The primary objective of this research was to evaluate the potential for monitoring forest change using Landsat ETM and Aster data. This was accomplished by performing eight change detection algorithms: pixel post-classification comparison (PCC), image differencing Normalized Difference Vegetation Index (NDVI), Soil-Adjusted Vegetation Index (SAVI), Transformed Difference Vegetation Index (TDVI), principal component analysis (PCA), multivariate alteration detection (MAD), change vector analysis (CVA) and tasseled cap analysis (TCA). Methods, Post-Classification Comparison and vegetation indices are straightforward techniques and easy to apply. In this study the simplified classification with only 4 forest classes namely close forest, open forest, bare land and grass land was used The overall classification accuracy obtained were 88.4%, 91.9% and 92.1% for the years 2000, 2003 and 2006 respectively. The Tasseled Cap green layer (GTC) composite of the three images was proposed to detect the change in vegetation of the study area. We found that the RBG-TCG worked better than RGBNDVI. For instance, the RBG-TCG detected some areas of changes that RGB-NDVI failed to detect them, moreover RBG-TCG displayed different changed areas with more strong colours. Change vector analysis (CVA) based on Tasseled Cap transformation (TCT) was also applied for detecting and characterizing land cover change. The results support the CVA approach to change detection. The calculated date to date change vectors contained useful information, both in their magnitude and their direction. A powerful tool for time series analysis is the principal components analysis (PCA). This method was tested for change detection in the study area by two ways: Multitemporal PCA and Selective PCA. Both methods found to offer the potential for monitoring forest change detection. A recently proposed approach, the multivariate alteration detection (MAD), in combination with a posterior maximum autocorrelation factor transformation (MAF) was used to demonstrate visualization of vegetation changes in the study area. The MAD transformation provides a way of combining different data types that found to be useful in change detection. Accuracy assessment is an important final step addressed in the study to evaluate the different change detection techniques. A quantitative accuracy assessment at level of change/no change pixels was performed to determine the threshold value with the highest accuracy. Among the various accuracy assessment methods presented the highest accuracy was obtained using the post-classification comparison based on supervised classification of each two time periods (2000 -2003 and 2003-2006), which were 90.6% and 87% consequently.

The emergence of the Internet has made the global information communications much easier than before. Users can navigate the desired information over the Internet by means of search engines. Even though search engine can help users search specified topic in a primary way, users usually cannot gain the overall idea of what the entire navigated results mean. In addition, information over the Internet keeps changing. Users cannot even keep track of the changes, let alone to comprehend the meanings of such changes. Consequently, this research proposes a two-stage incremental approach to figuring out the concept structure that represents the main concepts of the search results in the first stage, and keeping track of the concept changes with time based on spreading activation theory to assist users in the second stage. Experiments are conducted to examine the feasibility of our proposed approach. The first experiment is to evaluate the results from the first stage. It shows that the performance on recall and precision is quite satisfactory based on human experts¡¦ results. The second experiment is to examine the changing results from the entire proposed approach. It shows that high degree of agreement with our results is achieved from domain experts. Both experiments justify the feasibility of our proposed approach in real applications. That is, applying our proposed approach, users can easily focus on the topic they are interested in and learn its trend with great support. Keywords: Internet, Concepts Extraction, Concept Change Detection, Spreading Activation Theory.

To manage the information flood in the Internet, we usually navigate specific information using the provided search engines. Search engines are convenient but with limited functions. For example, it is impractical and impossible to browse through the entire collected information for us to gain an overall picture about what the navigated information stands for. To do so, we need an appropriate approach to automatically extracting concepts from the navigated information to assist users to easily and quickly gain the primary understanding toward a topic that interests users. In this research, we propose an approach to extracting concepts from the navigated web information and detecting the concept changes over time. It basically includes two stages. In the first stage, information is decomposed into paragraphs and they are clustered with key terms identified through the aid of latent semantic indexing method. Concepts are represented in the form of paragraph summary and associated key terms, which allows the user to easily comprehend what they describe. The second stage is to adaptively modify the concept structure to detect concept changes. With new information added, the concepts could be merging, splitting, or even emerging with time. Three experiments are conducted in this research to verify the proposed approach. Results of the first and second experiments show both high recall and high precision that matches the predefined concept categories. The last one is an illustrated real case application on the tsunami event. It shows that we can easily grasp different concepts of the tsunami reports and realize their changes by using our approach. The feasibility of employing our approach is thus justified.

