Speaker
Description
Radiochromic dosimetry film plays an important role in understanding the behaviour and calibrating ion beam applications. We developed a software pipeline using computer vision to investigate inconsistencies in the intensity of colour change on the area of films irradiated using the 3 MeV Tandetron at DFNA. Images of the films were obtained via scanning and analysed using Python with methods such as edge detection, binary masking and individual colour channel analysis. The dataset comprises 35 irradiated film samples (50, 100, and 150 Gy) and 7 non-irradiated controls, split across two batches of different ages. Preliminary results show that individual channel curves behave similarly to those suggested by the manufacturer in the technical sheet. We are also developing a method to calculate the area of film that was not fully irradiated within the focus of the accelerator by employing a multi-threshold model for irradiated areas. Furthermore, we hope the software developed can be employed in further experiments to determine the effect of age on radiochromic dosimetry film. The results suggest that the software methods employed are accurate in determining the qualitative changes in colour and can be further applied to determine where discrepancies in focus area arise by the end of the internship.