(Aerogeophysical Remote Sensing Center, Ministry of Geology and Mineral Resources, Beijing)
Through image feature analysis and ground investigation, it is confirmed that there is a difference between copper mineralization alteration and surrounding rock in Lanping area of Lancang River in reflection spectrum characteristics, that is, altered rock has high reflectivity in TM5 (near infrared) band (1.55μ m ~ 1.75μ m). On this basis, the computer image processing technology for extracting TM remote sensing information related to copper mineralization and alteration is studied. Practice has proved that principal component analysis has the best effect in this respect. According to the hue theory of typical feature sample area in Lanping area image, the abnormal hue displayed in the image is compared with the geological exploration data of Hongtujian mining area. The geological significance of remote sensing anomalies is evaluated. Because of image preprocessing (geometric correction, brightness stretching, multivariate statistics, selection of the best band combination, etc. ) is carried out according to the characteristics of Lanping area of Lancang River, and its treatment method has certain applicability in the whole area. In this paper, the methods and techniques for extracting the remote sensing color tone anomaly information of copper mineralization alteration are introduced in detail, combining with the actual data such as Formula 2, Figure 4, Table 4, and color chart 1 1(5).
Key words: principal component analysis of remote sensing information of copper mineralization and alteration in Lanping area of Lancang River
First of all, the questions raised
With the gradual deepening of the application of space remote sensing data in geosciences, according to the spectral characteristics of hydrothermal alteration products of endogenetic metal deposits, remote sensing color anomaly information related to mineralization alteration is extracted through computer image processing of TM data to guide the exploration of metal deposits, which has aroused widespread concern at home and abroad [1][5]. 1in the spring of 1992, when Comrade Li participated in the field investigation of the potassium prospecting research project (No.:9 1-39) directly decided by the Ministry of Science and Technology, he became interested in the special geological and geomorphological features around some copper mines in Lanping-Yunlong area, Yunnan Province. For example, the attached color map of 1 1 was taken in the red soil tip area. The surrounding vegetation is sparse, and the residual slope deposits and soil are lighter, yellow and red than those in the peripheral areas, with large scale and obvious landscape characteristics, which can be observed by naked eyes. However, the mountain is high, the slope is steep, the valley is deep, the traffic is inconvenient, and the degree of geological work is low, so the idea of extracting remote sensing information related to copper mineralization alteration from TM data (which can be called copper mineralization alteration remote sensing information for short) came into being.
Therefore, according to the spectral curves of 22 rock samples in this area, the spectral characteristics of copper-mineralized altered rocks are studied. Select199065438+TM7 data collected on February 20th for image processing, select typical ground objects in the image, and analyze the reflection spectrum characteristics of typical ground objects. On this basis, through a lot of experiments, the method of extracting remote sensing information of copper mineralization and alteration by principal component analysis is discussed. The TM images of the 944km2 area from Xiaogela to Jinman in the western margin of Lanping Basin in Lancang River Basin and the sub-regions of Baofeng, Wenjing, Qiaohou, Shundangjing, Shijing and Hongtujian were processed by principal component analysis, and the abnormal images reflecting mineralization and alteration were obtained as reference, among which Hongtujian sub-region was confirmed by geological work of Team 8 14.
Comrade Li is very devoted to this work. He ignored 1993 for hemoptysis several times and insisted on field investigation. The ruthless cancer demon forced him to stop this fruitful research in the spring of 1994, and there was no time for field verification in other subregions except the red soil tip subregion. After Comrade Li's unfortunate death, the author took the top area of Hongtu as an example to organize the text according to a large number of records, images and unfinished reports left by him for communication and reference. Images obtained in other partitions cannot be published one by one, but they can be exchanged. Although the geological interpretation of the images in this area needs to be deepened, it will be a comfort to the dead if the very precious methods and experiences studied can be used for reference.
