Uma solução para análise de vegetação, separação e localização de zonas de manejo em imagens aéreas utilizando sistemas com baixo poder de processamento

• Ano de defesa: 2021
• Autor(a) principal: Santos, Suzane Gomes dos
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso embargado
• Idioma: por
• Instituição de defesa: Universidade Federal da Paraíba; Brasil; Informática; Programa de Pós-Graduação em Informática; UFPB
• Palavras-chave em Português: Processamento de Imagem; Sistema embarcado; VANT; NDVI; K- means; Image processing; Embedded system; UAV; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/123456789/20309

The use of modern technologies, such as sensors, portable monitoring platforms and Unmanned Aerial Vehicles (UAVs), in Precision Agriculture, has brought great contri- butions to the agronomic development process, by a more accurate field monitoring and detecting vegetation issues in almost real time in a non-intrusive way, with low cost and time consumption. This work presents an automatic system of vegetation monitoring and separation in management zones using aerial images captured by Unmanned Aerial Vehicles (UAVs). The system analyzes crop images and divides them into regions identified by colors for easy visualization of problematic areas. To do so, it uses the Normalized Difference Vegetation Index (NDVI) and the K-means clustering algorithm. Besides, the system allows locating each management zone through geographical coordinates. The algorithm was optimized, allowing its execution in embedded systems. As result, processing times of approximately 6 seconds were achieved for an image with 9,387,360 pixels using a conven- tional computer, and 0.54 seconds for an image of 870,400 pixels using an embedded system.