The resolution in images is a perceptible detail measure. If the resolution increases, perception of fine details, edges, clearness of the objects and image quality increases too. Video surveillance cameras usually have a standard resolution for video surveillance applications, commonly in VGA resolution (640 x 480 pixels). This video image in most of the cases does not provide enough information to identify a person or an object, the cameras with low resolution deliver poor data information and poor information in detailed images to maximize its size. If an area needs more resolution, it is necessary an algorithm that achieve this without the loss of inherent characteristics. We selected the fuzzy logic theory to solve these problems. This technique is used to improve image resolution. It helps in processes where ambiguity and vagueness in the data interpolation are present, this is due to the non-linearity of image information (edges, fine details, textures, etc.). The proposed Gaussian membership functions have non-linear characteristics, so they obtain good results in interpolation process.

A commonly problem of digital image processing systems that use video cameras for control navigation, as those used in cars or planes control, is that these systems depend on image contrast and the environmental pollution as fog, smog or rain. These environment characteristics, filters wavelengths of the light, which causes that the captured images, were modified by the video camera, decreasing its efficiency. It is proposed to improve the chromatic content of captured images, where environmental pollution is present, using the Retinex model. This algorithm implementation uses different characteristics such as lightness changes and color contrast; these characteristics produce different results for every Retinex model proposed showing differences in color and luminance modification of the captured image. In this paper are proposed and compared three different Retinex models; these models are the Simple Retinex, the Multi-scale Retinex and the Multi-scale Retinex with Color Correction.

Resumen. Una de las principales dificultades para una correcta captura de imágenes subacuáticas utilizando medios electrónicos (cámara fotográfica o de video), se presenta en el mismo ambiente subacuático en donde la iluminación y el tono de la fuente de luz cambian dependiendo de la profundidad del escenario para la captura de la imagen, esto debido a las diferentes longitudes de onda que se logran filtrar a diferentes niveles de profundidad, lo que trae como consecuencia la captura de diferentes tonos con condiciones de iluminación no uniformes, debido a la dispersión de la luz incidente y a las diferentes longitudes de onda capturadas por el sensor del dispositivo de adquisición, lo que provoca que los colores sean modificados. Como solución, se propone el uso del modelo Retinex, este modelo se basa en el entendimiento del Sistema Visual Humano (SVH-HVS) y hace uso del fenómeno de constancia del color, esto para disminuir el efecto producido mediante los problemas identificados anteriormente. Éste modelo se utilizará como una etapa de pre-procesamiento para que de esta manera se manipule la escena capturada, los resultados se verán mediante la modificación del contenido cromático de la imagen, principalmente en las áreas con poca iluminación. Finalmente, en este artículo se evaluarán los resultados experimentales de dos modelos aplicados a imágenes subacuáticas, Retinex simple y Retinex multiescala.Palabras clave: cromaticidad, brillantez, fuente de iluminación, sistema de Visión, modelos Retinex.