This document presents an ensemble classification algorithm for hyperspectral images aimed at improving accuracy in thematic mapping for environmental monitoring and land cover classification. The proposed method incorporates morphological profiles, local binary patterns, support vector machines, and genetic algorithms, achieving classification accuracies of 93% for the Indian Pines dataset and 92% for the Pavia University dataset. The work highlights the challenges of high-dimensional data in hyperspectral imaging and the effectiveness of combined classifiers to enhance data representation and classification outcomes.