A monitoring device includes: a moving object image generation unit which receives, on a frame-by-frame basis, an overall image captured by a camera and performs inter-frame differential processing on the overall image or background differential processing between the overall image and a background image that is previously prepared; a density calculation unit which transforms the differential processed image into one-dimensional information and calculates a density indicating a degree of density of moving objects or of a crowd through frequency analysis; a model generation unit which calculates a reference density of the moving objects or of the crowd based on the density of a predetermined date and time; and a situation determination unit which compares between the density at the current time and the reference density, determines whether or not the density at the current time is different from the reference density, and generates a determination result.