Dear Tugrul, thank you for your very interesting presentation on in-process defect detection during FFF AM of polymeric cubes. We have a similar research project for on-line defect recognition during Selective Laser Melting (SLM) AM of hardening steel plates using camera images and CNNs. One problem we have had is that the camera could not take images perpendicularly to the upper surface of the plate because of the presence of teh laser beam also operating perpendicularly. So we had to take images at an angle from the vertical axis nd this creates some problems in image analysis and consequently in CNN processing. Have you had the same kind of problem? If not, how did you solve the issue of perpendicularity of the image acquisition be the camera?
Dear Roberto, thank you for this great question. In-situ process monitoring of AM is an on-going research interest in my lab group and we use mostly off-axis camera monitoring. On -axis camera monitoring in FFF is almost impossible but laser-based AM can use optical techniques to “look through” laser beam onto melt pool area. In this paper we installed a Raspberry PI camera perpendicular to the cube geometry so that we can get images consistently.
Dear Tugrul, thank you for your very interesting presentation on in-process defect detection during FFF AM of polymeric cubes. We have a similar research project for on-line defect recognition during Selective Laser Melting (SLM) AM of hardening steel plates using camera images and CNNs. One problem we have had is that the camera could not take images perpendicularly to the upper surface of the plate because of the presence of teh laser beam also operating perpendicularly. So we had to take images at an angle from the vertical axis nd this creates some problems in image analysis and consequently in CNN processing. Have you had the same kind of problem? If not, how did you solve the issue of perpendicularity of the image acquisition be the camera?
Dear Roberto, thank you for this great question. In-situ process monitoring of AM is an on-going research interest in my lab group and we use mostly off-axis camera monitoring. On -axis camera monitoring in FFF is almost impossible but laser-based AM can use optical techniques to “look through” laser beam onto melt pool area. In this paper we installed a Raspberry PI camera perpendicular to the cube geometry so that we can get images consistently.