J.M. Daida

Observation of sea-ice dynamics using synthetic aperture radar images: automated analysis

Two techniques for automated sea-ice tracking, image pyramid area correlation (hierarchical correlation) and feature tracking, are described. Each technique is applied to a pair of Seasat SAR sea-ice images. The results compare well with each other and with manually tracked estimates of the ice velocity. The advantages and disadvantages of these automated methods are pointed out. Using these ice velocity field estimates it is possible to construct one sea-ice image from the other member of the pair. Comparing the reconstructed image with the observed image, errors in the estimated velocity field can be recognized and a useful probable error display created automatically to accompany ice velocity estimates. It is suggested that this error display may be useful in segmenting the sea ice observed into regions that move as rigid plates of significant ice velocity shear and distortion. >

Object-oriented Feature-tracking Algorithms For SAR Image Of The Marginal Ice Zone

An unsupervised method that chooses and applies the most appropriate tracking algorithm from among different sea-ice tracking algorithms is reported. In contrast to current unsupervised methods, this method chooses and applies an algorithm by partially examining a sequential image pair to draw inferences about what was examined. Based on these inferences the reported method subsequently chooses which algorithm to apply to specific areas of the image pair where that algorithm should work best.

Mathematical Morphology for Automated Analysis of Remotely Sensed Objects in Radar Images

Previous studies have established that mathematical morphology may meaningfully transform remotely sensed images, provided that the desired objects to be extracted are well-defined both in terms of apriori information and image representation. However, ill-defined, fractal, objects are common in remote sensing. Naturally-occurring fractal objects-such as lakes, islands or ice floes- exhibit a diversity of shape and orientation that hinder characterization. Speckle, common to imaging radar, also hinders characterization. The reported method addresses the problem of ill-defined objects by considering symbiotic forms of segmentation and morphological transformation. Traditionally, segmentation decomposes an image into objects and background, while morphological transformations operate on the resulting binary image to extract needed information. The reported method incorporates morphological transformation interdependent upon pyramidal segmentation. Results indicate that the reported method performs with low misclassification-error rates (1.8% for a 4-look benchmark) and correct

Object-Based Feature-Tracking Algorithms for Sar Images of the Marginal Ice Zone

Synthetic Aperture Radar (SAR) provides an excellent means of observing the movement and distortion of sea ice over large temporal and spatial scales. Consequently, the European Space Agencys ERS-1 satellite will carry a SAR over the polar regions in late 1990. A key component in using arctic SAR data is an automated scheme for extracting sea-ice displacement fields from a sequence of SAR images of the same geographical region. Although automatic sea-ice tracking algorithms do exist, analyzing the Marginal Ice Zone remains challenging. The wide variety of ice movements and diverse of seascapes have led to the development of hybrid schemes. An important element of these schemes consists of a feature-based algorithm. Another important element usually consists of a statistically-based algorithm, e.g. correlation. Computer understanding of what features to look for, when to apply correlation and where to look is still subject to investigation. Our research focuses on a method for providing computer understanding of a Marginal Ice Zone scene. The method parses a complex ice scene into individual objects. We define an object as a closed boundary and its interior. Consequently, we highlight in this paper how the recognition of objects facilitates sea-ice tracking. We also describe our solution to the problem of weakly co�ected regi?ns as one facet in automatically creating objects from bItmaps. We provide an example of matching floes taken from a synoptic image pair. The example employs a simple object construct using invariant moments, although object contructs are not limited to them.

Automated Extraction Of Pressure Ridges From Sar Images Of Sea Ice: Comparison With Surface Truth

The authors estimate the characteristics of ridges and leads in sea ice from SAR (synthetic aperture radar) images. Such estimates are based on the hypothesis that bright filamentary features in SAR sea ice images correspond with pressure ridges. A data set collected in the Greenland Sea in 1987 allows this hypothesis to be evaluated for X-band SAR images. A preliminary analysis of data collected from SAR images and ice elevation (from a laser altimeter) is presented. It is found that SAR image brightness and ice elevation are clearly related. However, the correlation, using the data and techniques applied, is not strong.

Measurement of Pressure Ridges in Sar Images of Sea Ice: Preliminary Results on Scattering Theory

Sea ice ridges and keels (hummocks and bummocks) are important in sea ice research for both scientific and practical reasons. A long-term objective is to make quantitative measurements of sea ice ridges using synthetic aperture radar (SAR) images. The preliminary results of a scattering model for sea ice ridge are reported. The approach is through the ridge height variance spectrum Psi(K), where K is the spatial wavenumber, and the two-scale scattering model. The height spectrum model is constructed to mimic height statistics observed with an airborne optical laser. The spectrum model is used to drive a two-scale scattering model. Model results for ridges observed at C- and X-band yield normalized radar cross sections that are 10 to 15 dB larger than the observed cross sections of multiyear ice over the range of angles of incidence from 10 to 70 deg.