Abstract: With the continual expansion of Indian nuclear power program there has been a quantum jump in inspection requirements of PHWR fuel pellets at Nuclear Fuel Complex, Hyderabad. Every year over 30 million UO2 pellets have to be subjected to visual inspection. Carrying out inspection for such huge numbers is cumbersome, strenuous and requires enormous workforce. NFC has adopted, automation to meet the ever increasing inspection load. This involved development and deployment of vision based automated pellet visual inspection system. UO2 pellets are rotated mechanically under CCD cameras to acquire pellet surface image and this image is processed by custom made software for inspection. As on date around 0.3 million pellets have been cleared by using this system. The paper brings out the details of our experience, issues and subsequent actions being taken to address them which include revision of image processing algorithms, change of imaging devices etc.