Modern aviation gas turbine engines must be designed to meet increasingly stringent regulations on the emission of nitric oxides (NOx), carbon monoxide (CO), unburned hydrocarbons (UHCs), soot, etc. Lean Direct Injection (LDI) is a promising combustion strategy that meets these standards by burning globally lean near the blow-off limit, where NOx emissions are low. In order to characterize the LDI fuel injection process, the Diffuse Background Illumination (DBI) optical diagnostic is used to obtain a whole field of view, line-of-sight statistical description of liquid fuel presence in both non-reacting and reacting environments. First, DBI is applied to analyze a non-reacting, non-evaporating spray from a twin-fluid atomizer, providing liquid probability across the whole field of the spray to quantify the dispersal characteristics in the radial and axial directions. Next, the spatial resolution of the optical setup is varied from 25 μm/px to 200 μm/px to explore its implications on the depth of field and observable flow phenomena. It was determined that increasing the spatial resolution revealed smaller gaps between liquid segments/droplets. This interpretation was used to relate liquid probability to cumulative probability of gap sizes. Consequently, histograms of gap width at various locations were obtained to further describe the spray. This methodology provides a novel metric to quantify spray dispersal characteristics, capable of describing the hydrodynamic behavior of an injector for design validation and improvement for optimization of combustion performance. Second, DBI is implemented to study two distinct thermoacoustic instabilities in a reacting LDI flame, including both a longitudinal and swirling/hydrodynamic instability. The two operating conditions were selected to determine the suitability of DBI to detect liquid fuel in flames where the instabilities drive different droplet behavior. Simultaneous DBI and CH* chemiluminescence are phase-reconstructed via proper orthogonal decomposition (POD) to demonstrate the relationship between liquid fuel presence, heat release rate, and flame presence/stability. In the case of a longitudinal instability, it was found that DBI can be effective in characterizing fuel spray subject to poor atomization on a cyclic basis. In the case of a swirling/hydrodynamic instability, phase-averaged DBI was applied to obtain a time-resolved, line-of-sight description of liquid fuel dispersal in a reacting environment, to include liquid particle shape, size, and distribution for the whole field of view. In this case, it was also shown through the effective implementation of CH* chemiluminescence that the instability observed in the flame structure was unrelated to fuel dispersal in the combustor.