Search

The automated generation of a digital elevation model over the Antarctic using stereo matching high-resolution satellite images is a challenging task. Moreover, the homogeneous radiometry in the icy environment and the strong geometric dissimilarity between stereo pairs over the steep terrain limit the use of area-based matching techniques. To overcome this issue, we propose template matching with image transformation in order to reduce the geometric dissimilarity. First, we generated epipolar resampled images to ensure the ease of estimation and handling of image dissimilarities from various viewing directions. We then utilized the normalized cross-correlation (NCC) and transformed the image patches within the matching window along the sample, line, and diagonal directions in order to improve the match rates within the steep areas. Furthermore, we tested the proposed method using Antarctic IKONOS stereo images and found that the overall matching success rate improved from 93.5% to 97.0% for all image pixels. We then computed the success rates over an area in which the NCC produced a low elevation point density and observed a more significant improvement from 58.7% to 79.26%. When compared to the manually generated elevation, the maximum vertical difference improved from 11.4 to 4.7 m. With these improvements, we can build a 1-m resolution elevation model over the glaciated high relief terrain. [ABSTRACT FROM PUBLISHER]

Change detection was one of the earliest and is also one of the most important applications of remote sensing technology. For multispectral images, an effective solution for the change detection problem is to exploit all the available spectral bands to detect the spectral changes. However, in practice, the temporal spectral variance makes it difficult to separate changes and nonchanges. In this paper, we propose a novel slow feature analysis (SFA) algorithm for change detection. Compared with changed pixels, the unchanged ones should be spectrally invariant and varying slowly across the multitemporal images. SFA extracts the most temporally invariant component from the multitemporal images to transform the data into a new feature space. In this feature space, the differences in the unchanged pixels are suppressed so that the changed pixels can be better separated. Three SFA change detection approaches, comprising unsupervised SFA, supervised SFA, and iterative SFA, are constructed. Experiments on two groups of real Enhanced Thematic Mapper data sets show that our proposed method performs better in detecting changes than the other state-of-the-art change detection methods. [ABSTRACT FROM AUTHOR]