1

Topic: aerial photo scan resolution (mm) IO question

Hello-

I am trying to orthorectify and resample an old aerial photo that does not have a calibration report.  I know the focal length of the camera and have entered in 4 fiducial point coordinates by using the pixel values of the fiducial marks.  And according to our image library I also know that the image was scanned in at 600 dpi or 2360 dpmm.  The dimensions of the file are 7200 x 7122 x 1 [BSQ]. I am not sure how to enter in the x and y scan resolution properly as needed in the 'Interior Orientation Setup, no camera calibration' step.  I have tried entering 2360 and 1/2360 (which is 0.00042373) but no luck so far. The problem becomes evident when I try and optimize my gcps- the master image window looks good but the slave window does not and after cycling through the gcps roughly once the program eventually crashes. The error message I get says:

AERIAL_GCPS_OPTIMIZATION:AERIAL_GCPS_OPTIMIZATION_CORE:RESAMPLING_IMG_SINC_CORE: Array has too many elements

Thank you very much for your help,

RP

2

Re: aerial photo scan resolution (mm) IO question

Hello,

sorry to come back to you so late, got too busy lately smile
The crash that you get comes from your orthorectification matrices that are corrupted. Generally the cause is to be found in either (or both) the DEM that contains outliers/gap or an incorrect Interior Orientation.
If you are using a DEM, make sure that it does not contain any outliers or missing data on the area acquired by the aerial image.
Regarding the Interior Orientation, the x and y scan resolution must be entered in mm. For example, if the scan was operated at 15 microns resolution (i.e. 1000/15 dpmm), you need to enter 0.015 for the x and y resolution.  That said, I am confused with the scan resolution of you image. You said that it was scanned at 600dpi or 2360dpmm. This is not the same. 600dpi (dot-per-inch) corresponds to a scan resolution of 42 microns, whereas 2360dpmm corresponds to 0.42 microns. You need to make sure what is the correct resolution. My guess is that the 2360dpmm is not correct as the resulting film size would be (7200x0.00042= 3mm !!!), and the 600dpi seems more probable (7200x0.042=30cm).

Hope this helps solve your problem. Let me know if it doesn't...

Cheers,
Francois

3

Re: aerial photo scan resolution (mm) IO question

Hello-

Thank you for your response.  I tried it again, entering in 0.042 for the x and y scan resolution and got the same error.   I checked over the DEM and didn't see any issues- it's the same DEM I successfully used to orthorectify a different image of  the same area (this other image has a camera report) so I don't think the DEM is the problem.

In trying different things, I decided to orthorectify the image I had successfully done in the past, but this time going through the process as if I didn't have the camera report.

Unfortunately I got the same error. 

So I can orthorectify an image if I use the camera report information but get an error if I try and do it without that information.

I entered in the fiducial coordinates in sequence- point (1) and opposite point (2), point (3) and opposite point (4), not sure what else I could be doing wrong. I'd be happy to send you my images and associated files, etc if that would be helpful.

Thank you very much for your help.


Best,

RP

4

Re: aerial photo scan resolution (mm) IO question

Hello,

Hum, that is strange indeed !! I can't see the problem from there... Could you please send me (fayoub[at]gps.caltech.edu) the IO file of the image with a calibration report, the IO file of the same image assuming no calibration report, and the GCPS file (I assume that you used one). Thanks. I'll look into this asap.

All the best for 2009 !!!

Cheers,
Francois

5

Re: aerial photo scan resolution (mm) IO question

Hello Ryan,

OK, so I got to look at your data. The problem does not come from you, it's a bug in COSI-Corr, the same that was reported in the forum (James_Hollingsworth topic): The image reference system is indirect compared to the ground's one. I should have guessed it !!!  Anyway, the turnaround that was proposed in the James_Hollingsworth case cannot be applied to your problem (no camera report), and there is no easy fix unfortunately. However, we are supposed to release soon (by the end of the month) a new COSI-Corr version with that problem corrected. If you can wait until then that would be perfect, but if you are in a hurry you can send me your data and I will process the GCPS optim and EO computation for you.

Thanks anyway for reporting the problem to us, and sorry about the inconvenience.

Francois

6

Re: aerial photo scan resolution (mm) IO question

Dear Francois,

Greetings from UCSB!  I just installed the new version of cosi-corr and tried to again orthorectify an image without its camera report.  I actually do have the report but wanted to use this image as a test run for other images lacking reports. Unfortunately I am still having a problem, not sure if it is me or still a bug in cosi-corr.  The problem becomes evident when I try and optimize my GCPS. 