Information acquisition over the Internet has become popular recently. Users, however, have difficulty in understanding the overall concept resulting from the searched information about a specific topic of their interests in the Internet. Moreover, such pieces of information keep changing over time. Therefore, in this thesis, an approach is proposed to help users further realize the searched results of their interested topic, and detect implications of the information changes over time. The first part of this approach is to gather information of a user-specified topic and analyze the overall meaning and the relations represented by those pieces of information. In this manner, users can gain the general concept of what the search results indicate. Here the keyword extraction approach, called RCBKE, is proposed to identify keywords with their relationships. Evaluations are performed and the results show that RCBKE can discover representative keywords. The second part is to track and investigate the information change of the topic in a certain time period. As a result, users can easily recognize the change patterns of the specified topic. An example to illustrate our approach is shown accordingly. The feasibility of our proposed approach is then justified.

Obtaining up-to-date information concernmg the environment at reasonable costs is a challenge faced by many institutions today. Satellite images meet both demands and thus present a very attractive source of information. The following thesis deals with the comparison of satellite images and a vector based land use data base of the City of Vienna. The satellite data is transformed using the spectral mixture analysis, which allows an investigation at a sub-pixel level. The results of the transformation are used to determine how suitable this spectral mixture analysis is to distinguish different land use classes in an urban area. In a next step the results of the spectral mixture analysis of two different images (recorded in 1986 and 1991) are used to undertake a change detection. The aim is to show those areas, where building activities have taken place. This information may aid the update of data bases, by limiting a detailed examination of an area to those areas, which show up as changes in the change detection. The proposed method is a fast and inexpensive way of analysing large areas and highlighting those areas where changes have taken place. lt is not limited to urban areas but may easily be adapted for different environments. (author's abstract) / Series: Research Reports of the Institute for Economic Geography and GIScience

Validation of global forest change detection databases Abstract The main aim of the thesis is to validate selected databases of changes in forest areas based on the analysis of satellite imagery time series in the Czech Republic. For this purpose we are using databases of M. C. Hansen and P. V. Potapov which are mapping the evolution of forest areas internationally. For the purposes of validation, we have proposed a methodology primarily based on historical ortophotographs from 2000-2012, the same time period which is documented in the validated databases. The results obtained were statistically processed, allowing to assess the accuracy of validated databases. At the end of the thesis, we are discussing the causes of identified inaccuracies and presented with recommendations for future improvements of detection of changes in forest areas. Keywords: validation, forest, land cover, change detection, Hansen, Potapov

Northwestern Arctic Canada is one of the most rapidly warming regions in the Arctic (Serreze et al., 2009). Retrogressive thaw slumps (RTS) are one of the most dramatic thermokarst features in permafrost terrain (Kokelj et al., 2013). Many studies have focused on describing the distribution of thermokarst landscapes (i.e., Olefeldt et al., 2016), as well as change in thermokarst terrain over the historical record (i.e., Kokelj and Jorgenson, 2013). However, improved high temporal and spatial resolution monitoring of thaw slump activity is required to enhance our understanding of factors governing their growth. Recent advances in aerial and ground-based Structure from Motion (SfM), a photogrammetry application, allow for temporal and spatial high-resolution characterization of landscape changes. This thesis explores two methods in SfM photogrammetry: 1) aerial imaging using an unmanned aerial vehicle (UAV) and 2) ground-based imaging using stationary multi-camera time-lapse installations, to derive high-resolution temporal and spatial data for change detection. A trend in mean elevation change was produced, and agrees with the RTS behaviour over the study period, which supports the viability of the proposed capture method. The lack of congruency in data range suggests need for further development in terms of analyses and differencing algorithms employed. The proposed method may be feasible for employment in other fields of science in which high temporal resolution change detection is desired. This proof of concept study was conducted at a small slump on the Peel Plateau, NWT, Canada, and aims to enhance understanding of the development and perpetuation of thaw slumps, to better anticipate landscape and ecosystem responses to future climate change.

