Timeline for Fitting of 2d data points with a function considering scaling, rotation and translation
Current License: CC BY-SA 4.0
9 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 17, 2019 at 11:21 | comment | added | kglr |
@mrz, re shear, you can use the estimated transformation, estimatedtransformation = trans[sx, sy, tx, ty, \[Theta]] /. nlm["BestFitParameters"], and follow the suggestion in Ulrich's answer and comment.
|
|
| Jan 17, 2019 at 11:17 | comment | added | kglr |
@mrz, re "most reliable" maybe you can compare some fit measure, e.g., RootMeanSquare of residuals.
|
|
| Jan 17, 2019 at 11:00 | comment | added | mrz | Is in your model also a shear possible? | |
| Jan 17, 2019 at 10:58 | comment | added | mrz |
All values which you receive are similar as of Ulrich Neumann and Cesareo except for the translation. Ulrich Neumann got {1502.57, -13.1302} and Cesareo got {1513.99, -7.23374}. You got {1496.06, -13.01}. My approximate estimation for the vertical translation is = Mean[data2[[All, 2]] - data1[[All, 2]]]=-11.95 and for the horizontal translation = Mean[data2[[All, 1]] - data1[[All,1]]]=1507.1.Which values of the there solutions are most reliable?
|
|
| Jan 17, 2019 at 10:40 | comment | added | kglr |
@mrz, theta / Degree transforms theta (in radians) to degrees.
|
|
| Jan 17, 2019 at 10:38 | comment | added | mrz | Great. How do you receive this value from theta? | |
| Jan 17, 2019 at 10:36 | comment | added | kglr |
@mrz, the estimated angle is 359.607 Degrees (6.27632/Degree).
|
|
| Jan 17, 2019 at 10:33 | comment | added | mrz | This is a very interesting solution. Thanky you. Only th angle I do not understand: In the two solutions below both get Ulrich Neumann and Cesareo get an angle of about 0.38 to 0.39 degree. | |
| Jan 17, 2019 at 10:27 | history | answered | kglr | CC BY-SA 4.0 |