Study on spectral characteristics of surrounding rock near 1. metal sulfide deposit
Figure 3.3. 1 is the reflection spectrum curve of several representative near-mine altered rocks and ore bodies in East Qinling. From the shape of spectral curve, it can be seen that the metal sulfide alteration zone has strong absorption characteristics in blue-green band (0.4 ~ 0.6 micron) and near infrared band (0.85 ~ 1. 1 micron, 2.2 ~ 2.4 micron). Strong reflection peaks appear in red (0.6 ~ 0.85 micron) and near red (1.28 ~ 1.46 micron) bands. The reflection spectrum curves of silicified altered rocks and hydroxyl-containing altered rocks are similar, and their absorption bands are still 0.4 ~ 0.55 micron, 0.85 ~ 1. 1 micron, 1.9 ~ 2.3 micron, but 0.6 ~ 0.85 micron and1.
Fig. 3.3. 1 reflection spectrum curve of near-ore altered rocks and mineralized bodies in East Qinling area
① Kaolinized altered rocks; ② Silicified altered rocks; ③ Metal sulfide altered rock
Fig. 3.3.2 shows the contrast curve of reflection spectrum between near-mine alteration zone and near-mine surrounding rock in East Qinling area. Compared with the spectral curve of metal sulfide belt, the spectral response tends to be flat, and there is no sharp jump peak or valley. From the shape of the spectral curve, it can be seen that the farther the metal sulfide zone (ore body) is (the mineralization and alteration are weakened), the more obvious the trend of the spectral curve is. The spectral shapes of andesite and marble are similar, except that there is a strong absorption band in the blue-green band of visible light (0.4 ~ 0.6 micron), and the curve from red light to near-infrared light is basically a smooth curve with the top of the arc upward. Although there are weak absorption at 1.4 μ m and 1.9μm, there are no obvious absorption valleys and reflection shoulders. Gneiss and granitoids are a smooth straight line, and there is basically no abnormal display. From the spectral comparison of different alteration types and different surrounding rocks near the mine, it can also be seen that the spectral curves are relatively dense in two bands: 0.4 ~ 1.4 micron and 1.9 ~ 2.5 micron, and the reflection difference is great only in 1.4 ~ 1.9 band.
Fig. 3.3.2 Reflectance spectrum curve of near-mine altered rocks and surrounding rocks in East Qinling area.
① phyllite; ② Gneiss; ③ Marble; ④ andesite; ⑤ Silicified altered rock
The above-mentioned characteristics of reflection spectrum curve show that there are obvious differences between the reflection spectra of altered rocks and non-altered rocks, and the dispersion of 1.4 ~ 1.9 micron band is the best, that is, Landsat-TM5 band should be the most basic band for extracting mineralization alteration information. 1.9 ~ 2.4μ m band also has a certain dispersion tendency, so Landsat-TM7 can be used as the auxiliary band. 0.4 ~ 0.6 micron band is a strong absorption band compared with 1.4 ~ 2.5 micron band, and Landsat-TM 1, 2 can be used as an ideal matching band.
Figure 3.3.3 shows the reflected wave curves of broad-leaved forest in different states. From the curve shape, it can be seen that with the aggravation of plant diseases, the absorption of 0.6 ~ 0.7 micron, 1.4 ~ 1.6 micron, 1.9 ~ 2.4 micron band gradually jumps, on the contrary, the reflection peak of 0.7 ~ 1.4 micron band decreases. From the point of dispersion and aggregation of curves, the band dispersion of 0.6 ~ 1.8 micron and 2.0 ~ 2.5 micron is the best. Therefore, Landsat-TM4 and the 7 bands between it can be used as the basic application band, and the 0.4 ~ 0.5 micron band is a relatively strong absorption band, among which Landst-TM 1 can be used as the basic matching band.
According to the principle of algebraic operation, when the reflectivity difference between bands is similar but the slope of the curve is different, the ratio of reflection band to absorption band can expand the spectral difference of ground objects to some extent and display the dynamic range. Table 3.3. 1 is a numerical table of different combination ratios calculated according to the reflectance spectrum data of ground objects. It can be seen from the table that Landsat-TM 5, 7 and 1 can be used as the basic application band, auxiliary band and matching band for information extraction of altered rock zone, which can show the spectral difference between altered rock and surrounding rock background characterized by gossan, silicification, kaolin and sericitization to the maximum extent. Need more areas.
