2

A computer-tomography device has a roentgen matrix of 20x500 dots with the resolution of 2mm in each direction. This matrix is rotating around a belt, shich transports items to be analysed. A special reconstruction algorithm produces 3D model of the items from many-many matixes captured from all 360 perspectives ( one image per 1° angle).

The problem is, the reconstruction algorithm is very sensitive to the belt speed/position. Measuring the belt position requires quite complicated and expensive positining sensors and very fine mechanics.

I wonder if it is possible to calculate the belt velocity using the roentgen-image itself. It has a width of 40mm and should be sufficient for capturing the movement. The problem is, the movement is always in 2 directions - rotation and X (belt). For those working in CT-area, are you aware of some applications/publishings about such direct measurement of the belt/table velocity?

P.S.: It is not vor medical application.

Valentin H
  • 7,240
  • 12
  • 61
  • 111
  • Position sensing expensive? Clearly you're doing something wrong then. Take a rotary encoder, attach a hard rubber wheel of known circumference to it and connect the output to a timer/counter. Pulses_per_second · circumference / pulses_per_rotation = belt_velocity. Most likely you can do numeric refocusing on the CT raw data, but I'd always go for the simpler method, engineering wise. – datenwolf Jul 10 '16 at 15:25
  • The weight of items is considerable - up to 200kg, legth of the belt is 4m, temperature may vary from -5° to 40°. The items may be pushed from one side and pulled from the other on the belt. Given this, do you still consider controling the speed with +/- 1% tolerance is a simple engineering problem? If it would work at all, the measurement of speed using CV-methods would be for free. The machine is powefull enough to perform a little bit more calculations. Btw, there are few encoders on the prototype, all together >10k$ – Valentin H Jul 11 '16 at 01:13
  • 2
    I'm not suggesting to control the speed, but just to *measure* it close to the scanning plane. The problem with trying to perform velocity measurement from the CT data is, that the belt moves perpendicular to the scanning plane and that makes velocity estimations quite difficult. In my research (group) we're dealing with the same kind of problem: Measuring fluid velocity using OCT. Movements axial to the OCT probe spot are easy because they cause a doppler shift, but perpendicular movements boil down to speckle tracking (highly inaccurate and takes a lot of averaging). – datenwolf Jul 11 '16 at 07:46

1 Answers1

2

Hmm, interesting idea.

Are you doing a full 180 degree for the reconstruction? I'd go with the 0 and 180 degree cone beam images. They should be approximately the same, minus some non-linear stuff, artifacts, Poisson noise and difference in 'shadows' and scattering due to perspective.

You could translate the 180 image along the negative x-axis, to the direction opposite of the movement. You could then subtract images at some suitable intervals along this axis. When the absolute of the sum hits a minimum the translation should be approx at the distance the object has moved between 0 and 180, as the mirror images cancel each other out partially.

This could obviously be ruined by artifacts and wonkily shaped heavy objects. Still, worth a try. I'm assuming your voltage is pretty high if you are doing industrial stuff.

EDIT: "A special reconstruction algorithm produces 3D model of the items from many-many matixes captured from all 360 perspectives ( one image per 1° angle)."

So yes, you are using 180+ degrees. You could then perhaps use multiple opposite images for a more robust routine. How do you get the full circle? Are you shooting through the belt?

Tapio
  • 1,502
  • 1
  • 12
  • 24
  • 360° actually. The rentgen generator and the sensors are rotating around the belt constantly. I've done some evaluations using movement tracking examples from OpenCV and also experimented with some other ideas. The outcome (accuracy) is very dependent on the content of the belt. However, if the belt had some plum marks, it would work perfectly. – Valentin H Aug 15 '16 at 12:39
  • using perpendicular perspectives 0 and 180 is also used. It is not that exact, but has not some disturbing artifacts from calculations. Actually it is not enougch to use only 2 shots from 0 and 180, because the sensor is not plain but half-elliptic. – Valentin H Aug 15 '16 at 12:44
  • So it's not symmetric? Could you provide us with some images? I'd be interested in seeing the 0 and 180 deg raw images from before the reconstruction. – Tapio Aug 15 '16 at 13:30
  • "Are you shooting through the belt?" Not working on this project any more, just noticed not answered question. Yes, the roentgen goes through the belt and all the constructions holding the belt. A special adjustment procedure captures the images while the belt is empty (detected by light barriers) and "calculated the away" - creates an offset tensor having offset matrix' for each perspective. – Valentin H Jun 28 '18 at 14:19
  • Interesting, nice to know that's possible. Did you end up using software or hardware to solve your problem? – Tapio Jun 29 '18 at 06:51
  • At the time I left, it was hardware. Unfortunatelly I can't provite any images. To be honest, some parameters are also changed for security reasons. In fact the resolution is higher. The system is still in development in the final testing phase. It is actually for scanning the bags for explosives etc. at the airports. – Valentin H Jun 30 '18 at 09:22