Global Vision Media Focus
The company was featured on CCTV a total of 15 times, and its reach through Party media, central media, and local official media reached a total of 50 million people.
Integrated Satellite Remote Sensing Mineral Exploration Applications in Toksun Area
Background Introduction
Since 1981, when Goetz et al. first identified limestone and clay minerals from space orbit using the 10-band SMIRR (Space Shuttle Multispectral Infrared Radiometer), remote sensing technology has revolutionized the scope, scale, content, and methods of human geoscientific research. With the successful development of a series of hyperspectral resolution remote sensing instruments (such as China's GF5 AHSI, the US's AVIRIS, and Italy's PRISMA) and the vigorous development of satellite technology, hyperspectral remote sensing technology has made significant progress in data acquisition, radiometric calibration, spectral reconstruction, data processing, and ground feature identification. More and more countries and regions are widely applying hyperspectral remote sensing technology to mineral spectral identification and mapping. This article will analyze hyperspectral remote sensing mineral exploration technology using a case study.
Data and Methods
1. Data
By analyzing the lithology and mineralization characteristics of the study area, multispectral data such as Landsat9 OLI and ASTER were used. In order to improve the spatial resolution of the data, Airbus CNES imagery was also selected.
（1）LandSAT 9 OLI numberaccording to
Landsat 9 OLI data includes multispectral bands (VNIR-SWIR, 30m spatial resolution) and one panchromatic band (15m spatial resolution). Remote sensing geological interpretation and mineralization/alteration information extraction based on principal component analysis were performed using the OLI data.
(2) ASTER data
ASTER, a remote sensor carried by the Terra satellite jointly launched by the United States and Japan in December 1999, has 14 bands in the visible/near-infrared (VNIR, spatial resolution 15m), shortwave infrared (SWIR, 30m), and thermal infrared (TIR, 90m) ranges, significantly improving spectral resolution. Compared to Landsat 9 OLI multispectral data, VNIR has one less band, but the SWIR range, suitable for studying the "diagnostic spectral absorption features" of altered minerals, has been expanded by 4 bands (1.60~2.43μm), bringing the total to 6 (Landsat data only has 2 bands). Various false-color composite combinations can be selected for interpretation based on the spectral characteristics of rocks and minerals.
(3) Hyperspectral data
The GF-5 hyperspectral imagery features 330 bands in the visible-shortwave infrared range (400-2500nm) with a spatial resolution of 30 meters. It is suitable for detailed lithological and mineral mapping. Before use, the hyperspectral imagery requires atmospheric correction, geometric correction, and RPC orthorectification.
2 Methods
(1) Principal component analysis
Principal Components Analysis (PCA) is a method for extracting alteration information from multispectral data. Its principle is to perform a linear transformation on the variables while maintaining the total amount of information, compressing multiple variables into a few independent variables. Each principal component has a different geological significance, and the amount of information decreases progressively with each principal component. This process reduces redundancy and overlap, making the data richer and more focused, better highlighting useful information, and improving the accuracy of information extraction.
(2) Spectral angle mapping method
Spectral angle mapping technology estimates the similarity between a test spectrum (pixel spectrum) and a reference spectrum (laboratory spectrum, etc.) by calculating the angle between them (Figure 1).
Figure 1. Schematic diagram of spectral angle matching
Assuming the image data has been converted to apparent reflectance after dark radiation or path radiation cancellation, the spectral dimension is equal to the number of bands. SAM calculates the similarity between the test spectrum *ti* and the reference spectrum *ri* using the following formula:
Where: nb equals the number of bands. The similarity between two spectra is not affected by vector length and gain, thus reducing the influence of topographic contrast. a ranges from 0 to 90°.
The SAM mineral mapping method can quickly identify all known minerals in a spectral library and display them using categorized color images, making it easy to observe the distribution range of minerals of interest directly with the naked eye.
Analysis of extraction results
The extracted results were segmented using a threshold method of "mean ± N * standard deviation". Specifically, for iron staining alteration anomalies, N was set to 1.5, 2, and 2.5; for mud alteration anomalies, N was set to 2, 2.5, and 3, classifying them into three levels: Level 1, Level 2, and Level 3. A 3×3 Gaussian low-pass filter was then applied to the graded alteration anomalies to eliminate salt-and-pepper noise. The final results of the extracted alteration anomaly information for the study area are shown in Figure 2.
Figure 2. Results of alteration information extraction.
Quartz, sericite, chlorite, and epidote were segmented using a threshold of "mean ± N * standard deviation". For non-ferrous alteration anomalies, N was set to 2, and the results were overlaid after being filtered by a 3×3 Gaussian low-pass filter to obtain alteration information extraction result 1, resulting in alteration information extraction result 2 for the study area (Figure 3).
Figure 3. Results of alteration information extraction in the study area.
Hyperspectral mapping
In addition to using multispectral mapping, this paper also uses GF5 data for hyperspectral mapping, which corroborates the results of multispectral mapping (Figure 4).
Figure 4. Hyperspectral mapping results
In summary, based on existing data and completed work, we obtained a comprehensive remote sensing alteration prospecting information map (Figure 3) and a hyperspectral mapping result map (Figure 4) for this region. The results show that the region has great prospecting potential and good source material and ore-guiding and ore-reservoir structural conditions.