Facultad Regional San Francisco

Permanent URI for this communityhttp://48.217.138.120/handle/20.500.12272/109

Browse

Author: search.filters.author.Flesia, Ana GeorginaHas files: Yes

Search Results

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Learning to Detect Vegetation Using Computer Vision and Low-Cost Cameras
    (IEEE, 2020-02-27) Redolfi, Javier; Felissia, Sergio Francisco; Bernardi, Emanuel; Araguás, Gastón; Flesia, Ana Georgina
    A problem of current agriculture is the large amount of agrochemicals used to boost production due to their cost and the environmental pollution they cause. A partial solution to this problem consists in developing selective spraying techniques through the measurement of a green index that allows the selection of the precise amount of pesticide to be applied according to the specific conditions of each part of the field. Some of the problems of the existing systems are the inability to discriminate between types of vegetation and to pinpoint its location, since they only detect general patches of vegetation. In this work, we introduce a system prototype capable of measuring the presence of vegetation in an area using low-cost devices combined with current computer vision techniques. The system allows to generate a mask with the presence of vegetation in a certain area and it is also capable of distinguishing between different materials unlike current methods, which only allow to distinguish between green and non-green areas. The presented method opens the door to future research which can allow distinguishing between crops and weeds to make an even more selective application. The output of the system can be used also to design another type of weeding method that is not based on the application of agrochemicals.
  • Thumbnail Image
    Item
    Fisher Vectors for PolSAR Image Classification
    (SADIO, 2018-09) Redolfi, Javier; Sánchez, Jorge; Flesia, Ana Georgina
    In this letter, we study the application of the Fisher vector (FV) to the problem of pixelwise supervised classification of polarimetric synthetic aperture radar images. This is a challenging problem since information in those images is encoded as complex-valued covariance matrices. We observe that the real parts of these matrices preserve the positive semidefiniteness property of their complex counterpart. Based on this observation, we derive an FV from a mixture of real Wishart densities and integrate it with a Potts-like energy model in order to capture spatial dependencies between neighboring regions. Experimental results on two challenging data sets show the effectiveness of the approach.