Neutron imaging has gained increasing attention in recent years. A notable domain is the \emph{in-situ} study of flow and concentration of hydrogen-rich materials. This demands precise quantification of the evolving concentrations. Several implementations deviate from the ideal conditions that allow the direct applicability of the Beer-Lambert law to assess this concentration. The objective of this work is to address these deviations by applying both calibration and correction procedures to ensure and validate accurate quantitative measurements during 2D and 3D neutron imaging conducted at the cold neutron source at the NeXT instrument of the Institute Laue-Langevin (ILL), Grenoble, France. Linear attenuation coefficients and non-linear correlations have been proposed to measure the water concentration based on the sample-to-detector distance. Furthermore, the effectiveness of the black body (BB) grid correction method, introduced by Boillat et al. [1], is evaluated which accounts for spurious deviations arising from the scattering of neutrons from the sample and the surrounding environment. The applicability of the Beer-Lambert law without any data correction is found to be reasonable within limited equivalent thickness (e.g., below 4 mm of water) beyond which the correction algorithm proves highly effective in eliminating spurious effects. Notably, this correction method maintains its effectiveness even with transmissions below 1%. We examine here the impact of grid location and resolution with respect to sample heterogeneity.