Técnicas y usos en la clasificación automática de imágenes

Laburpena

The production and generation of visual information through mobilephones and cameras is massive. Also and mainly using remote sensing, through the acquisition of images of the earth's surface by means of planes, spacecraftand satellites that capture and serve data on meteorology, oceanography,geology, geography, geolocation, security, etc. These image capture instrumentsgenerate visual information every day that is impossible to process manually,and therefore various techniques and methods are used to automatically extractuseful knowledge. This literature review aims to understand the techniques anduses of automatic classification of images. To this goal, the Scopus and WoSdatabases were used to locate documents on the automatic classification ofimages published between 2008 and 2018. The resulting records were searchedfor their full texts, analysing the content to find out the most recurrenttechniques and their applications. With all this, it has become clear that thethree most commonly used techniques for the automatic classification of imagesare decision trees, neural networks and support vector machines, with theapplication of a wide variety of automatic classification, which seeks toautomate repetitive processes, complex monitoring and inspection, urbancontrol and development or recognition and assessment after natural disasters,among other matters.