I hope this message finds you well. I am writing to inquire about certain features of MTFMapper and to seek guidance on its compilation under the Windows operating system:
1. Firstly, I am interested in knowing whether MTFMapper can not only generate a marked chart for chart orientation results but also export additional data such as the center distance between two symbols on the chart, along with the Pitch, Roll, and Yaw values, into a text file simultaneously.
2. Secondly, I noticed that the 'ubuntu_build_notes.txt' provided in the source code directory facilitates smooth compilation on Ubuntu 22.04.2. However, I am currently working on a Windows platform and was wondering if there exists any documentation or guidelines specifically addressing the compilation process for Windows.
I am aware that you may receive numerous inquiries, and I apologize for any inconvenience this may cause.
Best regards,
Last edit: Rany 2024-08-06
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Some answers:
1. When you run mtf_mapper with the --chart-orientation option (I'm referring to command-line operation) it produces a file called fiducial_correspondence.txt which provides both the image space coordinates (in pixels) and the chart space coordinate (in mm) of the circular fiducial targets. The chart space coordinates are simply the expected x and y positions for a given target design, so you can use the contents of fiducial_correspondence.txt to perform any camera calibration or pose estimation you desire.
2. I'd love to be able to give you documentation on building on Windows, but I do not have any. The ubuntu instructions are still a decent guide, i.e., you have to build OpenCV locally in Visual studio, and install a suitable QT SDK, and a few of the other dependencies such as tclap. Then you just have to keep guiding cmake until it can locate all these components. I'm currently travelling so I do not have access to my Windows build machine, but if you still need help with this in two weeks from now, just remind me and I can provide you with an example of a working cmake configuration.
I would like to add a warning regarding the chart orientation functionality. It has been pointed out to me that this feature does not work reliably with different focal lengths and chart sizes; my own testing of this feature was too limited to really run into these issues, so for a given chart size, lens focal length, and working distance the feature may work just fine. I think I know how to fix this (there are some constraints on the optimization of the extrinsic parameters that are likely the cause of the errors), but I have not had time to fix or test this. If you are willing to do your own evaluation of the accuracy if this feature I would be happy to provide you with the required changes to the code to remove the aforementioned constraints. Just let me know!
Regards,
Frans
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi Frans,
Thank you very much for your help.
I have now obtained the desired results through fiducial_correspondence.txt！
Regarding the issue of chart orientation you mentioned, I am currently testing with an A3-sized chart at 30 cm, and the results appear to be reliable.
I plan to use MTF Mapper to test sharpness in the automated testing of a camera's production，I've just started learning about this area, so there's a lot of knowledge I'm still missing. There are some documents where I could use your guidance：
1.How should I choose the size of the chart according to the lens, will the size of the test chart affect the results of the sharpness test at the same test distance?（Currently, I am using the 'lensgrid_a3.pdf' file from the 'mtfmapper_sample_test_charts_0.5_v5' package,It is printed with positive film）
2.Also, concerning the --focal-ratio ratio, with the lens parameters below, the ratio=3.3/7.2, is that right?
Max Sensor 1/2.5" COMS
EFL(Focal Length) 3.3+5%
Max lmage Height 7.2mm
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Dear [Developer/Support Team]:
I hope this message finds you well. I am writing to inquire about certain features of MTFMapper and to seek guidance on its compilation under the Windows operating system:
1. Firstly, I am interested in knowing whether MTFMapper can not only generate a marked chart for chart orientation results but also export additional data such as the center distance between two symbols on the chart, along with the Pitch, Roll, and Yaw values, into a text file simultaneously.
2. Secondly, I noticed that the 'ubuntu_build_notes.txt' provided in the source code directory facilitates smooth compilation on Ubuntu 22.04.2. However, I am currently working on a Windows platform and was wondering if there exists any documentation or guidelines specifically addressing the compilation process for Windows.
I am aware that you may receive numerous inquiries, and I apologize for any inconvenience this may cause.
Best regards,
Last edit: Rany 2024-08-06
Hi Rany,
Some answers:
1. When you run mtf_mapper with the --chart-orientation option (I'm referring to command-line operation) it produces a file called
fiducial_correspondence.txtwhich provides both the image space coordinates (in pixels) and the chart space coordinate (in mm) of the circular fiducial targets. The chart space coordinates are simply the expected x and y positions for a given target design, so you can use the contents offiducial_correspondence.txtto perform any camera calibration or pose estimation you desire.2. I'd love to be able to give you documentation on building on Windows, but I do not have any. The ubuntu instructions are still a decent guide, i.e., you have to build OpenCV locally in Visual studio, and install a suitable QT SDK, and a few of the other dependencies such as tclap. Then you just have to keep guiding cmake until it can locate all these components. I'm currently travelling so I do not have access to my Windows build machine, but if you still need help with this in two weeks from now, just remind me and I can provide you with an example of a working cmake configuration.
I would like to add a warning regarding the chart orientation functionality. It has been pointed out to me that this feature does not work reliably with different focal lengths and chart sizes; my own testing of this feature was too limited to really run into these issues, so for a given chart size, lens focal length, and working distance the feature may work just fine. I think I know how to fix this (there are some constraints on the optimization of the extrinsic parameters that are likely the cause of the errors), but I have not had time to fix or test this. If you are willing to do your own evaluation of the accuracy if this feature I would be happy to provide you with the required changes to the code to remove the aforementioned constraints. Just let me know!
Regards,
Frans
Hi Frans,
Thank you very much for your help.
I have now obtained the desired results through
fiducial_correspondence.txt！Regarding the issue of chart orientation you mentioned, I am currently testing with an A3-sized chart at 30 cm, and the results appear to be reliable.
I plan to use MTF Mapper to test sharpness in the automated testing of a camera's production，I've just started learning about this area, so there's a lot of knowledge I'm still missing. There are some documents where I could use your guidance：
1.How should I choose the size of the chart according to the lens, will the size of the test chart affect the results of the sharpness test at the same test distance?（Currently, I am using the '
lensgrid_a3.pdf' file from the 'mtfmapper_sample_test_charts_0.5_v5' package,It is printed with positive film）2.Also, concerning the --focal-ratio ratio, with the lens parameters below, the
ratio=3.3/7.2, is that right?Max Sensor 1/2.5" COMS
EFL(Focal Length) 3.3+5%
Max lmage Height 7.2mm