Recent advances in computing technology allow for collecting vast amount of data that arrive continuously in the form of streams. Mining data streams is challenged by the speed and volume of the arriving data. Furthermore, the underlying distribution of the data changes over the time in unpredicted scenarios. To reduce the computational cost, data streams are often studied in forms of condensed representation, e.g., Probability Density Function (PDF). This thesis aims at developing an online density estimator that builds a model called KDE-Track for characterizing the dynamic density of the data streams. KDE-Track estimates the PDF of the stream at a set of resampling points and uses interpolation to estimate the density at any given point. To reduce the interpolation error and computational complexity, we introduce adaptive resampling where more/less resampling points are used in high/low curved regions of the PDF. The PDF values at the resampling points are updated online to provide up-to-date model of the data stream. Comparing with other existing online density estimators, KDE-Track is often more accurate (as reflected by smaller error values) and more computationally efficient (as reflected by shorter running time). The anytime available PDF estimated by KDE-Track can be applied for visualizing the dynamic density of data streams, outlier detection and change detection in data streams. In this thesis work, the first application is to visualize the taxi traffic volume in New York city. Utilizing KDE-Track allows for visualizing and monitoring the traffic flow on real time without extra overhead and provides insight analysis of the pick up demand that can be utilized by service providers to improve service availability. The second application is to detect outliers in data streams from sensor networks based on the estimated PDF. The method detects outliers accurately and outperforms baseline methods designed for detecting and cleaning outliers in sensor data. The third application is to detect changes in data streams. We propose a framework based on Principal Component Analysis (PCA) that reduces the problem of detecting changes in multidimensional data into the problem of detecting changes in the projected data on the principal components. We provide a theoretical analysis, which is support by experimental results to show that utilizing PCA reflects different types of changes in data streams on the projected data over one or more principal components. Our framework is accurate in detecting changes with low computational costs and scales well for high dimensional data.

This thesis concerns with change detection problematics in digital images captured under indoor conditions with an ordinary integrated camera in two consecutive moments. All challenges that accompany this problem will be discussed, starting with preprocessing and arriving to evaluation of the results. Currently used methods from this field are described and compared with each other such as differencing and LCP (Local Correlation Peak). A novel method, based on LTP descriptors, effectively solving this problem is proposed in this work. The proposed method is then tested on real data. The results of this tests are discussed subsequently. Besides the change detection method a method for parallax error minimization is proposed here.

With an increasing urban population in Sweden, expecting to reach 90% by 2050 (UN World Urbanization prospects, The 2011 Revision), this high level of urban population put pressure on functioning infrastructure, sufficient housing and need to monitor the environmental effects such as pollution and the effects of land use change. Stockholm County currently holds 22% of the population and accounts for nearly half of the urban growth in Sweden (Svensk Handelskammare).   Previous research on change detection using remote sensing cover the use of data sets from optical sensors, infrared spectrum, radar data and the use of additional derived data sets such as indices and texture measure (implemented on pixel or feature level). There is not yet any consensus regarding which change detection methods that is superior to others. Comparative studies often only test a few algorithms on one particular data set. Change detection of Stockholm urban area has not been well investigated in previous literature. This thesis is focused on a change detection analysis of Stockholm area between 1986 and 2006 using remote sensing data fusion. The data set used is SPOT-1 HRV XS data at 20m resolution from 1986, SPOT-1 HRV Panchromatic data at 10m resolution from 1987 and SPOT-5 HRG XS data of 10m resolution from 2006. The first challenge was to fuse the multispectral and panchromatic images from 1986 and 1987 to inject the details of the 10m panchromatic image into the 20m multispectral so that the resulting images will have similar spatial details as the 2006 images. This was done by wavelet transform. Haar, Daubechies, Coiflet and Biorthogonal wavelet families were tested to find the optimal fusion and the corresponding parameters. The results showed that the Daubechies, Coiflet and Biorthogonal families did not differ significantly and that for this data set and analysis purpose more than one wavelet family fusion results showed satisfactory results. The correlation coefficient for these three families was all over 0,96 at decomposition level two.   Then change detection was performed using change vector analysis (CVA) and a supervised non-parametric classifier. A comparison is made between two inputs: one using only spectral information and the other adding textural information to the spectral information. The change detection analysis was undertaken in three steps: calculating texture measures from the original images, calculating change magnitude using Change Vector Analysis (CVA) and classifying change from no-change using Support Vector Machine (SVM). Three GLCM texture measures were chosen: Homogeneity, Mean and Entropy in the change detection analysis. These, as well as the spectral information, were input for change vector magnitude. Then SVM is used to classify changed pixels from no-change pixels. Two change results were obtained, the first using only spectral information, and the other using both spectral and textural information. The overall accuracy using only spectral information was rather high at 87, 86%. But the visual inspections indicate that using only spectral change magnitude is not sufficient for a good change detection result because there is an apparent overestimation of change. When adding the textural information the overall accuracy increase drastically to 97,01%, although at visual inspection there seem to be an underestimation of change.  Because of the high overall accuracy an independent validation was made causing the overall accuracy and kappa to decrease. Change detection using only multispectral data got an overall accuracy of 76, 12% and kappa coefficient 0,53. For change detection result with added texture measures the overall accuracy became 85,80%  and 0,72.  The results further confirm the general advantages using texture measure although the independent evaluation resulted in a lower accuracy than the author's evaluations.