Second, the geological background
The experimental area is located in the area from Xiaogela to Jin Man on the western edge of Lanping Basin in Lancang River Basin. Structurally, this area belongs to the Sanjiang structural belt, and there is an ancient plate suture line-a long-term active Lancang deep fault, which runs through the north and south. The area experienced multi-stage tectonic movements in Variscan, Indosinian, Yanshan and Himalayan periods, which led to the development of folds, faults and magmatic rocks, and formed many sedimentary and endogenetic metal deposits [2]. Especially in the Middle and Late Triassic, large-scale neutral magma (andesite) erupted along the Lancang River fault, which is closely related to the formation of copper deposits in this area.
The discovered copper mineralization is mainly distributed in Shanglan Formation of Middle Triassic, Huakaizuo Formation of Middle Jurassic and Nanxin Formation of Lower Cretaceous. Copper ore bodies are veined, lenticular and layered, and most mineralization is accompanied by hydrothermal alteration such as silicification and iron dolomite. In this area, the mountains are high and the valleys are deep, the surface is severely cut, the bedrock is well exposed, the mineralized alteration zone is widely distributed, the residual slope deposits are widely distributed, and the vegetation is sparse, which creates certain conditions for remote sensing geological exploration.
The geological work in this area is low, and only one medium-sized copper deposit (Jin Man) has been evaluated at present. Therefore, it is of great significance to delimit the prospecting target area by using remote sensing information to narrow the exploration scope and speed up the exploration process.
III. Collection of rock and mineral samples and determination of reflection spectrum
In order to study the possibility and basis of extracting abnormal tone information from remote sensing of copper mineralization and alteration by using TM data, 22 rock samples were collected in this area, and the reflection spectra of these samples were measured by IRIS spectrometer, and the average reflectivity was calculated according to TM band. Table 1 lists the reflectance data of seven representative lithologic samples, and draws it as a graph 1.
Table 1 Statistical table of TM average reflectance of typical rocks and minerals in each band (unit:%)
sequential
Note: Average reflectivity; δ-variance
Figure 1 Spectral average reflectance curves of typical rocks and ores (see table 1 for curve numbers and sample lithology).
According to the characteristics of reflection spectrum, the reflection spectrum curves of rocks and minerals in this area can be roughly divided into three categories: the first category (spectral curve No. 1, 2,3) is the reflection spectrum curve of copper-bearing rock (copper ore), which is characterized by little change in TM 1-TM4 band, the highest in TM5 band, and a decrease of about 65438 in TM7 band. Many literatures [1] explain this phenomenon that hydroxyl (OH-) exists widely in the mineralized alteration zone, while OH- and OH- have strong absorption of electromagnetic waves in the TM7(2.08μm-2.35μm) band, so the TM7 brightness value in the mineralized alteration zone is low. The second is the reflection spectrum of igneous rock or sedimentary rock, which is characterized by low reflectivity and no obvious reflection peak, which is completely different from the reflection spectrum of mineralized rock samples. The third is the reflection spectrum curve of iron-dyed or silicified sandstone (curves No.4 and No.5 in Figure 1), which shows that the reflectivity gradually increases from TMI to TM5, and the TM7 band slightly decreases, and its TM 1 reflectivity is lower than the first one, which may be caused by the strong absorption and absorption of 0.45μm band electromagnetic wave by Fe ions.
The characteristics of reflection spectrum curves of the above three kinds of rocks and ores show that the reflection spectrum of copper-mineralized altered rocks is different from that of normal rocks, which is the basis for extracting copper-mineralized altered information and guiding the search for copper deposits by using remote sensing data.
Fourthly, TM image preprocessing.
In order to effectively carry out spectral analysis and extract prospecting information from representative sample areas on the image, it is necessary to carry out a series of preprocessing on the image, such as geometric correction, dynamic range stretching of brightness values, optimal combination selection of synthetic image bands, scale calculation and so on. Moreover, in order to facilitate the mosaic of Lanping-Yunlong region, the comparison of brightness values in each region and the repetition of some numerical operations, preprocessing is carried out according to the characteristics of the whole region.
The image of Lanping-Yunlong area (about 3072×4096 pixels, equivalent to 6×8 frames of 5 12 pixels× 512 pixels) is selected from TM data of 7 bands collected by Beijing Satellite Ground Station on February 20th. Taking topographic map as control, the image is geometrically corrected. Then, the minimum and maximum values of pixel brightness of each band in the whole image range are counted, and the brightness values of each band are linearly extended to 0 ~ 255 respectively; Then, the TM data of seven independent bands after geometric correction and expansion are formed into TM image data files of seven bands. Taking this as the source, the 1 Lanping seven-band basic image file (1024 pixels × 1024 pixels) is intercepted as the key research image.
