I. Introduction
Currently there is an explosion of devices that perform the recording of images and videos, such as the use of surveillance cameras, which we find in commercial establishments, in the streets and even in homes, these cameras perform a continuous recording, so their analysis after registration, can be a computational task that requires multiple resources, In a review of the literature, works related to the analysis of satellite images are found, these images have a fundamental characteristic that has a lot of information, so the size of the image is heavy and requires workstations with a lot of capacity [1]. These satellite images are presented in a configuration of using many cameras integrated into a single device, which is called multi-band or multi-spectral, for which it is required to perform operations between them, as the images are represented in the form of data matrix, perform operations between them requires computational capacity, to perform these operations we resort to the use of GPUs, for which these matrix operations are performed [2]. These satellite images can be analyzed in different ways, one of them is through the color information they have, so we resort to the use of the analysis of color characteristics, to analyze their behavior [3].