Image Correction for custom RPC files

Aerial Images (frame and digital) and missing RPC parameters (missing) satellite image data can be determined by camera parameters, sensor parameters, out-of-range elements, and ground control points. The supported data includes scanned aerial photos, aerial digital photos projected from the frame center (such as vexcel ultracamd), and aerial digital photos projected from the line center (such as ads40), push Scan satellite (such as alos prism/avinir, Aster, CARTOSAT-1, FORMOSAT-2, GeoEye-1, IKONOS, IRS-C, KOMPSAT-2, moms, QuickBird, WorldView-1,
Spot ). The general steps for normal shooting correction are as follows:

(1)        In-line targeting (interior orientation, for aerial photos only)-in-line targeting will establish the relationship between camera parameters and aerial images. It uses the bar control points, frame points, and focal lengths of the aerial images for internal orientation.

(2)        Exterior orientation: the external orientation associates the points on the aerial or Weide with the actual known ground locations (geographic coordinates) and elevation. Select a Ground Control Point and enter the corresponding geographic coordinates for external orientation.

(3)        Use the digital elevation model (DEM) for direct shooting correction-this step will truly correct the aerial and Wei films. Orientation files, satellite location parameters, and Collinearity Equations will be used during the calibration process ). The linear equation is generated by the preceding two steps using the digital elevation model (DEM.

The following uses SPOT4 Pan data as an example to describe how to use the custom RPC method to correct satellite images in TIFF format. Other data has similar operations.

Step 1. Prepare data

   In addition to SPOT4 image data, more than six ground control points (including elevation information) and some image attributes, including focal length, pixel size, and incident angle are required. The DEM data file of the region where the image is located.

Step 2: Create an RPC File

First, open SPOT4 image data in ENVI and follow the steps below to build the RPC file.

I,      Internal targeting

In the main menu, select map-> build RPC. In the select input file dialog box, select the SPOT4 image data file and click OK to open the build RPC panel. The parameters in the build RPC Panel are set below.

L type (CAMERA type): pushbroom Sensor

L focallength (focal length): 1082.0

L principal point x0 (image center X coordinate) & principal point y0 (image Center Coordinate): 0

L x pixel size (MM) & Y pixel size (mm) (x/y): 0.013

L Incidence Angle Along Track (angle of incidence along the track direction): 0

L Incidence Angle wide SS track (vertical orbital direction incident angle): 16.8

L sensor line along axis (sensor forward direction axis): x

L polynomial orders (polynomial coefficient): select 2

 

Sensor Name	Focal Length (mm)	Image pixel size (mm)
Ads40	62.77	(0.0065, 0.0065)
ALOS AVNIR-2	800.0	(0.0115, 0.0115)
ALOS prism	1939.0	(0.007 cross-track, 0.0055 along-track)
Aster	329.0 (Bands 1, 2, 3n) 376.3 (band 3b)	(0.007, 0.007) Bands 1, 2, 3n, 3B
EROS-A1	3500	(0.013, 0.013)
FORMOSAT-2	2896	(0.0065, 0.0065) Pan
IKONOS-2	10000	(0.012, 0.012) Pan
IRS-1C	982	(0.007, 0.007) Pan
IRS-1D	974.8	(0.007, 0.007) Pan
KOMPSAT-2	900 pan 2250 multispectral	(0.013, 0.013)
Kodak dcs420	28	(0.009, 0.009)
Moms-02	660	(0.01, 0.01)
QuickBird	8836.2	(0.013745, 0.013745)
Spot-1 ~ 4	1082	(0.013, 0.013) Pan
Spot-5 hrs	580	(0.0065, 0.0065) Pan
Starlabo TLS	60	(0.007, 0.007)
Vexcel ultracamd	101.4	(0.009, 0.009) Pan
Z/I imaging DMC	120	(0.012, 0.012)

Table focal length and pixel resolution of common Sensors

II,      External orientation

(1)         On the build RPC panel, click the select GCPs in dsiplay button. In the select GCPs in display dialog box (Figure 1), select projection for GCPs and set the projection parameters of the control point, click OK to go to the control point selection page (exterior orientation GCPs ).

 

Figure 1 select GCPs in display dialog box

(2)         The selection of control points is the same as that in geometric correction.

(3)         Select 12 control points. The rms value is controlled at about 1 pixel. In the exterior orientation GCPs dialog box, Select Options> export GCPs to build rpcs widget, and calculate the out-of-class Element Based on the control point information.

 

Figure 2 out-of-range elements calculated using GCP

(4)         Return to the build RPC panel and you can see the calculated out-of-range element. Click OK. the maximum and minimum elevation selections (minimum elevation and maximum elevation) are displayed. The system automatically calculates a default value to obtain the elevation information of the region where the image is located.

(5)         Click OK to execute rpcs computing.

The calculated rpcs are automatically saved in the header file (. HDR) of the data file and associated with the image file.

 

Figure 3Result of rpcs parameter setting

Step 3: Execute positive shot correction

In this step, the generic RPC and RSM menu is used to provide two kinds of normal shooting Correction Methods: ground control points and no ground control points.

(1)        On the main menu, choose map> orthorectification> generic RPC and RSM> orthorectify using RPC or RSM. In the select file to orthorecloud dialog box that appears, select the SPOT4 file.

(2)        In the open orthorectification parameters dialog box, select the output parameters of the final correction, such as the DEM file.

(3)        Click OK to execute the normal shooting correction process.

 

Figure 4 Correction output parameter settings

 

From "ENVI remote sensing image processing method" Science Press 2012 third printing