In order to obtain a three-band color composite image with the largest amount of information, the smallest correlation between bands and the best display effect, the combination correlation factor Q in the formula (1) is taken as the scale and basis for selecting the best band combination, and the best band combination is selected by finding the maximum value of the combination correlation factor Q..
Zhang Yujun on new methods of geological exploration.
Where, Si is I-band variance or deviation, also called variation; Ri is the correlation coefficient between bands.
According to the formula (1), the best three-band combination of Lanping TM color composite image of Lancang River is TM5, TM4, TM3 or TM4, TM5 and TM7.
In this work, the research details are gradually improved and the enlarged images are intercepted step by step, and the Lanping sub-region image is intercepted from the Lanping regional image of Lancang River (color version11(5)); Intercept the drawstring region from the Lanping sub-region: then intercept the laterite creek sub-region from the image of the drawstring region (color version 1 1(4)). Its scale is also increasing step by step.
5. Characteristic analysis of brightness values of main terrain images.
From the 65,438+0 ∶ 200,000 color composite images of TM4(R), TM5(G) and TM7(B) in Lanping, 65,438+065,438+0 land-based image sample areas were selected (see attached drawings of color swatches 65,438+065,438+0 The brightness values of pixels in each sample area are averaged according to the brightness values of all pixels in the sample area, as shown in Figure 2.
As can be seen from Figure 2, the shapes of curves No.4, No.5 and No.6 are very similar to the reflection spectrum curves of the first type of rocks in Figure 1, that is, there is a reflection peak in TM5, while TM7 decreases slightly. The three curve sample areas all have different degrees of mineralization and alteration, which correspond to the surface rocks, but the morphology changes at TM4. For example, curve 6 sample area corresponds to the surface laterite copper mine point. And the surface is covered with certain vegetation (see 1 1( 1) in the lower left corner of the color plate drawing). Due to the "steep slope effect" of vegetation near infrared band, the brightness value of TM4 changes, which is higher than curves 4 and 5, while the brightness value of TM3 is lower than curve 5, which is caused by vegetation interference. In fig. 2, curve No.9 and curve 1 1 show the typical spectral characteristics of plant reflection. The sample areas of these two curves correspond to the lush vegetation region on the ground, but the brightness values of each band of the two curves are still different, and there may be some differences in vegetation types. In Figure 2, curves 10 and 3, the ground corresponding to the image sample area is marl, but they are located on the shady slope and sunny slope of the ground respectively, so the brightness values are different, but in
Fig. 2 TM brightness curves of several rock and vegetation samples in Lanping area (brightness values are stretched).
1- JBOY3 mudstone near baicun-yangcun; 2-K2 mudstone and siltstone near Songdeng; 3— J2 marl near Yangcun; 4- bare red bed, no vegetation cover; 5— Light-colored mineralized sandstone; 6— Copper mineralized altered dolomitic limestone with a small amount of vegetation in Hongtujian mining area; Ey red bed near ZK2; 8-T3 pyroclastic rocks near Potato Mountain; 9- lush vegetation coverage area; 10— J2 marl near Jixiang; 1 1- lush vegetation coverage area
There is a weak reflection peak at TM4. In fig. 2, curves 1 and 2 are characterized by high brightness, with TM3 as the reflection peak and TM4 as the absorption valley. The surface rocks corresponding to the curve sample area are mudstone and siltstone. In Figure 2, the brightness values of curves No.8 and No.7 are low and the curves are gentle. The surface rocks corresponding to these two curves are pyroclastic rocks and red salt-bearing strata. The above characteristics constitute the basis of extracting copper mineralization alteration information.
