 17 03 08 18 03 32 37 02 46 48
9 CCD 973 2. 3 1 CCD 13 4 4 CCD 4 8 1 2 2 2 m n 8 df /dx i = 1 α = 1 p = 2 m n m n df CCD dx 2. 4. 4 3 14 2005 USGS 3 GLS 2005 Global Land Survey H = - 7 TM n P i log 2 P i i = 1 3 9 CCD P i i TM 2. 4 2. 4. 1 i j H i H j 2 ρ 2 H = H i + H j 1 - ρ 11 m = i p i 2. 4. 5 d = i "m 2 p i 16 s = i "m 3 p i /d 3 k = i "m 4 p i /d 4 i p i i ^F 2. 4. 2 ^F u v 2 = ^F u v F * u v u v ^F u v 2 F * 12 CCD Sum = m-1 n-1 ^F u v 2 ΔuΔv 4 u = 0 v = 0 1 m n u = 0 1 m - 1 v SNR = m /d 1 = 0 1 n - 1 m d 2. 4. 3 u v 3 3. 1 CCD 22
974 26 2 18 3. 2. 1 2 ~ 4 2 CCD TM CCD DN 2 10 m 0. 33 4 TM CCD CCD TM HJ1A-CCD1 155266 TM 124026 3 CCD TM 22 min 3 CCD GLS CCD TM 2005 0. 05% TM CCD HJ1A - CCD1 26. 80% 0. TM GLS2005 54 129. 41% 11. 66 TM 0 0. 00% CCD TM 0. 05% 4 CCD 4 HJ1A - CCD1 TM 2 300 600 m 10 20 4 ~ 269. 24% - 0. 18% ~ 113. TM 60 m2 34% - 3. 58% ~ 46. 00% TM - 1. 76% ~ 32. 05% CCD CCD 3. 2 CCD CCD TM HJ1A - CCD2 3-19. 07% - 45. 43% - 61. 75% - 83. 47% 1 3. 2. 2 e-science 5 4 0. 14% + 0. 06 1. 50% + 0. 15 4 6. 22% 4 CCD 6 2 /m1 Table 2 Analysis of inter - band displacement /m 154803 155266 156526 157070 124026 1 2 1 3 1 4 1 2 1 3 1 4 1 2 1 3 1 4 1 2 1 3 1 4 1 2 1 3 1 4 ΔX ΔY ΔX ΔY ΔX ΔY ΔX ΔY ΔX ΔY ΔX ΔY ΔX ΔY ΔX ΔY ΔX ΔY ΔX ΔY ΔX ΔY ΔX ΔY ΔX ΔY ΔX ΔY ΔX ΔY 36. 74-127. 82-46. 64-7. 70-41. 36 30. 58 22. 44 217. 80-20. 24 198. 66-55. 22 71. 28 65. 66 71. 73 35. 74 143. 68-15. 33 132. 03-1. 76 112. 47 31. 30 92. 14 55. 98-37. 30-1. 32-1. 30-10. 34-43. 78 73. 70 11. 88 1. 67-5. 81-2. 12-0. 35-1. 88 1. 39 1. 02 9. 90-0. 92 9. 03-2. 51 3. 24 2. 98 3. 26 1. 62 6. 53-0. 70 6. 00-0. 08 5. 11 1. 42 4. 19 2. 54-1. 70-0. 06-0. 06-0. 47-1. 99 3. 35 0. 54 RMS 9. 45 10. 06 3. 65 3. 61 5. 37 6. 37 5. 58 6. 87 9. 33 7. 74 7. 86 10. 83 6. 44 9. 79 4. 57 9. 23 7. 40 6. 79 5. 32 12. 63 3. 04 3. 86 5. 33 7. 03 6. 02 5. 88 3. 67 4. 31 5. 20 7. 46 RMS 13. 81 5. 14 8. 33 8. 85 12. 12 13. 38 11. 57 9. 39 8. 70 12. 71 2. 75 8. 29 8. 41 5. 66 9. 09 1 CCD www. cresda. com TM NASA2009 17.
