Application of the Mean-shift Segmentation Parameters Estimator (MSPE) to VHSR satellite images: Tetuan-Morocco

[EN] Image segmentation is considered as crucial step dealing with Object-Based Image Analysis (OBIA) and different segmentation results could be achieved by combining possible parameters values. Optimal parameters selection is usually carried out on the basis of visual interpretation; therefore, de...

Full description

Bibliographic Details
Authors: Benarchid, O., Raissouni, N., Sobrino, J.A., El Ayyan, A.
Format: article
Publication Date:2015
Country:España
Institution:Universitat Politècnica de València (UPV)
Repository:RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Language:Spanish
OAI Identifier:oai:riunet.upv.es:10251/80573
Online Access:https://riunet.upv.es/handle/10251/80573
Access Level:Open access
Keyword:EPSM
Satélite
Muy alta resolución espacial
Tetuán (Marruecos)
MSPE
Satellite
Very high spatial resolution
Tetuan (Morocco)
Description
Summary:[EN] Image segmentation is considered as crucial step dealing with Object-Based Image Analysis (OBIA) and different segmentation results could be achieved by combining possible parameters values. Optimal parameters selection is usually carried out on the basis of visual interpretation; therefore, defining optimal combinations is a challenging task. In the present research, Mean-shift Segmentation Parameters estimator (MSPE) proposed tool is applied to automate the selection of segmentation parameters values to Very High Spatial Resolution (VHSR) satellite images in the region of Tetuan city (Northern Morocco). MSPE estimates the parameters values for the Mean-shift Segmentation (MS) algorithm. However, this algorithm needs as inputs: i) existing vector database and, ii) spectral data to define automatically the segmentation parameter values. Finally, application of the MSPE method on different landscape’ types show accurate results with Under-Segmentation (US) values ≤0.20 for industrial, residential and rural zones, while for dense residential area values of 0.35.

Similar Items