A Study On Characteristics Of Remote Sensing Images Of Eruptive Products From The Tianchi Volcano In The Changbai Mountains

The terrain of the Changbaishan Tianchi volcano on the border between china and North Korea is complex, and traffic is not very convenient. The two third of the Tianchi Lake and one fourth of the cone of the Tianchi volcano belong to the Democratic Republic of Korea. In the present conditions, it's hard to study the spatial distribution of the volcanic eruptive products by conventional means from a large regional scope. In the Changbaishan Tianchi volcano, the predecessors are continuously carrying on the research work of the comprehensive geological investigation and petrology, and they have already obtained many research results in the determination of volcanic effusive stages and the volcanic eruptive products lithology and the boundary, but the studies in the spatial distribution of the eruptive products have been a weak link. The remote sensing technology can gain the wide range data with little limit, so it might help solve this problem. But the vegetation of the Changbaishan mountains grows well, which affects the quality of the remote sensing images, causes the difficulty to use the visible spectral remote sensing image to distinguish spatial distribution of the volcanic eruptive products. Therefore, by the technology of DEM terrain analysis, this thesis studies the geomorphologic characteristics of the Changbaishan Tianchi volcanic eruptive products, then, by the technology of remote sensing spectral analysis and texture analysis, studies the spectral and texture characteristics of the Changbaishan Tianchi volcanic eruption on the ASTER and IKONOS images. Finally, this work uses these image characteristics to speculate the spatial distribution of the volcanic eruptive products where the predecessors did not work. These results can provide a remote sensing basis for making the large-scale volcanic geological maps of the Tianchi volcano, in the Changbai Mountains, and especially for determining the scope and limits of some Holocene volcanic eruptive products.Making use of the terrain analysis technology of DEM, this work can gets the three-dimensional topographical map of the vicinity of the Changbaishan Tianchi volcanic cone, and through the interpretation of the study area it gains the distribution of the crater and volcanic cone and the distribution of the water systems and valleys. Through the landscape feature analysis of the rough areas of Pleistocene and Holocene rock eruption of the Tianchi volcano cone, this work comes to the preliminary conclusion: at the different stages of the Pleistocene , the Tianchi volcano cone had a ridge-like distribution of the rough rock, and the distribution of the four stages of the eruption had obvious levels, and often the cone eruption of a later stage accumulated on the previous stage of eruption. Generally speaking, the earlier eruption was located far away from the Burner and the ease of its slope, and the lower elevation. From the distributing character of the Chinese-made cone rough rocks mentioned above, this thesis speculates the Korean-made cone eruption of distribution in the study area. Tianchi volcano eruption of the Holocene has significant characteristics of the topography, such as the weather station of the base crumbled into liquid lava has a long ridge-like distribution, the overall flow unit features obviously; the Holocene landslide accumulation near the cone whose distribution mainly controlled by the terrain gradient; volcanic debris flow mainly confined to low-lying valley, of which flows have a significant performance in the 3-D perspective generated by the DEM . Therefore, using the DEM topography data to analyze the physiognomy is useful to determine the distribution of prominent morphology of the eruption, such as several cones produced in the Pleistocene and the eruption of new-style mud-rock flow.Processing the remote sensing image, inversing the radiation rate of Aster image by the formula, there are many ways to inverse the rate of reflection, but no very acceptable methodology. So this work takes advantage of the Log Residuals inversion and the FLAASH inversion two methods module ,and using the K-Means of unsupervised classification to class the above three inversion of the image. The inspection found that the classification accuracy rate of two reflectivity images is worse than the radiation images, so the radiance image is used in the spectral analysis. By the spectral analysis the characteristic curve of the rough and pumice spectra is obtained. Due to a high degree of vegetation coverage, only the range of the bare rock and the rough rocks in the eruption is identified, but the indirect method can be used to determine the eruption range, like the concave wall had already air slaked into the soil of the modern eruption deposits, so the distribution of the tundra vegetation is namely the distribution scope of the accumulation of the modern eruptions. At the same time researching the rough rock and pumice on the slope, the characteristics of the upward slope , it found out the reasons that a bare rock has dark and bright two tones. On the one hand the former was likely to face the sun, the latter mainly in the back-face, causing the different Spectrums of the same. On the other hand, probably because most of the former are located to the southeast of Burner , and the high degree of coverage of floating rock in the direction, may be the mixture of rough rocks and a small amount of pumice.Since the purchase of 1 m of the IKONOS images are true color, contrasting the field photographs with the distribution of vegetation of the Changbaishan zone, establishing a preliminary interpretation symbol, which selected eight types of training samples are selected and the spectral angle technology (SAM) is used to get the classification. The result of classification after confusion matrix evaluation is not satisfactory, maybe a serious loss of spectrum information in the synthesizing process, but the texture information is enough. This work introduces the texture analysis of the ENVI probability matrix method taking advantage of dissimilarity texture feature pictures to establish the symbol of the texture interpretation and interpret the distributing area of formation the modern Burner volcanic debris flow. Using the combination of spectrum and texture to do the classification, this work enhances the accuracy of classification and determines the distributing of the pumice and a distribution of its thickness, and terrain characters of different thickness pumices.The terrain of the Changbaishan Tianchi volcano is complex and highly vegetation-covered, it is often difficult to determine the precise spatial distribution region of the volcanic eruptive products only using the geological or visible remote sensing technology. The terrain analysis with DEM will be a certain extent compensate for this deficiency. Using high spatial resolution IKONOS images and multi-spectral imaging ASTER and combined with DEM analysis technology, this work obtains the terrain features and spectral characteristics and texture characteristics of eruptive products since the Pleistocen, and provides the support to determine the boundary of the spatial distribution of the volcanic eruptive products. In the future work, it is needed to further strengthen the field investigation and inspection of the terrain features of different types of eruptive products, so that it will truly become one of the tools which can be used to make a volcano geological map and evaluate volcanic hazards evaluation.