9 CCD 975 Table 3 3 /m Comparison of side length error between image control points / m 154803 155266 156526 157070 124026 1-91. 09-0. 22% N /A N /A N /A N /A 3. 26 0. 01% - 0. 11 0. 00% 2-23. 04-0. 03% N /A N /A N /A N /A - 48. 81-0. 12% 37. 70 0. 05% 3 60. 44 0. 08% 118. 45 0. 16% 65. 92 0. 16% 91. 35 0. 22% 9. 44 0. 01% 4-14. 12-0. 02% N /A N /A N /A N /A - 57. 32-0. 14% - 7. 98-0. 01% 5-121. 24-0. 15% 23. 97 0. 03% - 9. 08-0. 02% 8. 14 0. 02% 3. 85 0. 00% 6 65. 25 0. 15% N /A N /A N /A N /A 34. 90 0. 08% - 3. 13-0. 01% 7-17. 75-0. 02% - 55. 97-0. 07% 23. 45 0. 06% 40. 42 0. 10% - 32. 62-0. 04% 8 34. 19 0. 03% N /A N /A N /A N /A - 12. 94-0. 03% 14. 29 0. 01% 9-69. 61-0. 20% 62. 24 0. 18% 39. 57 0. 09% 79. 57 0. 19% - 18. 56-0. 05% 10 97. 61 0. 11% - 34. 95-0. 04% - 49. 31-0. 12% - 39. 31-0. 09% - 3. 70 0. 00% 11 21. 53 0. 03% 15. 21 0. 02% 73. 62 0. 18% 93. 62 0. 22% 29. 57 0. 04% 12 44. 86 0. 05% - 31. 37-0. 03% 29. 59 0. 07% 29. 59 0. 07% 16. 31 0. 02% 13 45. 71 0. 04% 21. 59 0. 02% - 82. 11-0. 20% 82. 11 0. 20% - 13. 62-0. 01% 14 80. 12 0. 14% 64. 17 0. 11% 112. 37 0. 27% 76. 15 0. 18% 9. 59 0. 02% 112. 86 0. 01% 183. 35 0. 03% 204. 01 0. 03% 380. 73 0. 04% 41. 03 0. 00% 8. 06 0. 01% 20. 37 0. 03% 22. 67 0. 01% 27. 20 0. 04% 2. 93 0. 00% RMS 66. 54 55. 79 61. 25 53. 22 18. 65 Table 4 4 /m The overall image displacement / m 154803 155266 156526 157070 124026 ΔX ΔY ΔX ΔY ΔX ΔY ΔX ΔY ΔX ΔY - 13067. 56-6318. 73-7207. 20-15782. 34-6560. 98-11613. 55-8483. 58-14442. 33-993. 30-86. 24-768. 68-371. 69-343. 20-751. 54-385. 94-683. 15-403. 98-687. 73-45. 15-3. 92 RMS 95. 32 100. 17 45. 57 37. 71 63. 42 21. 42 30. 59 74. 33 15. 90 14. 63 RMS 138. 27 59. 15 44. 50 64. 50 21. 61
976 26 5 Table 5 Analysis of radiation accuracy 6 Table 6 Analysis of SNR DN 154803 155266 156526 157070 124026 1 30. 49 55. 50 14. 50 55. 50 14. 50 2 24. 73 48. 94 14. 31 48. 94 14. 31 3 28. 31 38. 90 14. 40 38. 90 14. 40 4 48. 75 56. 82 10. 57 56. 82 10. 57 1 22. 13 47. 72 9. 19 47. 72 9. 19 2 17. 93 42. 31 10. 68 42. 31 10. 68 3 17. 91 33. 75 11. 74 33. 75 11. 74 4 35. 93 53. 99 9. 07 53. 99 9. 07 1 29. 55 54. 56 12. 10 54. 56 12. 10 2 21. 19 47. 53 13. 27 47. 53 13. 27 3 23. 66 42. 68 13. 37 42. 68 13. 37 4 34. 05 61. 50 12. 03 61. 50 12. 03 1 27. 81 53. 80 20. 31 53. 80 20. 31 2 23. 26 47. 98 20. 63 47. 98 20. 63 3 25. 48 43. 42 19. 53 43. 42 19. 53 4 39. 38 57. 33 15. 32 57. 33 15. 32 1 70. 17 44. 92 7. 91 44. 92 7. 91 2 35. 20 42. 39 10. 34 42. 39 10. 34 3 38. 87 38. 37 12. 18 38. 37 12. 18 4 61. 91 51. 85 9. 65 51. 85 9. 65 1 3. 83 5. 19 4. 51 2. 65 5. 68 2 3. 42 3. 96 3. 58 2. 33 4. 10 3 2. 70 2. 87 3. 19 2. 22 3. 15 4 5. 38 5. 96 5. 11 3. 74 5. 37 5. 11 12 1 25 7 CCD TM Table 7 7 Analysis of SNR for different land cover types 154803 155266 156526 157070 124026 8. 54 7. 48 6. 93 5. 36 7. 71 2. 04 1. 90 2. 91 2. 62 2. 07 2. 79 2. 21 2. 69 2. 99 2. 25 10. 54 8. 73 6. 08 4. 82 11. 85 1. 43 1. 84 3. 04 1. 45 2. 09 3. 2. 3 8 4 CCD TM 55. 37% 154803 40. 71% 155266 50. 25% 156526 61. 48% 157070 CCD TM TM 4 TM 36. 60% 66. 13% 55. 81% 49. 26% 4 Fig. 4 Distribution map of image radiance 5 40. 21% HJ1A - CCD1 TM "1. 74% "0. 08 2 4 3 2 155266 124026 5 40% TM 4 CCD