For initially creating the fiducial point file, in the ‘FiducialX’ and ‘FiducialY’ fields I have just been entering the values I see in the ImageX and ImageY fields when the cursor is centered over the marks, not sure if that is right or not.  To create an IO file for use in GSPC optimization, I now need to enter in a value for the focal length field (even if I don’t know what that is, not having the report). If I leave it unfilled I get an error when trying to start the optimization process, “AERIAL_GCPS_OPTIMIZATION: AERIAL_GCPS_OPTIMIZATION_CORE: EXTERIOR_ORIENTATION_CORE: Error in computing altitude(alt. null)

7

Re: aerial photo scan resolution (mm) IO question

Hello Ryan,

to process aerial images without having the camera calibration report, you'll need the focal length (usually the calibrated focal length is printed on the film's edge), the scan resolution in mm, and be able to locate the fiducial marks on the digitized image.

In consequence, the error you got when you left the focal length field empty is normal (although I admit, a "missing focal length" warning message should be displayed instead).

Regarding the fiducials points selection, if you do not have the camera report, FiducialX and FiducialY are not used and you can set them to a junk value (0 for example).

Now, regarding the GCPS optimization of your slave image, I assume that you orthorectified the master image on a UTM grid, selected tie-points between the orthorectified master and the raw slave, and converted these tie-points to gcps. If these steps were successfully taken, then the error comes most probably from the slave IO file or the GCPS files. Take a look a the IO file and see if the affine coefficients of the camera-to-image and image-to-camera transform look good. If you have some doubt, post the IO content on this forum and we can have a look at it together. Double check the GCPS file also (idem, post it on the forum if you are not sure of its correctness).

Let me know what you find out...

Francois

8

Re: aerial photo scan resolution (mm) IO question

Dear Francois,

Thank you for your reply and suggestion.  I am going to post text from 4 files here:
1.    the io file when all camera report information is included
2.    the io file when only focal length is used
3.    the optimized gcps file generated with the camera report info
4.    the optimized gcps file generated w/out the camera report info

sorry if this gets a little long. There are differences between the files, I am not exactly sure how to interpret them.
Thank you very much,
RP

;Interior orientation file generated the :Thu Apr 02 15:34:49 2009
;Camera Calibration Report Available :
yes

;Focal of the camera (in meters)
0.1532150000000

;Principal Point location in millimeter (x, y):
0.0190000000000    -0.0120000000000

;Fiducial Points
;Fiducial (x,y) (millimeter), Image (x,y)(pixel)
;Note: If Camera Calibration not available, fiducial (x,y) are not used
-104.9870000000000  -104.9870000000000   764.1100000000000   698.4400000000001
105.0110000000000   104.9930000000000 10766.7800000000010 10702.5599999999990
-104.9790000000000   104.9860000000000 10770.0000000000000   701.6700000000000
105.0000000000000  -104.9870000000000   764.3300000000000 10701.6700000000000
-112.9950000000000     0.0080000000000  5768.1099999999997   318.4400000000000
112.9940000000000     0.0140000000000  5765.7799999999997 11082.7800000000010
-0.0040000000000   113.0080000000000 11149.2199999999990  5701.7799999999997
0.0060000000000  -112.9820000000000   384.7800000000000  5700.1099999999997

;Coefficients for affine transformation
;from Camera to Image
;x(pix) = a0 + a1*x(mm) + a2*y(mm)
;y(pix) = b0 + b1*x(mm) + b2*y(mm)
5765.7513342888715
-0.0092712106530
47.6398855081383
5701.5622150314712
47.6317130906771
0.0083311130928

;Coefficients for affine transformation
;from Image to Camera
;x(mm) = a0 + a1*x(pix) + a2*y(pix)
;y(mm) = b0 + b1*x(pix) + b2*y(pix)
-119.6797935194056
-0.0000036717036
0.0209944150629
-121.0511014957197
0.0209908129856
0.0000040854621

;Distortions measured with radial distance
;- yes : radial distance in millimeters
;- no : field angle in decimal degree
yes

;Radial distortion values
0.0000000000000     0.0000000000000

;Radial distortion coefficients - lenses correction
0.0000000000000     0.0000000000000     0.0000000000000

;Tangential distortion coefficients - lenses correction
0.0000000000000     0.0000000000000     0.0000000000000

;Atmospheric distortion correction
yes





;Interior orientation file generated the :Thu Apr 02 16:01:17 2009
;Camera Calibration Report Available :
no

