RASOR QGIS plugins

From Rasor Wiki
Jump to: navigation, search

Contents

EO-based exposure feature extraction plugin

The RASOR project includes improved products coming from other FP7 risk-related projects like SENSUM. The synergy among all these products is fundamental in order to provide the end-users with a set of consistent products. A QGIS plugin along with source code is presented; the algorithms are capable to extract vulnerability indicators both from medium- and very-high-resolution optical datasets, e.g. built-up areas and building footprints [1][2].

The plugin can be downloaded from the Official QGIS repository. Source code is available on GitHub.

A user and developer guide including code syntax is available here.

Dependencies: Python, Numpy, Scipy, GDAL, OpenCV, OrfeoToolbox, Scikit-image, PyEphem.

Medium Resolution High Resolution
Unsupervised built-up area Building footprints
Age of built-up area Building height
Building regularity
Building density
Roof type
Building alignment

The table below summarizes the list of algorithms available from the QGIS plugin. A more detailed explanation of the techniques follows.

Icon Name Description
Pansharp.png Pan-sharpening Pan-sharpening algorithm from OrfeoToolbox
Classification.png Classification Unsupervised/Supervised classification from OrfeoToolbox
Segmentation.png Segmentation (including optimizer option) Segmentation algorithms from OrfeoToolbox, TerraAIDA (InterImage) and Skimage (python library)
Features.png Features Computation of spectral and textural features from segments
Coregistration.png Co-Registration Co-registration algorithm designed for medium resolution. SURF and FFT alternatives are included. While services of Open CV library are used in our approach, the FFT algorithm comes from Numpy (python library)
Stacksatellite.png Stack Satellite Stack satellite workflow including co-registration and built-up extraction with 5 different methodologies
Change.png Unsupervised change detection Automatic analysis of the outcome of the object-based built-up area extraction algorithms
Footprints.png Footprints extraction Supervised extraction of building footprints
Height.png Building height Combination of shadows and footprints with acquisition date for height extraction
Density.png Building density Calculation of the density of building in an area of interest
Regularity.png Building regularity and alignment Computation of alignment and regularity of buildings

Medium-resolution-related products

Medium-resolution imagery, in particular Landsat datasets, represent a very important source of information. The capability to cover large areas with a good level of detail combined with the rich archive and the open-data policy are among the main advantages. The algorithms working on medium resolution are in the process of being included into the ESA Grid Processing On Demand (G-POD) and Cloud Toolbox systems.

Unsupervised built-up area extraction

Urban area monitoring is a key-point for the vulnerability of an area of interest. The tool provided is capable to detect built-up areas with very limited user intervention [3][4][5]. Different methodologies are proposed to leave more choice to the end users and cover the different challenges found in test-cases. An example of the obtained results is shown in Figure 1.

(thumbnail)
Figure 1. Example of hybrid-based built-up extraction of the city of Izmir (Turkey).

Age of built-up areas

Built-up areas in developing countries are generally in a process of significant expansion, leading to changes in the vulnerability of the area. The revisit time of satellites gives a helpful contribution to monitor the evolution of urban areas. Newly developed areas can be highlighted and further tracked leading to a constantly-updated status of the region of interest [6]. An example of the process is displayed in Figure 2.

(thumbnail)
Figure 2. Results of the change detection process on Izmir (Turkey).

High-resolution-related products

High-resolution optical imagery provides a closer look to the area of interest, offering a chance to extract per-building vulnerability indicators. Unfortunately, very high-resolution datasets generally do not come with open data policies.

Building footprints

A formerly developed semi-automatic algorithm is capable of detecting and extracting footprints from very high-resolution imagery. Some user skills are required to define a training set for the supervised classification. A supervised approach was selected because it has proved more reliable in respect of a pure-unsupervised one. The outcome of the classification is also used for two more indicators: roof type and building height. An example is shown in Figure 3.

(thumbnail)
Figure 3. Example of footprints extracted in a high density area in the city of Izmir (Turkey).

Building height

A renowned approach to calculating the height of buildings is based on the shadow length. The supervised classification outcome of the building footprints extraction includes a class related to shadows. The proposed algorithm is able to combine footprints and shadows with the satellite acquisition time. Height computed from shadows is automatically associated to footprints during the process without any user intervention.

Building regularity

Shape regularity of buildings is computed as a by-product of the extracted footprints. Calculation is based on the ratio between lengths of the two main directions of the building: a building is considered irregular if the ratio exceeds 4.

Roof type

The outcome of the supervised classification mentioned earlier can be useful to determine the roof material. For example, red roofs are made of tiles with high probability; cement flat roofs are instead more likely to appear grey.

Building density

Building footprints represent a fundamental input for the determination of density. The proposed algorithm is capable of processing footprints and determining the density in a radius defined by the user. An example is shown in Figure 4.

(thumbnail)
Figure 4. Example of building density computed in the city of Izmir (Turkey).

References

[1] M. Harb, D. De Vecchi, F. Dell'Acqua "Phisical Vulnerability Proxies from Remotes Sensing: Reviewing, Implementing and Disseminating Selected Techniques", Geoscience and Remote Sensing Magazine, IEEE (Volume:3 , Issue: 1 ), pp.20-33, March 2015.

[2] D. De Vecchi, M. Harb, D. A. Galeazzo, F. Dell'Acqua "Exposure Monitoring from Optical Earth Observation data: an Open-Source and integrated set of Tools", Earth Observation Open Science 2.0, 10-12 October 2015, Frascati, Rome, Italy.

[3] M. Harb, F. Dell'Acqua, D. De Vecchi "Multi-Risk building exposure and physical vulnerability mapping from optical satellite images: developing an integrated toolset", Geoscience and Remote Sensing Symposium, 2014 IEEE International, pp.1164-1166, Quebec City, QC, 2014.

[4] M. Harb, D. De Vecchi, F. Dell'Acqua "Automatic hybrid-based built-up area extraction from Landsat 5, 7, and 8 data sets", Urban Remote Sensing Event (JURSE), 2015 Joint, pp.1-4, Lausanne, Switzerland, 2015.

[5] D. De Vecchi, M. Harb, F. Dell'Acqua "A PCA-based hybrid approach for built-up area extraction from Landsat 5, 7 and 8 datasets", Geoscience and Remote Sensing Symposium, 2015 IEEE International, Milan, Italy, 2015.

[6] D. De Vecchi, D. A. Galeazzo, M. Harb, F. Dell'Acqua "Unsupervised change detection for urban expansion monitoring: an object-based approach", Geoscience and Remote Sensing Symposium, 2015 IEEE International, Milan, Italy, 2015.

Flood Map Detection plugin

RASOR Layers QGIS plug-in

The RASOR QGIS plug-in was also expanded from the previous release. It is now possible to:

• download a layer form the rasor catalog;
Download layer .PNG
• create a new exposure layer, compliant with the RASOR taxonomy;
Create new exposure.PNG
• upload a new/modified layer;
Upload layer.PNG
(thumbnail)
Figure 1 The RASOR QGIS plug-in: function to download a new layer. The user selects the layer to download from the RASOR catalog according to his user profile. The downloaded layer will appear on the QGIS layer list .
(thumbnail)
Figure 2 The RASOR QGIS plug-in: function to create a new layer. The user selects the category of exposure and the type of impact that the layer will be used for.
Personal tools
Namespaces

Variants
Actions
Navigation
Tools