Background: Accuracy of brachytherapy source positions is an important part of radiotherapy quality assurance using two-dimension ion chamber array or film. It has already been confirmed that the mathematical morphology algorithm can be used for medical image processing. Purpose: This study aims to detect the dwell position of high-dose-rate brachytherapy source using EBT3 film based on mathematical morphology. Methods: EBT3 film was exposed for different step lengths of source and preset dwell position errors. The scanned digital images were denoised using contourlet transform. Black spots of 192Ir source were obtained on the basis of threshold selection and morphological operation. The positions of source were acquired by means of binary centroid algorithm. The calculated source position coordinates were compared with the reference data. Results: The error of source step length was less than ±0.2 mm

while the error of source dwell position was within ±0.5 mm. Conclusion: Mathematical morphology can be used in automatic detection of brachytherapy source positions. This method is simple and reliable. Abstract