Journal
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
Volume 7, Issue 10, Pages 4069-4080Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/JSTARS.2014.2308301
Keywords
Building detection; multivariate mathematical morphology; percentage occupancy hit-or-miss transform (POHMT); very high spatial resolution (VHSR) images
Categories
Funding
- Natural Sciences and Engineering Research Council of Canada
Ask authors/readers for more resources
The morphological hit-or-miss transform (HMT) was found to be efficient for the detection of buildings in panchromatic bands of very high spatial resolution images. The use of multispectral information was judged to be necessary to improve the results. The application of morphological operators to multispectral images is problematic, as no universal strategy for ordering the multivalued pixels of these images has been widely adopted. In this paper, we propose a new method to detect building locations based on a recently developed concept for the HMT to handle noise, called percentage occupancy HMT (POHMT). The parameters for the POHMT were defined with the aid of the top-hat by reconstruction transformation. To eliminate irrelevant locations, we applied a vegetation mask and verified locations by their proximity to shadows. The novelty of the method consists in the proposed vector-based strategy that allows for the application of the POHMT to multispectral images in order to detect building locations. Moreover, an original technique to automatically define the parameters for the POHMT was proposed. The method was tested on subsets from a pan-sharpened Ikonos image and from raw GeoEye and WorldView-2 images. The experimental results are promising.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available