Home Laser Controls & Instrumentation Division
Machine Vision Systems
Pellet Inspection System
A computerized Machine Vision Based Inspection System is developed in order to perform visual surface inspection of cylindrical fuel pellets.
This inspection system includes an imaging system that gives a high quality unfolded image of the pellet surface. The imaging system includes a high resolution camera, machine vision lens and custom designed illumination. To process the obtained image, an image processing software is developed for automatic detection of defects on the pellet surface, its quantification and analysis.
[Fig. : Machine Vision Based Inspection System]
The system generates a throughput of 2.5 second/pellet. Among the pellets accepted by the system, correct decision was made in 98.3 % of the pellets and among the pellets rejected by the system, correct decision was made in 85 % of the pellets.
The system is advantageous over human eye due to its objectivity, better reliability, high speed and accuracy, ability to quantify defects for qualitative assessment.
Development of FBTR fuel end plug metrology system
A machine vision based system has been developed for precision dimension measurement of FBTR fuel Top and bottom end plugs automatically using shadow graph technique. The imaging system is based on collimated illumination which is obtained by employing a telecentric lens and LED illumination source and a high resolution digital camera along with bi-telecentric lens to get a high quality, high contrast shadow image. The image processing software computes precise measurement of eleven different dimensions of the Top plug and seventeen dimensions for the bottom plug with an accuracy of ±10µm. A special sample holder fixture is designed by ED&D, BARC to hold the sample during measurement.
This system has been installed and commissioned at Radio Metallurgy Division (RMD), NFG, BARC. It is being used by RMD for quality assurance of FBTR fuel end plugs. The user has reported that system completes the entire measurement cycle in ~ 1 second whereas the human inspector can do few of the above measurements and still requires ~ 30 sec (minimum) time. The machine vision based systems score over human based inspection with high accuracy and speed, as machine vision based systems are independent of factors like monotony and fatigue.
[Fig. : ]
PHWR Fuel Tubes Inspection System
Machine vision based inspection system to detect surface defects on PHWR fuel tube is developed for NFC, Hyderabad. NFC has defined defects categorised as dents, pits, handling defects, tool mark, flattening, burning chattering and handling defects. The system comprises of 2 no. 8K line scan cameras and 2 no. 12” line lights to image and analyse full 500mm of the tube.
OCR based End plate Inspection System
Software with imaging setup is developed for the reading of end plate identification number. This system is standalone system where pre-processing and execution are done inside the smart camera. Results are transferred using serial communication to master PC or master data logger.