Remote Sensing

To understand the urban dynamics of Hanoi, the project team detects current and past urban structures and land use patterns on satellite images. The analysis of the current and historical land use pattern will give insights in the historical urban development of Hanoi. This task is subdivided in a number of analytical steps based on remote sensing technologies. The base of the task is a semi-automatic classification.

The following table shows the satellite images which are the basis for the semi-automatic classification.

The main preparatory/preliminary work for the land use classification is the definition of land use classes. Each class has been described according to cartographic rules. To illustrate the respective land use class, the project team added some examples taken from the RapidEye satellite images. All land use classes have been summarized in a mapping guideline.

The first classification has been done on the RapidEye satellite image of 2012. Then, the team conducted the validation on-site. In addition to the RapidEye data, OpenStreetMap data (OSM) has been used as additional thematic data to reach a better segmentation level. The image processing software eCognition was used to carry out an object-oriented, rule-based classification. Therefore for each thematic class features, domains and rules based on that have been defined. All different thematic classes are clearly separated from each other. The features correspond to spectral measures, texture, geometry und neighborhoods. Based on these rules, the classification has been carried out automatically. The very dense and heterogeneous areas in the city center of Hanoi are a challenge compared to European cities. As follow up to the classification of the status quo, a visual validation has been done based on satellite data.

Fig.: Example of the classification based on RapidEye satellite images from 2012 (Source: DELPHI 2013)

In addition to the satellite image analysis, ground-truth investigations are necessary (ground-truth campaign). The Vietnamese partners will do the collection of ground-truth-data on-site. A mapping guideline and the setup of a geo portal are essential for this procedure. The geo portal will help to exchange data and simplify the exchange of validation rules and final results. The web-based geo portal is implemented based on open layers as open source software.

Fig.: Geo portal for the REMON project (Source: DELPHI 2013)