PDI techniques applied to the automatic identification and extraction of anthropogenic activities

Abstract

The use of Digital Image Processing (PDI) consists in adjusting the visual characteristics of the image in order to provide contributions to its interpretation and generate products that can later be subjected to further processing. Among the DIP techniques, the segmentation of objects in the image is highlighted by the possibility of distinguishing the region of interest from the background of the image. Therefore, this work aimed to perform the segmentation of anthropic activities in an image of the city of Paracatu-MG that included both the urban network of the city and the area of open-pit mineral extraction. To this end, we used a satellite image that contemplated the reality of the municipality. Initially, the Otsu method was used to perform the segmentation of the tree and underbrush vegetation and the image histogram was calculated in gray levels for each input (red channel (R); green channel (G) and; blue channel (B)). From this, weights and averages for the front and background classes of the image were calculated until an optimal threshold was found. In order to verify the possibility of segmenting the open-pit mineral extraction area and the urban mesh of the Paracatu municipality, the Region Growth (RC) method was chosen. Finally, the morphological operation of closure was performed in the segmented regions, so that the gaps were added to the original regions of the image and compared the results of the segmentation of the Otsu method with the segmentation by threshold. It was verified that it was possible to achieve the proposed objective through the PDI techniques and that the segmentation of the images can be used as a previous form of processing that can assist in computational techniques.