Examinando por Autor "Mancera Gómez, Cristian Camilo"
Mostrando 1 - 1 de 1
Resultados por página
Opciones de ordenación
- ÍtemClasificación de imágenes de calles: distinguiendo entre limpias y sucias utilizando redes neuronales convolucionales mediante UAS(Fundación Universitaria Los Libertadores. Sede Bogotá., ) Mancera Gómez, Cristian Camilo; Gutiérrez Ramírez, Sebastián Alejandro; Melo Daza, Pedro FernandoThe project focuses on image classification using artificial intelligence developed in Python. This tool aims to distinguish between two categories from a set of over 500 images. These images are captured from the air with the help of an Unmanned Aerial Vehicle (UAS). The project involves the development of Python code that employs artificial intelligence to identify and classify a large number of images into two categories: dirty streets and clean streets. The acquisition of these images is done from the air using a UAS, which follows a predetermined flight plan, has the necessary airworthiness permits, and adheres to rigorous safety and flight mapping protocols. This tool is essential for the project as it facilitates data collection and information identification using integrated data augmentation techniques in the code, allowing for image modifications as detailed in the following sections.The process of developing the code also requires proper training to ensure its correct operation and the generation of valid information without errors or distortions. For training, previously collected images from online sources were used, as well as photographs taken during a session in the Sopó area, totaling 237 images uploaded into the code. Despite the challenges involved in this entire process, a reliability percentage of over 85% was achieved. This means that the system offers considerable credibility and experiences minimal information loss.