Silicified burrows and stylolite seams are common in the chalky pelagic limestones of the Kometan Formation (late Turonian-early Campanian) of the Dokan area in northeastern Iraq. The shape of individual and packed network branched chert nodules, which are the result of the early diagenetic mobilization of radolarian silica, are controlled chiefly by pre-existing Thalassinoides burrow system. Frequent appearance of Thalassinoides may reflect periodical formation of discontinuity (omission) surfaces on the canyon margin (or near the listric fault) of the Dokan area resulted from combination of tectonic and eustatic processes. It is also proven that many so-called bedding planes of the Kometan Formation are merely pressure solution stylolites in massive homogeneous pelagic limestones (pseudo-bedding). It is concluded that the association of silicified burrows (chert nodules) with stylolite veins refers to regressive discontinuity surfaces, while the separation of these two features indicates either winnowed permeable surfaces or true bedding planes. The clay constituent of such stylolites is found to be formed mainly by transformation of smectite during burial processes.

Hyperspectral image processing has been a very dynamic area in remote sensing\nand other applications in recent years. Hyperspectral images provide ample spectral\ninformation to identify and distinguish spectrally similar materials for more accurate\nand detailed information extraction. Wide range of advanced classification techniques\nare available based on spectral information and spatial information. To improve\nclassification accuracy it is essential to identify and reduce uncertainties in image\nprocessing chain. This paper presents the current practices, problems and prospects of\nhyperspectral image classification. In addition, some important issues affecting\nclassification performance are discussed.