SummaryDetection of tracks produced by α particles, protons or nuclear fission fragments in plastic detectors, viz., solid‐state nuclear track detectors, constitutes a very important tool in various areas. It is not easy for humans to count CR‐39 nuclear tracks manually, especially when the track density is very high. An automated computer program called KTTMS2, written in C++ and running with a user friendly interface, has been developed for recognition and parametric measurements of etched tracks in images captured from the surface of solid‐state nuclear track detectors. Well‐known edge detection methods were applied to estimate the precision and accuracy of nuclear track densitometry using the CR‐39 detector. Among the various routine edge detection methods, the Canny method was chosen because it was the most accurate technique. Because accuracy becomes more important as the track density increases, this allows more overlapping tracks to be detected. KTTMS2 (the proposed system) has an efficiency of 95% and can identify the noise as a background track (5%). Experimental results showed that the error percentage was reduced from 7.63% to 3.23% for high‐density tracks when the count was adjusted by the estimated overlapping tracks.