Ingeniería en Electrónica y Telecomunicaciones-Pregrado

Permanent URI for this collectionhttps://dspace-test.ucuenca.edu.ec/handle/123456789/43

Browse

Browsing Ingeniería en Electrónica y Telecomunicaciones-Pregrado by Author "Andrade Rodas, Juan Manuel"

Filter results by typing the first few letters
Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Evaluación de métodos de corrección radiométrica y geométrica de una cámara de campo de visión amplio
    (Universidad de Cuenca, 2025-09-24) Palaguachi Dután, Felipe Sebastián; León Guanoquiza, Franklin Josué; Andrade Rodas, Juan Manuel; Palacio Baus, Kenneth Samuel
    This work evaluates the application of geometric and radiometric correction techniques on sky images captured with wide-field cameras, with the aim of mitigating distortions introduced by wide-angle lenses (140° and 180°) and improving the quality of these images for analysis in computer vision. The main objective of this work is to obtain accurate representations of sky conditions, which are essential for applications such as short-term photovoltaic forecasting. A comprehensive review of the state of the art is presented, exploring geometric models such as Kannala-Brandt, Scaramuzza, UCM, EUCM, and Mei, as well as radiometric methods including HDR image generation using Debevec, Mitsunaga, and Robertson techniques, including vignetting correction using approaches by Gao, Lopez-Fuentes, and Zheng. Comparative experiments have been designed to evaluate the performance of these methods on synthetic and real images, using metrics such as PSNR, SSIM, RPE, and radiometric uniformity. The results highlight that the Kannala-Brandt model excels in the correction of synthetic images, while EUCM and Mei offer greater accuracy in real scenarios; in the radiometric domain, the Debevec technique preserves key HDR details, and the Lopez-Fuentes method optimizes vignetting correction. Finally, the implications of these findings for sky image enhancement are discussed, and future lines of research are proposed, such as the development of hybrid models and the expansion of datasets, with an independent focus on their application for photovoltaic prediction studies.