;Focal of the camera (in meters)
0.1532150000000

;Scan Resolution in millimeter (x, y):
0.0420000000000     0.0420000000000

;Fiducial Points
;Fiducial (x,y) (millimeter), Image (x,y)(pixel)
;Note: If Camera Calibration not available, fiducial (x,y) are not used
764.0000000000000   698.1100000000000   764.0000000000000   698.1100000000000
10766.7800000000010 10702.5599999999990 10766.7800000000010 10702.5599999999990
10770.0000000000000   701.7800000000000 10770.0000000000000   701.7800000000000
764.4400000000001 10701.4400000000010   764.4400000000001 10701.4400000000010
5768.1099999999997   318.4400000000000  5768.1099999999997   318.4400000000000
5765.6700000000001 11082.6700000000000  5765.6700000000001 11082.6700000000000
11148.8899999999990  5701.6700000000001 11148.8899999999990  5701.6700000000001
384.6700000000000  5699.8900000000003   384.6700000000000  5699.8900000000003

;Coefficients for affine transformation
;from Camera to Image
;x(pix) = a0 + a1*x(mm) + a2*y(mm)
;y(pix) = b0 + b1*x(mm) + b2*y(mm)
5766.5699999999997
23.8095238095238
0.0000000000000
5700.8199999999997
0.0000000000000
23.8095238095238

;Coefficients for affine transformation
;from Image to Camera
;x(mm) = a0 + a1*x(pix) + a2*y(pix)
;y(mm) = b0 + b1*x(pix) + b2*y(pix)
-242.1959400000000
0.0420000000000
0.0000000000000
-239.4344400000000
0.0000000000000
0.0420000000000

;Distortions measured with radial distance
;- yes : radial distance in millimeters
;- no : field angle in decimal degree
yes

;Radial distortion values
0.0000000000000     0.0000000000000

;Radial distortion coefficients - lenses correction
0.0000000000000     0.0000000000000     0.0000000000000

;Tangential distortion coefficients - lenses correction
0.0000000000000     0.0000000000000     0.0000000000000

;Atmospheric distortion correction
yes




;File generated the : Thu Apr 02 15:56:55 2009
;GCPS optimization report:
;GCP / Easting misregistration (meter) / Northing Misregistration (meter)
;     1   -13.232325    47.451948
;     2   -17.849489     8.887216
;     3     2.670258     7.704268
;     4     1.328219    -3.140744
;     5    -4.681032    10.572638
;Easting mis-registration in meters (Average / Standard Deviation):     -6.3674987     8.0513609
;Northing mis-registration in meters (Average / Standard Deviation):     14.3117880    17.2885513
;Norm mis-registration in meters (Average / Standard Deviation):     15.6643645    19.0714031
;
;Files:
;- Image: U:\SCI_NAPP_USGS\AB1NAPPW0006921_146_a.tif
;- Reference Image: R:\NYSA\ryan\Channel Islands\Santa Cruz\imagery\Aerial photos\2005 naip\naip_2005_sci_1pt5m
;- DEM: U:\SCI lidar\sci_gnd.tif
;- IO file: U:\SCI_NAPP_USGS\1994_cosi_corr\1994_camera_report_io.txt
;- GCPS/Tied Points file: U:\SCI_NAPP_USGS\1994_cosi_corr\gcps_5pts.txt
;
;Resampling: Sinc
;Correlator: Frequential
;- Window Size: 128
;- Mask Treshold:0.900000
;- Nb Robust Iteration:2
;- Resampling:0
;Easting ouput pixel size (meter):   1.50
;Northing ouput pixel size (meter):   1.50
;
;Convergence quality report:
;Loop number, Avg x, Avg y, Norm Avg xy, Stdev x, Stdev y, Norm Stdev xy
;     1    -3.939759    -7.800326     8.738809     1.783210     4.939354     5.251386
;     2    -0.594143     0.881010     1.062631     2.012902     1.523311     2.524332
;     3    -0.773973    -3.153500     3.247090     5.013931     7.962898     9.409954
;     4     7.825186     6.723936    10.317211    19.125842     6.206662    20.107722
;     5    -6.367499    14.311788    15.664365     8.051361    17.288551    19.071403
;
;Original GCPS:
;GCP number, Longitude (degree), Latitude (degree), Altitude (meter), X (pix), Y (pix), SNR, Opti, dE (meter), dN (meter), dA (meter)
;     1   -119.8653338224     34.0093152764      63.7039    3416     752   1.00000   1     0.0000     0.0000     0.0000
;     2   -119.8464700860     33.9781001557     114.2673    5429    4653   1.00000   1     0.0000     0.0000     0.0000
;     3   -119.8023416106     34.0010671921     208.0801   10209    1576   1.00000   1     0.0000     0.0000     0.0000
;     4   -119.8165244321     33.9611983120     -22.9403    8634    6790   1.00000   1     0.0000     0.0000     0.0000
;     5   -119.8522402633     33.9906194821      52.4246    4812    3082   1.00000   1     0.0000     0.0000     0.0000
;
;Optimized GCPS:
-119.8652347771     34.0095932275      55.0607    3416     752   0.96372   1     0.0000     0.0000     0.0000
-119.8460168596     33.9781866859      96.6407    5429    4653   0.96648   1     0.0000     0.0000     0.0000
-119.8024606538     34.0011342845     208.8339   10209    1576   0.96439   1     0.0000     0.0000     0.0000
-119.8176755390     33.9609054614     -20.4626    8634    6790   0.96198   1     0.0000     0.0000     0.0000
-119.8520853925     33.9907752407      53.4021    4812    3082   0.94511   1     0.0000     0.0000     0.0000