Six, copper mineralization alteration remote sensing information extraction
Although remote sensing technology has the advantages of strong macro-generalization ability, large coverage and repeated observation on a regular basis. However, it is very rare to directly indicate the occurrence of ore deposits or ore bodies. Firstly, the geological and metallogenic process is extremely complicated, and secondly, the spatial and spectral resolution of remote sensing technology is still limited. Therefore, the mineral information reflected by remote sensing data is often very weak, while the background geological information is very strong, so extracting mineral information has become the primary task and difficult problem of remote sensing geology, and trying to establish mathematical formulas (simple or complex) to extract mineralization information from remote sensing images by computer often fails. We use trial and error methods commonly used in image processing technology, such as ratio method, color coordinate system transformation, unsupervised classification, six-band or seven-band KL transformation (that is, principal component analysis) and so on. [4] Try to blast the remote sensing information of weak mineralization contained in TM data, and the effect is not good. Finally, four-band principal component analysis (TM 1-TM4-TM5-TM7 and TM 1-TM3-TM4-TM5) was successfully used to extract the remote sensing information of copper mineralization and alteration.
6. 1 KL transformation
Principal component analysis is realized by KL transform in image processing technology. As we all know, almost every multivariate analysis method usually needs to simplify complex problems, that is, to reduce the dimension of complex sets at the expense of some information, or to achieve the purpose of "seeing the forest from the tree" by transforming some secondary parameters [3]. Although the definition and operation of principal component analysis are strict in mathematics, the geological significance reflected by KL conversion results of TM data is very complicated. The top principal components reflect the widely distributed information of strata, lithology, structure, vegetation and other characteristics, while the principal components with large serial numbers reflect some macro and small information. It is found that these secondary information often contains mineralization and alteration information in several areas. Therefore, the principal component analysis method used in this paper is different from the conventional principal component analysis, which aims at compressing the dimension and highlighting the main information. Instead, it adopts the method of avoiding the main information and using the weaker secondary component information to explore its special geological significance. Therefore, we don't adopt the idea of "seeing the forest from the trees", but adopt the idea of "seeing the pests from the changes of leaves".
Table 2 Eigenvalues and Eigenvectors of KL Transform
The specific method is KL transform of the image. The linear correlation coefficient between each principal component and the original band pixel brightness value is the component of statistical feature vector, and the relative variation of each principal component is the statistical feature value. For two four-band images of 1024 pixel×1024 pixel in Lanping area of Lancang River, TM 1, TM4, TM5, TM7 and TM 1, TM3. The geological significance of TM 1, TM3, TM4 and TM5 transformation results is easier to clarify. Although the combination of TM 1, TM4, TM5 and TM7 contains more information from the perspective of optimal combination, there is no TM3 in this combination, which is of special significance for vegetation suppression. See Table 2 for the statistical results of KL transformation of pixel brightness values in TM 1, TM3, TM4 and TM5 in Lanping area of Lancang River. The information contents of the first to fourth principal components are 87%, 9.7%, 2.8% and 0.5% respectively.
6.2 abnormal image mapping
Because the purpose of this study is mainly to study the geological significance of those components whose serial numbers are large and the information amount takes a secondary position in the results of principal component analysis, the general abnormal information is mostly contained in the fourth (KL P4) and third (KL P3) components, and their linear functional relationship with the brightness value of TM band pixels is shown in Formula (2):
Zhang Yujun on new methods of geological exploration.
Therefore, an abnormal image is formed by color synthesis, that is, color synthesis is performed using KL P4(R), KL P3(G) and TM3/TM4(B). The attached chart 1 1(3) is an image of Lanping sub-region intercepted from the Lanping abnormal image of Lancang River. The significance of the ratio of TM3/TM4(R34) is to reduce the influence of topography and restrain the disturbance of vegetation. In the process of color synthesis, it is beneficial to set off the geological background by giving it blue. Although the information composition of KL P4 and KL P3 is clear, its geological significance is not intuitive. According to the color theory calculation of color synthesis of various image sample areas, the geological significance of various tones is judged, and the tone of abnormal images is qualitatively evaluated by combining with the field verification results.
6.3 Theoretical Calculation of Color Synthetic Tone in Typical Ground Objects Sample Area
According to Formula (2), calculate KL P4, klp3 and klp2 of pixel brightness values in four bands (TM 1, TM3, TM4 and TM5) in Lanping area (see Table 3).
