Drainage Network Planning and Analysis Based on Image Segmentation - an Application in the Bandeirinha Stream - Formosa, State of Goiás, Brazil

Abstract

The present research aimed to demonstrate the feasibility of using image segmentation for better drainage extraction, automated drainage network hierarchization and main channel definition based on Horton (1945). The test area for the applications was the Bandeirinha Stream Hydrographic Basin (BSHB), which is located in the municipality of Formosa, in the state of Goiás, Brazil. The study used the Shuttle Radar Topography Mission (SRTM), with a spatial resolution of 1" x 1" (one arc second). The SRTM segmentation was performed using the Python language ODR_Obia algorithm, inserted with a QGIS plugin. Through the QGIS platform, the ODR_Hidro plugin was used to perform the delimitation of the BSHB, the extraction of the drainages and the definition of the main channel. The results were validated in the field and evaluated from the manual definition of the main channel. That was done using both the data produced by the application and the drainage data contained in the vector base of the topographic maps of the Army of Geographic Service (AGS), at the scale of 1:100,000 and 1:25,000. From the results obtained, it can be inferred that the ODR_Hidro model used for the delimitation of the BSHB, the automated hierarchization of the drainages and the definition of the main channel from the segmented SRTM image is very efficient. The field validation demonstrated its effectiveness by proving the existence of first-order drainages which are not mapped on the DSG chart at a scale of 1:100.000. This leads to the conclusion that the use of the proposed tools and the inclusion of image segmentation, aiming for the extraction of drainage network and geomorphometric characterization of watersheds, can efficiently contribute to the provision of morphometric data.