Http://bbs.esrichina-bj.cn/ESRI/thread-58227-1-1.html
In my initial Remote Sensing Study, I could not tell the differences and relationships between sensor calibration, radiation calibration, radiation correction, and atmospheric correction. In addition, different terms are interpreted differently in different documents. For example:
Calibration is to convert the measured value obtained by the sensor Absolute brightness Or convert Physical quantities such as surface reflectivity and surface temperature Processing of relative values (Principles and Methods of Remote Sensing Application analysis, such as Zhao yingshi)
Remote Sensor calibration is to establish a quantitative relationship between the output value of each detector and the actual radiant brightness of the detector. Outbound DN In the corresponding field of view Radiant Brightness Value The quantitative relationship between them (declarative Peng ).
Radiation calibration records the Sensor Voltage or numeric value Convert Absolute radiation brightness (Quantitative remote sensing, Liang shunlin, 2009)
In fact, in simple terms, the radiation calibration is to convert the recorded original DN value to the atmospheric outer surface reflectivity, in order to eliminate the error generated by the sensor itself. There are multiple methods: lab calibration, on-star calibration, and site calibration. Formula 1 Is to set the initial DN value to radiant brightness Where lb is the radiation brightness value in the unit of W/cm2. μm. Sr (Watt / Square centimeter . Micron . The Unit is the same as the radiation brightness value. It can be seen that the radiation brightness and DN value are linear. Formula 2 Yes The radiant brightness value is converted to the atmospheric apparent reflectivity. , Formula: L λ Is the radiation brightness value, D The day-to-day distance of the astronomical unit, Esun λ Is the mean of the apparent radiant rate of the sun, θ s The height angle of the sun is measured in degrees. But in general, this part of work basically does not need to be done by the user, the correlation coefficient is included in the data header file or metadata. For example, ENVI can be used to enable the data of the modemet, that is, the reflectivity (the apparent reflectivity of the outer atmospheric layer), the radiation brightness, and the emission rate (see dsbin: Sensor calibration ). Http://bbs.esrichina-bj.cn/ESRI/viewthread.php? Tid = 56191 ).
Atmospheric correction converts the brightness or apparent reflectivity of radiation to the actual reflectivity of the earth surface to eliminate the error caused by atmospheric scattering, absorption, and reflection. There are two types: Statistical and physical.
The statistical model is based on the correlation between land surface variables and remote sensing data. The advantage is that it is easy to establish and can effectively summarize the data obtained from the local area, such as the empirical linear calibration method, for details about the internal field method, refer to the Yuni xiaoju Sina Blog: the statistical model of Radiation Correction.Http://blog.sina.com.cn/s/blog_5f4afe870100da1w.html.
On the other hand, physical models follow the physical laws of remote sensing systems. They can also establish a causal relationship. If the initial model is not good, you can add new knowledge and information to know which part of the model should be improved. However, the process of establishing and learning these physical models is long and tortuous. Models are abstract to reality. Therefore, a realistic model may be very complex and contain a large number of variables. For example, 6 s model and mortran model.
Radiation Correction refers to the correction of all radiation-related errors (including radiation calibration and atmospheric correction) generated during the acquisition of optical remote sensing data ). Relationship
Landsat TM5 Radiation Calibration and atmospheric correction (reproduced)
1. Radiation Calibration
1. As ENVI 4.4 has a dedicated radiation calibration module, the actual operation is very simple. After opening the original TM image, select
Basic tools-preprocessing-calibration utilities-Landsat TM
2. Go to the next step and select Landsat 4, 5, or 7 Based on the sensor type. Obtain the data acquisition time, sun elevation, from the header file of the remote sensing image. If you use file-open external file-Landsat-fast to open header. dat, sun elevation is complete. Here, pay attention to the choice of calibration type as radiance. Output file. The calibration is complete.
Ii. Atmospheric correction
atmospheric correction aims to eliminate the influence of atmospheric, light, and other factors on Ground Reflection, obtain real physical model parameters such as the reflectivity, radiation rate, and surface temperature of the Ground Object to eliminate the influence of water vapor, oxygen, carbon dioxide, methane, and ozone on the reflection of the earth object, eliminate the influence of atmospheric molecular and aerosol scattering. flaash can process any hyperspectral data, satellite data, and aviation data. (860nm/1135nm) , the data is composed of hymap , AVIRIS , and casi , hydice , epoch (EO-1) aisa , harp , dais , probe-1 , TRWIS-3 , sindri , mivis , OrbView-4 , Nemo . flaash You can also calibrate vertical imaging data and lateral view imaging data.
