RANdom sample consensus (RANSAC) algorithm for enhancing overlapped etched track counting

In this study, a method to enhance the accuracy of overlapped etched track analysis is proposed. Counting tracks by eye is not an easy task and automated tracks counting systems are attractive key for this problem. This method supplements the deficiencies of the conventional track analysis method. A computer programme named KoreaTech Track Measurement System written in C++, which is the authors’ previous method, has been upgraded. In the proposed track analysis method, the track images captured from solid state nuclear track detectors are geometrically analysed and the number of tracks is counted. A damaged etching track shape can be restored on the track image to improve the analysis accuracy. For track restoration, the effective points are differentiated from the damaged track image. The track image is then restored by estimating the radii (small object removal) or their axis (ellipse, circle and non-circle) using the RANdom sample consensus method. Using the restored track image, the track parameters are obtained from the ellipse and then approximated to the contour of the track image to analyse the track image. Then, the total number of tracks including the overlapped tracks is counted. To verify the proposed track analysis method, experiments using actual etching track images are conducted and the results are discussed.