;File generated the : Thu Apr 02 16:51:37 2009
;GCPS optimization report:
;GCP / Easting misregistration (meter) / Northing Misregistration (meter)
;     1    -9.231442    51.105833
;     2   -28.114144   -35.083309
;     3          NaN          NaN
;     4     1.583115   -77.738107
;     5          NaN          NaN
;Easting mis-registration in meters (Average / Standard Deviation):    -14.8798101     8.6992957
;Northing mis-registration in meters (Average / Standard Deviation):     24.8106818    39.8842882
;Norm mis-registration in meters (Average / Standard Deviation):     28.9305838    40.8219818
;
;Files:
;- Image: U:\SCI_NAPP_USGS\AB1NAPPW0006921_146_a.tif
;- Reference Image: R:\NYSA\ryan\Channel Islands\Santa Cruz\imagery\Aerial photos\2005 naip\naip_2005_sci_1pt5m
;- DEM: U:\SCI lidar\sci_gnd.tif
;- IO file: U:\SCI_NAPP_USGS\1994_cosi_corr\1994_no_camera_report_focal_lngth_io.txt
;- GCPS/Tied Points file: U:\SCI_NAPP_USGS\1994_cosi_corr\gcps_5pts.txt
;
;Resampling: Sinc
;Correlator: Frequential
;- Window Size: 128
;- Mask Treshold:0.900000
;- Nb Robust Iteration:2
;- Resampling:0
;Easting ouput pixel size (meter):   1.50
;Northing ouput pixel size (meter):   1.50
;
;Convergence quality report:
;Loop number, Avg x, Avg y, Norm Avg xy, Stdev x, Stdev y, Norm Stdev xy
;     1    -5.066423    -2.202323     5.524388    17.667791    31.495444    36.112516
;     2     3.469249    10.700889    11.249209    22.098107    44.674904    49.841482
;     3     8.849616     6.526080    10.995700    41.297215    53.914965    67.913794
;     4     8.361773    -6.289485    10.463119    23.895976    58.250940    62.961811
;     5   -14.879810    24.810682    28.930584     8.699296    39.884288    40.821982
;
;Original GCPS:
;GCP number, Longitude (degree), Latitude (degree), Altitude (meter), X (pix), Y (pix), SNR, Opti, dE (meter), dN (meter), dA (meter)
;     1   -119.8653338224     34.0093152764      63.7039    3416     752   1.00000   1     0.0000     0.0000     0.0000
;     2   -119.8464700860     33.9781001557     114.2673    5429    4653   1.00000   1     0.0000     0.0000     0.0000
;     3   -119.8023416106     34.0010671921     208.0801   10209    1576   1.00000   1     0.0000     0.0000     0.0000
;     4   -119.8165244321     33.9611983120     -22.9403    8634    6790   1.00000   1     0.0000     0.0000     0.0000
;     5   -119.8522402633     33.9906194821      52.4246    4812    3082   1.00000   1     0.0000     0.0000     0.0000
;
;Optimized GCPS:
-119.8648983915     34.0118445547      24.0249    3416     752   0.96389   1     0.0000     0.0000     0.0000
-119.8462154545     33.9774368292     106.9288    5429    4653   0.41635   1     0.0000     0.0000     0.0000
-119.8022424320     34.0003024930     167.6824   10209    1576   0.00000   1     0.0000     0.0000     0.0000
-119.8179969675     33.9590277203     -35.8826    8634    6790   0.00399   1     0.0000     0.0000     0.0000
-119.8525472027     33.9903663871      37.7200    4812    3082   0.00000   1     0.0000     0.0000     0.0000

9

Re: aerial photo scan resolution (mm) IO question

Hello Ryan,

OK, thanks for these files. It's clearer now. The problem does come from the IO file (without camera report).
There is an incoherence between the IO files and what you wrote in your first post.