FlaashAtmospheric correctionUsed Modtran 4 + Radiation Transmission ModelOfCodeThe pixel-Level Correction is used to correct the joint effect caused by diffuse reflection, including the classification map of volume clouds and non-transparent clouds, and to adjust the spectral smoothing caused by human suppression.
FlaashPairLandsat, Spot, AVHRR, Aster, MRM, MERIS, AATSR, IRSSuch as multi-spectrum, hyperspectral data, aerial images, and custom formats for rapid atmospheric correction analysis. It can effectively eliminate the influence of air, light and other factors on the reflection of the object, and obtain real physical model parameters such as the reflectivity, radiation rate, and surface temperature of the object.
Correction Process
ClickENVI--Basic tools-Preprocessing-Calibration utilities-Flaash
Spectral-Flaash.Or clickEnvi-Spectral-flaash
1, The input data must be the data after the radiation correction, and the radiation correction data must be convertedBilOrBIPFormat (Basic tools--Convert data);
2, Edit the header file of the input data, mainly for the wavelengthWavelenth(That is, the wavelength center value of each band)And wavelength widthFWHM(Wavelength range of each band). It is not a hyperspectral data.FWHM. (ENVI--File--Edit ENVI Header)
3, After entering the data, the following dialog box is displayed:
There are two options. If the input image has different conversion factors in different bands, select the first one, and vice versa. I usedIRSAll the bands of the image are the same factor. Therefore, the second factor is used.ENVIStandard
Unit conversion scale.
Radiance scale factorsIs a unit conversion factor, if yourRadiance(Spectral sensitivity) is the standard unitW/m2 * Um * Rad, AndFlaashRequired input isUW/cm2 * sr * nm, The factor is10.
1 m = 103 Mm = 106μM = 450nm = 1012 PM (Pimi)
1 W = 103 Mw = 106μW 1Megawatt= 106Tile
RadHorizontal radian Sr Three-dimensional AngleSphere
4, Set output parameters, including:Output reflectance file.,Output directory for flaash files, AndOutput
Directory for flaash files
5, Input imaging and sensor parameters
Scene Center lacation The longitude and latitude of the center of the image. You can open the image and view the longitude and latitude of the center (you can enter the row and column numbers of the center in the following window)
Sensor altitude The sensor height (orbital height) is displayed after the correct sensor is selected.
Ground elevation The average altitude (The unit of the selected region) isKm
6,Atmospheric ModelEarth atmospheric model And aerosol Models
6Standard atmospheric model
Select the atmospheric model of the corrected region according to the following table.
The selected atmospheric model based on the data longitude and latitude and acquisition time
Moisture Inversion settings (Water retri)
Water vapor inversion settings, using two methods to remove water vapor
A. Use the Moisture Removal Model to restore the water volume of each pixel in the image
The data must have 15nmThe above spectral resolution covers at least one of the following spectral ranges:1050-1210nm (Priority),770-870nm,870-1020nm. For most sensors, the default display of moisture inversion is NoBecause most sensors do not have proper bands to compensate for the effect of moisture.
B.A single moisture factor is used for the overall image. The default value is1The moisture inversion model can be used for multi-spectral data.
Manually set the moisture band
Aerosol model (Aerosol retri )
Data band coverage is required using the aerosol model 660nmAnd 2100nmSpectrum.
A. Provides four standards Modtran Aerosol model
Rural(Country ),Urban(City ),Maritime(Ocean ),Tropospheric
B. Two aerosol inversion methods
2-band(K-T) Method(Similar to the fuzzy Reduction Method)If the adaptive black value (usually the shadow area or water body) is not found, the system uses the visibility value for calculation. Therefore, this option is also required.
Relationship between weather conditions and visibility
7Spectral Polishing(Hyperspectral) Spectral Polishing
Spectral Polishing(Hyperspectral Data)
Make the spectral curve more similar to the spectral curve of real objects
Fine-tune the spectral Curve
8. Multi-spectral data parameter settings
When the water vapor Inversion Model and aerosol model are set in the Basic settings, the parameters are set in the modify multi-spectral settings box.
Moisture Removal Model Parameters
Parameter settings of the aerosol model (Data band coverage is required for the aerosol model)660nmAnd2100nmSpectrum.) The setting values are shown in the following table:
9. Hyperspectral Data parameter settings
Automatically select Channel Definition(Recommendation)
Set Channel Definition
10. Advanced Settings
Spectral definition file: built-inAVIRIS,Hymap,Hydice,Hyperion,Casi,Aisa
Aerosol height
CO2Hybrid ratio:390ppm
Use domain correction
Use previousModtranModel Calculation Result
SetModtranSpectral Resolution of the model (recommended value)Cm-1)
SetModtranMulti-scattering model
Corner of the sky"Azimuth (for non-star sensor)