Semi-automated analysis of three-dimensional track images

In the past, three-dimensional (3-d) track images in solid state detectors were difficult to obtain. With the introduction of the confocal scanning la ser microscope it is now possible to record 3-d track images in a non-destr uctive way. These 3-d track images can latter be used to measure typical tr ack parameters. Preparing the detectors and recording the 3-d images howeve r is only the first step. The second step in this process is enhancing the image quality by means of deconvolution techniques to obtain the maximum po ssible resolution. The third step is extracting the typical track parameter s. This can be done on-screen by an experienced operator. For large sets of data however, this manual technique is not desirable. This paper will pres ent some techniques to analyse 3-d track data in an automated way by means of image analysis routines. Advanced thresholding techniques guarantee stab le results in different recording situations. By using pre-knowledge about the track shape, reliable object identification is obtained. In case of amb iguity, manual intervention is possible. (C) 2001 Elsevier Science Ltd. All rights reserved.