Boltzmann entropy (also called configurational entropy) has been recently adopted to analyze entropy of landscape gradients (Gao et al. (2017, 2018, 2019)). The goal of belg is to provide an efficient C++ implementation of this method in R. It also extend the original idea by allowing calculations on data with missing values (Nowosad and Gao (2020)). This function accepts a RasterLayer, RasterStack, RasterBrick, matrix, or array object as an input. It allows for calculation of the relative (the relative argument equal to TRUE) and absolute Boltzmann entropy of a landscape gradient. As a default, it uses a logarithm of base 10 (log10), however log and log2 are also available options for the base argument.