Fig. 3.3.3 Reflectance spectrum curves of vegetation (Sambucus williamsii) in Shanggong Gold Mining Area of Henan Province in different states.
1-normal; 2- weak poisoning period; 3-toxic period
According to the types of alteration zones, the ratios of 7/ 1, 7/2, 5/ 1 and 5/2 can highlight the information of altered rocks dominated by Fe3+ to some extent. According to the experimental research, the mixed proportional processing [(tm3× tm4)-k]/tm7 greatly reduces the unaltered background information, and proportionally synthesizes tm5/1(r)+tm7/1(b)+[(tm3× tm4)-k]/tm7 (g).
Table 3.3. 1 List of ratios of characteristic bands of remote sensing mineralization in western Henan.
Table 3.3. 1 is the value of different combination ratios calculated according to the reflectance spectrum data of vegetation in different states. It can be seen from the numerical value of the combination ratio that as the extraction information of heavy metal ions poisoning vegetation, Landsat-TM4 and 7 bands have obtained better spectral differences. If the severely damaged area is highlighted, mixed band ratio processing, such as COSTM4×TM7-K, is needed. According to the experimental research, the ratio composite [cost M4× tm3-k] (r)+tm4/TM1(b)+tm7/TM1(g) image shows that the severely damaged area is orange, the slightly damaged area is close to white, and the normal area is close to cyan. You can also use TM4(R)+TM7(B)+[COSTM4×TM7-K](G) images. The normal area is red, the lightly damaged area is close to pink, and the severely damaged area is close to white.
2. Application principle of reflection spectrum characteristic mode.
Because most of the rock reflectance spectrum tests are carried out under indoor conditions, the rock reflectance recorded by remote sensing sensors is a reflection of natural conditions. Vegetation, soil, water and other factors in the natural environment often distort the reflection spectrum characteristics of rocks. For the near-ore alteration zone, the mineralization alteration information provided by remote sensing satellites is often related to the information of water, soil and vegetation in the mineralization alteration zone. When the mineralized alteration zone contains high toxic elements such as arsenic, mercury and lead and heavy metal elements, it can poison vegetation on a large scale. When the alteration zone is subjected to strong weathering and denudation, the surface residues are mostly siliceous and hydroxyl-containing clay mineral aggregates, and the scale of such weathering residues can far exceed the distribution range of the alteration zone. Therefore, in the process of extracting alteration information, we must consider the existence of interference factors, determine the extraction targets (vegetation poisoning, Fe3++ and OH-) according to the outcrop characteristics of locally mineralized altered rocks, and adopt different calculation formulas to remedy and screen according to the characteristic mode of reflection spectrum.
For example, the vegetation coverage rate in Qinling Mountains is about 70%, and the forest-rich areas can reach more than 90%. The main vegetation types in this area are mixed forests and shrubs of oak, Pinus tabulaeformis, Pinus armandii and Betula platyphylla. According to regional geochemical research, the nonferrous and precious metal gold deposits in Xiaoshan and Funiushan areas are often accompanied by harmful elements mainly arsenic and heavy metal elements mainly copper, lead, zinc and molybdenum, which will produce limited vegetation toxic halo within the radius of dispersed halo in the gold alteration zone. According to this feature, the northern slope of Funiu Mountain, with thin weathered residue and relatively poisoned vegetation, was selected as the experimental area, and the mixing treatment was carried out based on Landsat-TM3, 4, 5 and 7 bands sensitive to vegetation growth. The working procedure is shown in Figure 3.3.4.
Fig. 3.3.4 Program Diagram of Vegetation Poisoning Halo Extraction in Funiu Mountain Area of Henan Province.
