Forum

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - hagorms

Pages: [1]
1
Bug Reports / Re: calibration p1 p2
« on: March 23, 2018, 02:03:01 PM »
so change them in xml. 
Your xml file and gui have different names for same values.
Let's say after I calibrate my camera I have p1 = 0.001 and p2 = 0.003. If I load them through GUI and through XML, i will have different callibrations. 

2
Bug Reports / Re: calibration p1 p2
« on: March 23, 2018, 01:54:18 PM »
Yes, they  stay the same after importing xml back.
I think you just forgot to change labels in GUI

3
Bug Reports / calibration p1 p2
« on: March 23, 2018, 01:19:43 PM »
Hello.
I try to precalibrate cameras using other software and I think I found a bug.
After I define calibration as it is shown in the picture below in the gui I set p1 = 7 and p2 = 8, but when I export cameras it gives me :
<p1>8</p1>
 <p2>7</p2>
 <p3>9</p3>
  <p4>10</p4>

4
General / Agisoft Calibration
« on: March 14, 2018, 03:41:33 PM »
Hello.
II have two synthetic images img1 and img2.
I know exact coordinates where images were taken img1(-0.125, 0 , 0) img2(0.125, 0, 0)
I know exact angle between axis and cameras.
How do I load this parameters to agisoft and lock them?
   
Best regards.

5
Python and Java API / loading calibration
« on: January 23, 2018, 01:01:23 PM »
Hello.
After i calibrated cameras with multiple frames I use this script to save calibration:

   
Quote
def exportCalibration(self, path, filename):
        self.superChunk.exportCameras(path + filename, format=PhotoScan.CamerasFormatXML)
        i = 0
        for camera in self.superChunk.cameras:
            camera.sensor.calibration.save(path +str(i)+ ".xml")
            i = i + 1

To load Calibration i use this script:

 
Quote
  def loadCalibration(self, calibrationPath, calibName, calibFilesPrefix):
       
        for frame in self.superChunk.frames:
            frame.importCameras(calibrationPath + "\\" + calibName, format=PhotoScan.CamerasFormatXML)
            cameraId = 0
            for camera in frame.cameras:
                camera.open(self.cameraFolders[0] + self.fileVec[0][cameraId])
                sensor = frame.addSensor()

                calibration = PhotoScan.Calibration()
                calibration.load(calibrationPath +"\\" + calibFilesPrefix + str(cameraId) + ".xml")
                sensor.calibration = calibration
                sensor.width = camera.sensor.width
                sensor.height = camera.sensor.height
                sensor.focal_length = calibration.f
                sensor.pixel_width = 0.0055
                sensor.pixel_height = 0.0055
                sensor.type = camera.sensor.type
                sensor.fixed = False
                camera.label = sensor.label
                camera.sensor = sensor
                cameraId = cameraId + 1
           

So my question:
When I don't see aligned cameras as on the screenshot,
,
 does it mean, that cameras were not loaded correctly?
What is the correct way to load cameras? Shall I rename images first?
 

6
General / Multi frame calibration
« on: November 15, 2017, 01:26:12 PM »
Hello!
I have several cameras and would like to calibrate their position alltogether.

I use this answer http://www.agisoft.com/forum/index.php?topic=4076.msg21014#msg21014  to calibrate cameras individually.

But I have some question about agisoft:

1) Is camera position  optimized for all camera frames?

2) if i use your gui to load photos to chunks and than combine chunks in frames, how does agisoft understand which photo belongs to which camera?

3) Can I somehow calibrate camera position, if i don't have any frame for the timestamp (the calibration board can't be seen at the same time from camera 1 and camera

Best regards. Matvey.

7
General / Agisoft and opencv calibration.
« on: November 14, 2017, 01:47:48 PM »
Hello. I used opencv calibration to calibrate cameras, using aicon markers.
Now I have  4x4 transform matrix for each camera which contains a rotation matrix and a translation matrix ( 4x4 matrix with 0001 in the last line).
Now what I want is to take my opencv calibration and put it straight to agisoft, but your y and z are negative (for 1rst camera your rotation is eye with 1 -1 -1 1) .
So maybe you can help me with the next question: which transfrom shall I make  to get a valid agisoft matrix?
 next the same question in russian:

Доброго времени суток. Я использовал опенсв калибровку, чтобы откалибровать камеры, используя аикон маркеры.
После калибровки имею 4на4 матрицу ткоторая состоит из матрицы поворота и матрицы сдвига (всего 4на4 с 0 0 0 1 в последней строке)
Борюсь с тем, чтобы взять эту калибровку опенсв и засунуть ее в агисофт, однако в агисофте z and y оси отрицательны (как понял: для первой камеры матрица eye  имеет 1 -1 -1 1 вместо стандартных единиц)
Может быть Вы мне поможете со следующим вопросом: какие действия нужно произвести, чтобы перейти в систему поворота в агисофте.

if need, i can provide opencvxml and agisoft xml and all scripts i did to parse it

8
General / Allignment after calibration loading
« on: November 13, 2017, 01:38:41 PM »
Hello.
I calibrated 4 cameras using opencv and after that i made a script that transfoms my opencv calibration to agisoft calibration.
The question I have: do i need to allign photos after i uploaded calibration?  What is generic preselection? does it fix camera positions and distortion parameters?
After I import my calibration, i don't have any tie points which makes me feel sad, but after allignment, everything looks well.
Best regards. Mat

9
General / Re: agisoft circular marker detection
« on: October 05, 2017, 05:31:27 PM »
Thank you. It worked.
So the main question now  is:
Does the quality and accuracy of agisoft aicon markers detection (for allignment ) influence on circular non coded marker detection?
Can I use markers I found with circle marker detection to improve the allignment or there is no need to do this?
Best regards Matvey (hagorms)

10
General / agisoft circular marker detection
« on: September 29, 2017, 12:37:13 PM »
Hello.
How do I detect circular non coded markers? Unfortunatelly can't detect any circles =(

u can use this test images :
https://drive.google.com/open?id=0B8c5PQjVtSaTQXI3Z0tGNG9RQ1U
https://drive.google.com/open?id=0B8c5PQjVtSaTQlh3WjFZVWFuVGs
https://drive.google.com/open?id=0B8c5PQjVtSaTZmJ1TG1HOG1xLTg


I use this python string line :
chunk.detectMarkers (type = PhotoScan.TargetType.CircularTarget, tolerance = 50, inverted = False, noparity =False)

Everything works for CircularTarget12bit .




11
Python and Java API / Re: calibration load
« on: July 05, 2017, 03:28:02 PM »
Solved. Instead of loading calibration for each sensor, I used "import cameras for chunk"
Thank you for good documentation.
I just looked in the wrong place

12
Python and Java API / calibration load (solved)
« on: July 05, 2017, 02:13:31 PM »
Hello.
I think I don't understand something with load for calibration.
According to python api:

class PhotoScan.Calibration has method
load(path, format=’xml’)
Loads calibration from file.
Parameters
• path (string) – path to calibration file
• format (string) – Calibration format in [’xml’, ‘australis’, ‘photomodeler’, ‘calibcam’,
‘calcam’, ‘inpho’, ‘usgs’].
Returns success of operation
Return type boolean
page 11

my code :
   calibration_4cameras = PhotoScan.Calibration
   calibration_4cameras.load("C:\\Users\\OSLab\\Desktop\\autoprocess\\agisoft_calib.xml", format = "xml")
gives me an error
TypeError: descriptor 'load' requires a 'PhotoScan.Calibration' object but received a 'str'
What do I do wrongly ?

Pages: [1]