According to the data of P4(R), P3(G) and R34(B) listed in Table 3, we can roughly estimate the hue that each region should present on the color composite image. Then mark the above-mentioned sample area with 1:200000 on the remote sensing color tone abnormal image of Lanping area (color plate attached figure1(3)). As expected, its color tone is basically the same as that theoretically estimated in Table 3, so it can be judged that the red and purplish red colors can be interpreted as abnormal areas of copper mineralization alteration on the map. Yellow and green tones mainly belong to vegetation and marl distribution areas, and other lithology shows white, blue and blue tones (Table 3).
Table 3 KL transform principal component values and TM3/TM4 list of TM images in each sample area
6.4 Preliminary geological verification of laterite top zoning
Cut the sub-region of red soil flow (1 1(3)) from the image in the color version, and enlarge it by 4 times to get the abnormal image of red soil flow (1 1(4)). In the geological sketch of 1 ∶ 50, the light-colored copper mineralization layers Ⅲ, Ⅳ and Ⅴ are as high as 6. 12%, and minerals such as malachite, chalcocite, chalcocite can be seen on the surface, which are disseminated, scattered, thin-film and veinlets distributed in light-colored timely sandstone particles and fractures, and the mineralization continuity in the ore-bearing layer is good. Length and thickness of light-colored copper mineralization layer in the study area. IV is greater than no. ⅲ, but the copper mineralization is poor, and the grade is low, only 0.02% ~ 0.04%, and a small amount of copper spots and malachite can be seen locally on the surface. The copper mineralized layer V is located in the interlayer fracture zone, with a length of 650m and a width of about 40m and a grade of 2.08% ~ 12.77%. The main copper mineral in the deep part is chalcocite, and the main minerals in the surface are malachite, azurite and black copper. The ore bodies are vein-shaped, pod-shaped, beaded and layered.
Table 4 Abnormal coincidence rate
Fig. 3 Geological survey of Lajing (Hongtujian) copper mine in Lanping County, Yunnan Province
The dotted box is the corresponding position of the remote sensing color tone anomaly image of copper mineralization alteration in the red soil peak area (color version attached figure 1 1(4)) (according to the data of team 8 14).
1- Paleocene Guolang Formation; 2- Eocene Yunlong Formation; Upper Cretaceous 3- Mankuanhe Formation; Lower Cretaceous 4- Nanxin Formation; 5- Upper Jurassic Bajulu Formation; 6- Middle Jurassic Huakaizuo Formation; 7— axis of anticline; 8— Normal fault; 9- Reverse fault; 10-Fault with unknown nature; 1 1- copper deposit; 12-light layer (copper mineralization)
Part of the light-colored copper mineralization layer I coincides with the light rose area on the abnormal image. The grade of the mineralized layer is low, ranging from 0.27% to 0.8%. The mineralization is discontinuous, and copper minerals such as malachite and azurite can be seen intermittently on the surface.
Light-colored mineralized layer II and part of mineralized layer I have no abnormal purple-red tone in the attached chart 1 1(4), which may be blocked by mountain shadows and need further study.
In addition, the northeast corner of the yellow box in the attached drawing 1 1(3) of the color plate has a triangular deep rose tone area, which corresponds to the geological map of the north of Rama Mountain (beyond the scope of Figure 3). Team 8 14 tracked two light-colored copper-bearing layers, 2400 meters long, 4 meters to 4.2 meters thick and with a grade of 0.34.
According to the contrast statistics between the abnormal distribution of hue and the area of copper mineralization layer, the coincidence rate between them is shown in Table 4.
As can be seen from Table 4, the coincidence rate between geological exploration results and remote sensing color anomalies in Hongtujian mining area is 89.3%. However, there is no statistical study on abnormal problems without mineralization. Obviously, the anomaly range is larger than the mineralization range. This problem is very complicated. At present, due to the failure to verify and study the anomalies one by one, the degree of coincidence between them needs further discussion. But generally speaking, prediction can't be as demanding as exploration, and it is expected that the prediction results will be completely accurate, just like the multiplicity of any geophysical information.