In your first post you wrote:
- the image size is 7200x7122 and the film was scanned at 600 dpi or 2360 dpmm. From there I suggested the correct scan resolution to be 600dpi (i.e., 42 microns per pixel) as it would correspond to a film size of ~30cm.

Now, when I look at your IO file, I can see 2 things:
- the top-left and bottom-right fiducial points have respectively pixel coordinates of (764,598) and (10766,10702), which indicates an image size much larger than 7200x7122.

- the scan resolution cannot be 600dpi nor 2360dpmm !! The affine transformation from Image to Camera coefficients (given by IO with calibration report) suggests a scan resolution of 21 microns (1210dpi). Indeed, if you look closely at the coefficients, you will see that the pixel size is around 0.021mm.

The scan resolution (42 microns) you used with the IO without calibration is therefore incorrect, and causes the GCPS optim to fail. The scan resolution should be 21 microns instead. May be you have 2 sets of scanned image (one at 600dpi and the other one at 1210dpi) and mixed them up ? smile

Let me know if it works better now...

Francois

10

Re: aerial photo scan resolution (mm) IO question

Dear Francois,

Thank you for your response- I indeed had 2 files of the same image with different scan resolutions!  So that solves that part. But when I entered a scan resolution of 0.021 and created a new io file (posted below) and tried to use it to optimize my points, the slave and the master screens still did not match up properly. I went back and looked at the affine coefficients, and while the coefficients are about the same as the ones found in the io file created using the camera report, their order is not the same.  For example, the original (correct) io file with the camera report contains the following:

;Coefficients for affine transformation
;from Camera to Image
;x(pix) = a0 + a1*x(mm) + a2*y(mm)
;y(pix) = b0 + b1*x(mm) + b2*y(mm)
5765.7513342888715
-0.0092712106530
47.6398855081383
5701.5622150314712
47.6317130906771
0.0083311130928

If you compare that to the coefficients below you'll see what I mean.  As an experiment, I manually edited the order of the coefficients in my new 0.021/no camera report io file to match those of the camera report io file and tried the GCPS optimization again- this time it worked. So I'm making progress, but I'm still not sure if it's a problem with me or a remnant bug in cosi-corr.

Thanks again for all your help, this is wonderful feedback/support.

RP






;Interior orientation file generated the :Mon Apr 06 09:13:48 2009
;Camera Calibration Report Available :
no

;Focal of the camera (in meters)
0.1532150000000

;Scan Resolution in millimeter (x, y):
0.0210000000000     0.0210000000000

;Fiducial Points
;Fiducial (x,y) (millimeter), Image (x,y)(pixel)
;Note: If Camera Calibration not available, fiducial (x,y) are not used
764.0000000000000   698.1100000000000   764.0000000000000   698.1100000000000
10766.7800000000010 10702.5599999999990 10766.7800000000010 10702.5599999999990
10770.0000000000000   701.7800000000000 10770.0000000000000   701.7800000000000
764.4400000000001 10701.4400000000010   764.4400000000001 10701.4400000000010
5768.1099999999997   318.4400000000000  5768.1099999999997   318.4400000000000
5765.6700000000001 11082.6700000000000  5765.6700000000001 11082.6700000000000
11148.8899999999990  5701.6700000000001 11148.8899999999990  5701.6700000000001
384.6700000000000  5699.8900000000003   384.6700000000000  5699.8900000000003

;Coefficients for affine transformation
;from Camera to Image
;x(pix) = a0 + a1*x(mm) + a2*y(mm)
;y(pix) = b0 + b1*x(mm) + b2*y(mm)
5766.5699999999997
47.6190476190476
0.0000000000000
5700.8199999999997
0.0000000000000
47.6190476190476

;Coefficients for affine transformation
;from Image to Camera
;x(mm) = a0 + a1*x(pix) + a2*y(pix)
;y(mm) = b0 + b1*x(pix) + b2*y(pix)
-121.0979700000000
0.0210000000000
0.0000000000000
-119.7172200000000
0.0000000000000
0.0210000000000

