ABSTRACT
The ability to detect and quantify land-cover and land-use changes in the earth’s environment using remote sensing depends on sensors that can provide accurate and consistent measurements of the earth’s surface features over time. A critical step in providing these measurements is having a process to standardize image data from different sensors onto a common scale. To take full advantage of remote sensing, the data must be inherently sound. This implies an ongoing need for calibration, validation, stability monitoring, and quality assurance. To use remotely sensed data and ensure science observations of high quality, scientists need to know the following:
• What part of the electromagnetic (EM) spectrum they are looking at (spectral) • How much energy the instrument is receiving (radiometric) • Where the energy is coming from:
• Center of pixel location (geometric) • Bounds of the area from which the energy is coming (spatial)
8.1 Introduction .......................................................................................................................... 113 8.2 Need for Sensor Calibration ................................................................................................. 114
8.2.1 Applications Based on Temporal Analysis ............................................................... 115 8.2.2 Applications Based on Absolute Calibration ............................................................ 115 8.2.3 Applications Based on Mosaics ................................................................................ 115 8.2.4 Applications Requiring Surface Re«ectance Correction ......................................... 116
8.3 Typical Preprocessing Chain ................................................................................................ 116 8.4 Radiometric Calibration ....................................................................................................... 116
8.4.1 Relative Radiometric Calibration ............................................................................. 117 8.4.2 Absolute Radiometric Calibration ............................................................................ 117
8.4.2.1 Prelaunch Calibration ................................................................................ 118 8.4.2.2 Postlaunch, Onboard Calibration ............................................................... 118 8.4.2.3 Postlaunch, Vicarious Calibration ............................................................. 119
8.4.3 Cross-Calibration ...................................................................................................... 119 8.5 Radiometric Calibration Variables ....................................................................................... 120
8.5.1 At-Sensor Spectral Radiance for L0Rp Products (Q-to-Lλ) ..................................... 120 8.5.2 At-Sensor Spectral Radiance for L1 Products (Qcal-to-Lλ) ....................................... 121 8.5.3 Conversion to TOA Re«ectance (Lλ-to-ρP) ............................................................... 121
8.6 Ready-to-Use Images ............................................................................................................ 122 8.7 Summary .............................................................................................................................. 122 References ...................................................................................................................................... 122
The earth-observing (EO) sensors’ calibration accuracy and consistency over time are critical performance parameters and have a direct effect on the quality of the land-cover data products derived from on-orbit observations. As more satellite observations become available to the science and user communities, the number of science data products and the applications derived for these products continue to increase. Long-term land-cover and land-use data are often constructed based on observations made by multiple EO sensors over a broad range of spectra and on a large scale in both time and space. These sensors, either of the same type or of different types, can be operated on the same platform or different ones. Even sensors of the same type can be developed and built with different technologies by different instrument vendors and operated over different time spans. Some sensors may have been built without adequate onboard calibration and may not have gone through a comprehensive system-level prelaunch characterization; therefore, they cannot ‡rmly establish their calibration traceability or consistently maintain calibration stability.