9 CCD 977 8 3. 2. 5 Table 8 Sharpness results of experimental images 154803 155266 156526 157070 124026 1 5. 01 3. 67 4. 89 5. 64 13. 12 2 5. 66 4. 28 4. 90 6. 40 8. 03 3 7. 00 5. 07 6. 50 7. 64 11. 74 4 6. 11 4. 41 5. 44 6. 68 11. 49 Fig. 5 5 4 3 2 Comparison of image Sharpness Standard false color composite of band 4 3 2 Δ = 1 2 m 4 6 10 CCD TM 4 TM 98. 59% 154803 98. 94% 155266 100. 30% 156526 100. 01% 157070 CCD TM 4 TM 95. 29% 101. 54% 100. 91% 100. 10% 4 TM 4. 3 db 3. 2. 4 9 4 TM 88. 92% 104. 80% 94. 83% 80. 30% CCD TM CCD 4 TM 81. 24% 96. 31% 95. 22% 91. 97% CCD TM 4 3 2 6 Table 9 9 Amount of information results 154803 155266 156526 157070 124026 1 4. 25 3. 82 4. 20 4. 14 5. 05 2 4. 30 3. 93 4. 26 4. 46 4. 40 3 5. 06 4. 57 4. 89 5. 02 5. 13 4 4. 84 4. 40 4. 72 4. 95 5. 14 18. 45 16. 73 18. 07 18. 57 19. 72 7. 22 8. 51 7. 70 6. 52 8. 12 Fig. 6 Comparison of power spectral analysis 10 /db Table 10 Summation of each component of image power spectral /db 154803 155266 156526 157070 124026 1 87. 34 86. 29 87. 19 87. 04 91. 26 2 86. 47 85. 34 86. 20 86. 04 84. 70 3 85. 66 84. 64 86. 70 86. 52 85. 10 4 87. 53 92. 18 93. 11 92. 59 91. 26
978 26 4 1 CCD 0. 05% 11-14. 2 10 20 2 CCD DN TM DN 4 Thomson G. Evaluation of Russian Arkon-2 Earth observation sat- TM ellite J. The Imaging Science Journal CCD CCD TM CCD 2009 6 37-42. DN 6 Zhang Zengxiang Wang Xiao Wang Changyao et al. National land cover mapping by remote sensing under the control of interpre- CCD ted data J 216-224. 7 Gutman G Byrnes R Masek J et al. Towards monitoring landcover and land-use changes at a global scale The global land sur- 3 CCD 4 3 2 11 2 216-224. CCD TM vey 2005 J / somg 2008 74 1 6-10. CCD 4 TM CCD 4 TM 5 CCD 4 2 by our landsat ground station and U. S. NOAA J References 1 Zhao Shuhe Feng Xuezhi Zhao Rui. Evaluation on data quality and geometric correction of China-Brazil resources satellite No. 1 data in Nanjing area J. Remote Sensing Technology and Application 2000 15 3 170-174.. 2000 15 3 170-174. 2 Wang Qinjun Tian Qingjiu. Quality evaluation of LISS3 image from IRS-P6 satellite J. Geography and Geo-Information Science 2007 23 3 11-14.. IRS-P6 LISS3 J. 2007 23 3 3 Yang Zhongdong Gu Songyan Qiu Hong et al. CBERS-1's CCD image quality evaluating and cross calibrating study J. Journal of Remote Sensing 2004 8 2 113-120.. CCD J. 2004 8 2 113-120. 2005 53 3 163-173. 5 Jia Fujuan Wu Yanlin Huang Ying et al. Design and on-orbit application of CCD camera on HJ-1A /1B Saterllites J. Spacecraft Engneering 2009 6 37-42.. -1A 1B CCD J.. Journal of Geo-Information Science 2009 11 2. J. 2009. Photog Rammetric Engineering and Remote Sem- 8 Franks S Masek J Headley R et al. Large Area Scene Selection Interface LASSI methodology of selecting landsat imagery for the global land survey 2005 J. Photogrammetric Engineering and Remote Sensing 2009 75 11 1 287-1 296. 9 Wang Xinmin Zhang Lei. Analysis of geometric fidelity of systematically corrected Landsat-5 TM image data J. Remote Sensing of Environment China 1989 4 4 293-299.. TM J. 1989 4 4 293-299. 10 Wang Xinmin Shao Beien Dai Zixin et al. An accuracy comparison between the geodetically corrected TM images produced. Remote Sensing of Environment China 1987 2 1 73-79.. NOAA TM J. 1987 2 1 73-79. 11 Zhang Xia Zhang Bing Zhao Yongchao et al. Image quality assessment for the infrared Multi-Spectral scanner of the Chinese- Brazil Earth resources satellite J. Journal of Image and Graphics 2002 7 A 6 581-586.. J. 2002 7 A 6 581-586. 12 Liu Jiangui Zheng Lanfen Tong Qingxi. Estimation of signal to J. noise ratio of remote sensing images J. Journal of Basic Science
