The VISTA Variables in the Vía Láctea survey is a multi-epoch, near-infrared (NIR) search for RR Lyrae stars, which trace the structure and evolutionary history of the Galactic Bulge. To measure accurate light-curves for these and other variable sources in the crowded fields of the Bulge, we have built a pipeline to perform difference imaging analysis (DIA) on VVV data. This is the first time DIA has been carried out with a large-scale NIR survey, which presents certain specific challenges such as the large pixel scales and high variable background. This thesis begins with an investigation into the wider problem of reference image selection for DIA. Through this study, we found that an alternative formulation of traditional DIA - Target Image Convolution - yields better results for VVV. This technique is employed in the core image subtraction engine for the VVV DIA pipeline. This work also details the pipelines for carrying out calibration, period-finding and Fourier modelling on the light-curves of DIA sources, as well as for obtaining their dereddened NIR magnitudes and colours. To validate the pipelines, the magnitude calibration is checked against another NIR survey (in the most challenging Galactic centre field) and analysis results are crosschecked against the OGLE-IV catalogue of RR Lyrae (in a low-extinction region of the Galactic bulge). Both studies are used to determine the efficacy of the existing pipeline and identify areas for improvement. DIA generally shows a significant improvement over existing aperture photometry. Using the NIR properties of VVV DIA counterparts to the OGLE-IV RR Lyrae, we demonstrate that it is possible to identify new variables with this pipeline, including RR Lyrae.

No description available.

No description available.

No description available.