;Distortions measured with radial distance
;- yes : radial distance in millimeters
;- no : field angle in decimal degree
yes

;Radial distortion values
0.0000000000000     0.0000000000000

;Radial distortion coefficients - lenses correction
0.0000000000000     0.0000000000000     0.0000000000000

;Tangential distortion coefficients - lenses correction
0.0000000000000     0.0000000000000     0.0000000000000

;Atmospheric distortion correction
yes

11

Re: aerial photo scan resolution (mm) IO question

Dear Ryan,

OK, glad to hear that the first part is solved. Now, regarding the error during GCPS optim,  I am the one to blame. This is a remnant error in COSI-Corr similar to the one in the James_Hollingsworth  topic but in the case of no camera report. Sorry about that. It will be corrected in the next COSI-Corr release.
For now, what you'll need to do is to express the affine coefficients in a camera reference system that is positive counter-clockwise, which is not the case currently (positive clockwise). So pratically, you need to invert the x and y of camera reference system. It means x(mm) = y(mm) and y(mm)=x(mm), so if I consider the IO file just above, you will correct the affine transformation coefficients this way:

;Coefficients for affine transformation
;from Camera to Image
;x(pix) = a0 + a1*x(mm) + a2*y(mm)
;y(pix) = b0 + b1*x(mm) + b2*y(mm)
5766.5699999999997
0.0000000000000
47.6190476190476
5700.8199999999997
47.6190476190476
0.0000000000000


;Coefficients for affine transformation
;from Image to Camera
;x(mm) = a0 + a1*x(pix) + a2*y(pix)
;y(mm) = b0 + b1*x(pix) + b2*y(pix)
-119.7172200000000
0.0000000000000
0.0210000000000
-121.0979700000000
0.0210000000000
0.0000000000000

Let me know how it works out, and sorry for the trouble.

Francois

12

Re: aerial photo scan resolution (mm) IO question

Dear Francois,

Just so I’m clear.  If the original io file looks like this, with coefficient lines numbered in parentheses,

Coefficients for affine transformation
;from Camera to Image
;x(pix) = a0 + a1*x(mm) + a2*y(mm)
;y(pix) = b0 + b1*x(mm) + b2*y(mm)
(1) 5766.5699999999997
(2) 47.6190476190476
(3) 0.0000000000000
(4) 5700.8199999999997
(5) 0.0000000000000
(6) 47.6190476190476

;Coefficients for affine transformation
;from Image to Camera
;x(mm) = a0 + a1*x(pix) + a2*y(pix)
;y(mm) = b0 + b1*x(pix) + b2*y(pix)
(7) -121.0979700000000
(8) 0.0210000000000
(9) 0.0000000000000
(10) -119.7172200000000
(11) 0.0000000000000
(12) 0.0210000000000

The fix is to switch the order so the new file looks like:

Coefficients for affine transformation
;from Camera to Image
;x(pix) = a0 + a1*x(mm) + a2*y(mm)
;y(pix) = b0 + b1*x(mm) + b2*y(mm)
(1) 5766.5699999999997
(3) 0.0000000000000
(2) 47.6190476190476
(4) 5700.8199999999997
(6) 47.6190476190476
(5) 0.0000000000000

;Coefficients for affine transformation
;from Image to Camera
;x(mm) = a0 + a1*x(pix) + a2*y(pix)
;y(mm) = b0 + b1*x(pix) + b2*y(pix)
(10) -119.7172200000000
(11) 0.0000000000000
(12) 0.0210000000000
(7) -121.0979700000000
(8) 0.0210000000000
(9) 0.0000000000000

So the way the lines switch is different for the ‘camera to image’ and ‘image to camera’ coefficients?

Thank you very much,

RP

13

Re: aerial photo scan resolution (mm) IO question

Hello Ryan,

Exactly !!
You need to invert the x and y of the camera reference system (x(mm) and y(mm) in the IO), so if you look at the affine transformation:

;from Camera to Image
;x(pix) = a0 + a1*x(mm) + a2*y(mm)
;y(pix) = b0 + b1*x(mm) + b2*y(mm)

;from Image to Camera
;x(mm) = a0 + a1*x(pix) + a2*y(pix)
;y(mm) = b0 + b1*x(pix) + b2*y(pix)

inverting the x(mm) and y(mm) corresponds to a different switch for ‘camera to image’ and ‘image to camera’ coefficients.

Francois