9 CCD 979 and Engineering 1999 7 4 360-365. J. 2006 31 7. J. 569-572. 1999 7 4 360-365. 16 Balboa R Grzywacz N. Power spectra and distribution of contrasts 13 Wang Hongnan Zhong Wen Wang Jing et al. Research of measurement for digital image definition J. Journal of Image and Graphics 2004 9 7 829-831. 17 Chander G Markham B L Helder D L. Summary of current ra-. J. diometric calibration coefficients for Landsat MSS TM ETM + 2004 7 829-931. 14 Cheng Jicheng Guo Huadong Shi Wenzhong. Uncertainty of Remote Sensing Data M. Beijing Science Press 2004.. M. 2004. 15 Lin Zongjian Zhang Yonghong. Measurement of information and uncertainty of remote sensing and GIS data J. Geomatics and Information Science of Wuhan University 2006 31 7 569-572.. 2011 26 1 66-74. of natural images from different habitats J. Vision Research 2003 43 24 2 527-2 537. and EO-1 ALI sensors J. Remote Sensing of Environment 2009 113 5 893-903. 18 Zhu Yunqiang Sun Jiulin Song Jia et al. E-Geoscience research and practice A case show of north eastern Asia joint scientific exploration and cooperative research platform J. Advances in Earth Science 2011 26 1 66-74.. e-science J. Data Quality Evaluation of Chinese HJ CCD Sensor Liu Rui 1 2 Sun Jiulin 2 Wang Juanle 2 Liao Xiuying 2 1. College of Geographical Science Chongqing Normal University Chongqing 400047 China 2. State Key Laboratory of Resources and Environmental Information Systems Institute of Geographical Sciences and Natural Resources Research Chinese Academy of Sciences Beijing 100101 China Abstract The Chinese environmental mitigation HJ satellite CCD sensors are capable of large area and alltime monitoring and have a great advantage in coverage and frequency of repeated observations. However as a relative newly remote sensing sensor it is essential to evaluate the data quality for further applications. Based on the introduction of HJ CCD sensor and combined with LandSat TM data the geometry accuracy and radiance quality of HJ CCD data was quantitative evaluated in this paper. The results showed that 1There is no deviation between the 4 bands of HJ CCD and compared with TM very small geometric distortion is detected 0. 05%. However there's large geometric error in both X and Y direction. 2There is no obvious difference in radiation range and signal to noise ratio between the two sensors but the image sharpness of CCD data can only a- chieve about half of TM which leads to application limitation for object recognition and feature extraction etc. 3 The CCD data is rich in amount of information and standard false colour composition of band 4 3 2 is the best band combination choice. Meanwhile the power spectral of CCD data guaranteed the abundant texture feature which will ensure the application accuracy in studies such as classification. 4Near infrared and red band showed the best quality between the CCD bands and followed by green and blue band. Key words Geometric accuracy Radiance quality Image sharpness Information entropy Power Spectral.