SEMI-AUTOMATIC METHOD FOR SUGARCANE CROPS EXTRACTION AND REGULARIZATION USING LANDSAT-5 IMAGES, PARALELEPIPED CLASSIFIER AND DOUGLAS-PEUKER ALGORITHM
DOI:
https://doi.org/10.14393/19834071.2013.19789Abstract
This paper proposes a semi-automatic method for the extraction and regularization of sugarcane crops from Landsat-5 image. The method uses a enhancement function for increase contrast between objects in the image, the quadtree structure used for subdividing the image into homogeneous regions, parallelepiped classifier used for extract pixels that belongs to the desired pattern, edge detection from binary image with segmented regions, polygonization and regularization of contours by Douglas-Peuker algorithm. Experiments carried out with test area show that the proposed method is appropriate for applications involving semi-automatic sugarcane areas extraction, as it provided contour information with approximately 94.95% of pixel correctly classified, 1.64% of false positive and 3.4% of false negatives. Keywords: sugarcane, remote sensing, contours regularization.Downloads
Download data is not yet available.
Downloads
Published
2014-04-01
Issue
Section
Mathematics