Abstract The Qilian Shan, at the northeastern frontier of the Tibetan Plateau, is a key area for studying the expansion mechanism of the Tibetan Plateau. Although previous thermochronology and paleomagnetic studies indicate Neogene northward expansion of the northern Qilian Shan, there is a distinct temporal gap in knowledge relative to the tectonic history of the southern Qilian Shan. This has hindered a complete understanding of the Cenozoic deformation pattern of the entire Qilian Shan. To study the growth history of the southern Qilian Shan, apatite fission track (AFT) data have been acquired from Zongwulong Shan and the Huaitoutala section. AFT thermal history modeling from the former shows a rapid cooling episode occurred at ~18–11 Ma, which is interpreted as marking the onset of intensive exhumation in the southern Qilian Shan. Within the Huaitoutala section, detrital grain up‐section shows progressively decreasing peak AFT ages followed by an age increase from midsection, implying that a sediment‐recycling event occurred at approximately 7 ± 2 Ma. Together with a shift in paleocurrent directions, this change marks the onset of Late Miocene deformation of the northern Qaidam Basin. Combined with previous studies on the deformation time of the Qilian Shan, our findings suggest that both the northern and southern Qilian Shan region grew outward synchronously in opposite directions during the Neogene. This resulted in the formation of a flower structure, which had an important impact on the deformation pattern of north Tibet. The synchronous outward expansion may have been triggered by the removal of mantle beneath north Tibet.
SUMMARY Modern geophysical data acquisition technology makes it possible to measure multiple geophysical properties with high spatial density over large areas with great efficiency. Instead of presenting these co-located multigeophysical data sets in separate maps, we take advantage of cluster analysis and its pattern exploration power to generate a cluster map with objectively integrated information. Each cluster in the resulting cluster map is characterized by multigeophysical properties and can be associated with certain geological attributes or rock types based on existing geological maps, field data and rock sample analysis. Such a cluster map is usually high in resolution and proven to be more helpful than single-attribute maps in terms of assisting geological mapping and interpretation. In this paper, we present the workflow and technical details of applying cluster analysis to multigeophysical data of a study area in the Trøndelag region in Mid-Norway. We address the importance of carefully designed pre-processing procedures regarding the input data sets to ensure an unbiased data integration using cluster analysis. Random forest as a supervised machine learning method for classification/regression is strategically employed post-clustering for quality evaluation of the results. The multigeophysical data used for this study include airborne magnetic, frequency electromagnetic and radiometric measurements, together with ground gravity measurements. Due to the nature of these input data, the resulting cluster map carries multidepth information. When associated with available geological information, the cluster map can help interpret not only bedrock outcrops but also rocks underneath the sediment cover.
Deep learning-based methods, especially deep convolutional neural network (CNN), have proven their powerfulness in hyperspectral image (HSI) classification. On the other hand, ensemble learning is a useful method for classification task. In this letter, in order to further improve the classification accuracy, the combination of CNN and random forest (RF) is proposed for HSI classification. The well-designed CNN is used as individual classifier to extract the discriminant features of HSI and RF randomly selects the extracted features and training samples to formulate a multiple classifier system. Furthermore, the learned weights of CNN are adopted to initialize other individual CNN. Experimental results with two hyperspectral data sets indicate that the proposed method provides competitive classification results compared with state-of-the-art methods.
An approach for pearl shape classification using fuzzy pattern recognition based on Zernike moments was proposed. After a series of preprocessing on the pearl image obtained, transformed it to the polar coordinates, then calculated eighth-derivative Zernike moments means as each type of typical shape 's characteristic parameter. The utilization of fuzzy pattern recognition achieved the shape effective distinction of pearl single image, finally, by seeking and comparison of characteristics image through multi-angles of view, realized the pearl shape classification. Experiment results indicated that, the method is of good distinction ability.
The recent interest in broadband seismic technology has driven us into searching for new and improved deghosting methods. The ghost effects in recorded towed streamer data are caused by the reflection of data back down from the sea surface. The interference between the up-going primary wavefields and the down-going ghosts creates notches in the recorded spectrum which reduce the useful bandwidth of the spectrum especially at increasing streamer depth. The purpose of this abstract is to propose a new approximative deghosting method for towed streamer data that works well for streamer depths down to around 15m. Compare to other proposals for deghosting, our method is local, which means the recorded pressure can be deghosted using only itself and its horizontal spatial derivatives at that receiver location together with the receiver’s depth. Numerical example we are showing in this abstract indicates a promising result in deghosting data between the first and second notch frequency.
Abstract Grid size has a significant influence on the computation efficiency and accuracy of finite-difference seismic modeling and can change the workload of reverse time migration (RTM) remarkably. This paper proposes a non-orthogonal analytical coordinate system, beam coordinate system (BCS), for the solution of seismic wave propagation and RTM. Starting with an optical Gaussian beam width equation, we expand the representation on vertically variable velocity media, which is the most common scenario in seismic exploration. The BCS based on this representation can be used to implement an irregular-grid increment finite-difference that improves the efficiency of RTM. Based on the Laplacian expression in Riemannian space, we derive the wave equation in the BCS. The new coordinate system can generate an irregular grid with increment increasing vertically along depth. Through paraxial ray tracing, it can be extended to non-analytical beam coordinate system (NBCS). Experiments for the RTM on the Marmousi model with the BCS demonstrate that the proposed method improves the efficiency about 52% while maintaining good image quality.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)
