A microcomputer-based image processing system has been developed to use with solid-state nuclear track detectors for number counting, area measurement, and location identification. The image digitizer is designed to interface with a microcomputer and a standard black and white TV camera for data acquisition. A spatial dependent segmentation algorithm is proposed for track identification under nonuniform illumination and the low track-pixel population conditions. The system has been tested using Lexan samples irradiated by a Cf-252 spontaneous fission source. The counting accuracy is approximately 3% with a long term stability of 5%. Efficiency calibration was performed using samples with track densities up to 104/mm2. Results indicate that the proposed track identification scheme tends to reduce the track overlap effect and the system behaves as a nonparalyzable counting system. For track densities below 3000/mm2, either a paralyzable or a nonparalyzable counter model can be established with an accuracy better than 0.5%. For track densities below 1500/mm2, the track registration efficiency is in good agreement with manual counting results.
Read full abstract