Dániel Kristóf

Changes in a Wetland Ecosystem: A Vegetation Reconstruction Study Based on Historical Panchromatic Aerial Photographs and Succession Patterns

This paper presents the methodology and results of a vegetation reconstruction method based on botanical sampling, the knowledge of succession pattern, digital photograph-interpretation and automatic delineation via image segmentation. The aim is to provide a methodology for interpretation of archived black-and-white aerial photographs, which can be applied at other study sites. Our study area was the Nyíres-tó mire in the Bereg Plain (NE Hungary). Initially, botanical sampling was carried out, and this was followed by separation and identification of current vegetation types. In our study we selected automatic delineation using multi-resolution image segmentation as the method for vegetation mapping. Based on the present-day vegetation map produced and the known successional pathway of the mire, archive aerial photographs were analyzed separately in reverse chronological order to derive plant associations present at the different photograph acquisition dates. With this method we were able to make a chronological sequence of digital vegetation maps over a period of almost fifty years (1956–2002). The analysis of vegetation maps showed that forest cover increased steadily until 1988. After an artificial water supply was introduced (in 1986), the spread of tree-dominated associations became slower, and the relative cover of the different vegetation types reached a stable state.

Using big geospatial data for fast flood detection: Developments from the IQmulus project

This paper presents the first achievements of the IQmulus EU FP7 research and development project with respect to processing and analysis of big geospatial data in the context of flood and waterlogging detection.

Integrating MODIS Surface Reflectance Products into the Processing of Medium and High-resolution Satellite Images: Difficulties and Solutions through Two Hungarian Case Studies

The first aim of this study is to investigate the possibilities of integrating temporal information from Moderate Resolution Imaging Spectroradiometer (MODIS) daily time series and spatial information from high resolution satellite images by means of subpixel unmixing Our second objective is concentrated on the use of these images in order to carry out direct radiometric correction of high-resolution optical imagery of corresponding spectral wavelengths over the same area. Our results show that although the preprocessing of the MODIS surface reflectance time series diminished the noise, it was insufficient for carrying subpixel analysis. Therefore, we propose a method for the polygon-based representation of MODIS data for more accurate subpixel analysis. On the other hand, our radiometric rectification results with MODIS reference data yielded convincing results.

Environmental impact assessment of a barrage system using novel change detection methods

This paper deals with the environmental and ecological effects of a barrage system realized near the Hungarian- Slovakian border. The study area is the region of Szigetkoz in the north-western part of Hungary. Altogether 23 images, dating from 1981 to 2001, are processed and analyzed. Ten images were selected to analyze seasonal vegetation dynamics. Fourteen images are used in long-term vegetation change analysis. Multiple supplementary datasets are included in zonal statistics calculations and are also used as reference. Radiometric correction and normalization are carried out by combining traditional and novel methods, including radiance and reflectance calculations, atmospheric corrections and multi- temporal principal component analysis. Multiple methods are used and compared for change analysis such as vegetation index differencing and change vector analysis. Image segmentation, unsupervised and supervised classification methods are used to analyze land cover changes. The results are then compared and amalgamated to have precise description of the ongoing processes.

Retrospective analysis of long-term landscape evolution based on archive satellite imagery and historical maps

This study deals with the long-term retrospective analysis of the evolution of landscape features by exploiting the potential of long-term archive optical satellite imagery time series. Resulting temporal trajectories and disturbance/stability features are then used as an input to assess the current state of local ecosystems.

Integrated method for long-term environmental change detection by remote sensing

Remote sensing methods make it possible to analyze and describe landscape changes. However, one can hardly acquire sufficient data for direct long-term analysis. Multiple sensors, geometric distortions, phenological phase differences, atmospheric conditions, different solar angles and many other effects cause inter-scene variability. Furthermore, the temporal distribution of available data sets is often inhomogeneous, which tends to amplify the above-mentioned problems. In our work, we propose a methodology to cope with these difficulties for long-term environmental monitoring and quantitative change detection. A complex approach was chosen with the objective of integrating different methods and disciplines (radiometric and geometric correction, classification, image segmentation and GIS analysis, among others) to extract the maximum of information from the available data. This methodology is presented and tested on an interesting case study that deals the environmental effects of a barrage system in the northwestern part of Hungary.