Implementación de diferentes técnicas de clasificación de imágenes satelitales multiespectrales para la identificación de coberturas y zonificación del complejo de paramo de chingaza en la jurisdicción de la corporación autónoma regional del guavio corpoguavio

Abstract

The páramo biome is a collection of high-mountain tropical wetland and grassland ecosystems that provide important ecosystem services, including hydrological buffering and water supply (Patiño et al., 2021) are located in tropical high-altitude regions. mountain included between the forest areas and the perpetual snows in the Andes Mountain range limited by the super paramos with little vegetation and the high Andean region considered as the area with the greatest diversity and the highest levels of plant endemism. (Garavito, 2015). In search of the protection of the Paramos in 2016 through resolution 710, the Ministry of the Environment carried out the delimitation of the Paramo de Chingaza complex located in the jurisdiction of several municipalities in the departments of Cundinamarca, Boyacá and Meta. Likewise, it established that in accordance with the provisions of Law 1753 of 2015, the competent environmental authorities should carry out zoning and determination of use regimes according to the guidelines defined by the ministry. (MADS, 2015) In view of the foregoing, it is the obligation of CORPOGUAVIO to carry out the zoning of the Chingaza páramo in accordance with the guidelines established by the norm within its jurisdiction; The development of this research arises from the need for the corporation to carry out the zoning of the complex and its objective is the precise identification of coverage and its subsequent zoning. The development of the research included the use of multispectral satellite images SENTINEL level L1C and the application of different digital processing techniques such as atmospheric, radiometric corrections, image fusion, filtering, calculation of vegetation indices, Tasseled Cap transformations among others and the evaluation of different supervised classification methods (Minimum Distance, Maximum Probability, Mahalanobis Distance) methods unsupervised (k-means and Isodata) and vector support machines, allowed obtaining optimal results not only for coverage classification but also for zoning and allocation of use regimes that are inputs of great importance for the preparation of Plans management for the protection and conservation of protected areas. As a result, optimal coverage classifications were obtained, being the Support Vector Machine method the one that presented the best results with precision greater than 90% contrasted with the test regions, which later allowed the zoning and the preliminary assignment of use regimes.