The ratio of TM5/TM4 and TM4/TM3 is the best vegetation index. Their response to the degree of vegetation toxicity is an inverse sequence, that is, the reflectance of vegetation from normal development to toxic metamorphosis decreases at 0.36 ~ 0.61~1.20 (tm5/tm4) and 6.25 ~ 3.3 ~ 1.3 (tm4/tm3), respectively. TM5/TM 1 is sensitive to the reaction of iron hat, especially to the moderate or above toxicity of vegetation, and their ratio coefficients are 3.38 and 2.90, which are more than twice the background value. Vegetation poisoning area and alteration area are in high frequency domain, so low frequency background information can be compressed by filtering, and high frequency domain can be enhanced and highlighted by bright colors.
As shown in Figure I.1,the high value area of TM4/TM3(B) is located in the vegetation toxic altered rock area, and the bright yellow patches with nearly equal R and G components represent the information related to mineralization. It should be pointed out that this yellow dot cannot accurately indicate the location of mineralization and alteration. Due to the poor location caused by the migration of dispersed halo, this point is often located downstream of mineralized altered rocks.
In Xiong 'ershan-Waifangshan area, both mineralization and alteration develop along the structural fracture zone. In most cases, the structural fracture zone is negative topography, and most areas are filled with limonite and clay except for the silicified mineralized alteration zone that is intermittently exposed. The main ore-forming surrounding rocks are Archean Taihua Group granite-greenstone series and Proterozoic Xiong 'er Group andesite. Therefore, the information that needs to be enhanced is the hydroxyl-rich clay mineral aggregate developed in the tectonic belt.
According to the reflection spectrum data of ground objects, the targets and backgrounds in Landsat-TM3, 5 and 7 bands are well dispersed. If the alteration zone is further divided, it is necessary to process mixed pixels based on Landsat-TM3, 5 and 7 bands. The working procedure is shown in Figure 3.3.5 below:
Figure 3.3.5 Process Diagram of Halo Extraction of Hydroxyclay Minerals
According to the characteristic pattern of rock reflection spectrum, the results of TM7/TM 1 and TM5/TM4 ratio highlight the core of alteration halo zone-metal sulfide iron oxide cap, and the results of TM3/TM4 ratio highlight the information that vegetation is poisoned by heavy metals. Fourier transform transforms the ratio data into frequency domain, and then suppresses the low-frequency background by high filtering, so that the alteration information is highly enhanced; The information is restored to the spatial domain by inverse transform (IFT), and the threshold of color segmentation is determined by using the known altered region as a sample. Based on this threshold, we can do false color roaming or false color density segmentation, and get the reflector I.2..
The processing methods of compressed background mixed pixels proposed in the characteristic mode of reflection spectrum are [(tm3× tm4)-k] and [cos-tm4× tm7]-k, where the value of k represents the reflectivity (background value) of the surrounding rock in the mineralized alteration zone. The method of background compression is called "piecewise linear expansion" in image processing, that is, the whole dynamic range of image brightness value is divided into several segments, and each segment is expanded to different degrees (Figure 3.3.6).
In the figure, L 1 is the brightness value variable of the original image, and L2 is the brightness value variable of the transformed image. A 1, a2 and a3 are the selected segmentation breakpoints respectively. The slope between breakpoints controls the transformation of brightness values in the part.
In the figure, k 1, k2 and k3 are the slope of the transformation curve of the corresponding road section respectively. By properly selecting the breakpoint and slope, the contrast of the target image in a specific brightness value area can be enhanced or the contrast of some targets can be compressed.
Figure 3.3.6 Schematic diagram of piecewise linear expansion
Plate Ⅰ. 3 is an image obtained by image processing program, and its purpose is to extract the information of ore-guiding faults. In order to highlight the characteristic information of vegetation water abundance and change in different sections of the fault structural belt, TM4/TM3, TM5/TM 1 and TM5/TM2 were used for ratio processing respectively. The comparison value image data is scale to obtain a gray image with a gray level of 0-255. Then, the brightness value intervals of fracture feature information (192, 128, 1 15) are found as breakpoints in the ratio image, and the non-target area is compressed to 0, and the target area is given a larger expansion slope, and the expanded image is synthesized into a linear body with fracture structure significance, which is accurately displayed in bright tones. If the altered halo spot is superimposed on it, the ore-controlling significance of the fault will be more clear.