To sum up, it can be considered that it is feasible and effective to extract the remote sensing information of copper mineralization and alteration in Lanping area of Lancang River by principal component analysis.
Improvement of abnormal image of Lanping in Qiliangjiang River
In order to further reduce the correlation between P3 and P4 components of the first KL transform and "purify" the anomaly, the negative value of TM6 was taken as a negative value when making an abnormal improvement image of Lanping area of Lancang River with 1024 pixels×1024 pixels (color picture1(5)). That is, TM 1, TM3, TM4, TM5 and TM 1, TM4, TM5, TM7 are respectively subjected to KL transformation, and then two principal components (PP 1, PP2) are respectively selected from the obtained principal components for KL transformation. Select PP2 from the two principal components obtained by the double KL transform of TM 1, TM3, TM4 and TM5, and select PP 1 from the two principal components obtained by the double KL transform of TM 1, TM4, TM5 and TM7, and use TM6 to perform false color synthesis on R, B and G respectively to generate Lanping in Lancang River area.
Fig. 4 Flow chart of TM anomaly image processing in Lanping area of Lancang River
See Table 2, Table 5, Table 6 and Table 7 for principal component eigenvalues and eigenvectors obtained by dual KL transformation of two band combinations.
Table 5 Eigenvalues of TM1,TM3, TM4 and TM5 quadratic KL transforms
Table 6 Eigenvalues and Eigenvectors of KL Transform
Table 7 Eigenvalues of TM1,TM4, TM5 and TM7 quadratic KL transforms
The geological significance of the red tone on the swatch drawing 1 1(5) has been discussed in detail in sections 6.4 and 6.3. It is also the remote sensing information of copper mineralization and alteration. The yellow tone in the attached chart 1 1(5) is the composite tone in the high-value area after the negative value of PP:TM6, that is, the tone related to the low-temperature area, which mainly reflects the information of the shady slope of the mountain. Because there are few examples of abnormal copper mineralization on the shady slope of the mountain discussed in section 6.4, it is not completely certain that the yellow tone is also the display of copper mineralization alteration information. The attached figure 1 1(5) shows the tone of the green background (-TM6) on the color board, which sets off the general situation of the geological terrain. Due to color synthesis, the third and fourth components after KL transform of TM 1, TM4, TM5 and Tm7K are subjected to secondary KL transform to obtain the first principal component (PP 1).
From the above discussion, we are mainly interested in the abnormal red tone on the abnormal image. These red tone anomalies mainly reflect the alteration information of copper mineralization, and they are mainly distributed in feather clusters along the transverse faults of Lancang River. These anomalies directly show the size of copper mineralization alteration area, which is positively related to the scale of vein or seam exposure. The color version of the attached map 1 1(5) is also marked with some place name codes, which is convenient for comparison with geographical and geological maps. Because the picture is compressed and displayed, scattered anomalies are also lost, so it will be clearer to use four pieces of stitching or enlarged scanning. Except Jin Man (color code 1 1(5)) and Hongtujian (15), there are deposits in Xiangbicun (3), Kedengjian (4), Yantou (5) and Wendeng.
The geological research object is extremely complex and changeable, and the information of mineralization and alteration is relatively weak; Imagine using one or several mathematical methods (image processing technology) to calculate and get a complete and universal solution, which is beyond the current level of science and technology. However, we can still simplify the problem for some specific areas and find a set of suitable image processing techniques (mathematical tools) to extract rice relatively purely from weak mineralization and alteration information.
This paper has received valuable opinions from Comrade Ding Qun of our center. Zhang Xin from Yunnan Remote Sensing Station, Li Jinxing and Liu Jifu from 8 14 Geological Team participated in the interpretation and calculation of the photos, and we would like to express our gratitude.
refer to
[1] Liu Yanjun et al. Study on Multi-information Prediction of Hidden Orebodies in Dongping Gold Mine, Remote Sensing of Land and Resources, 1994, (1): 15 ~ 22.
Xiao et al. Mesozoic and Cenozoic Geology and Minerals in Yunnan. Beijing: Ocean Press, 1993.
[3] M. Kendall et al., Multivariate Analysis. Beijing: Science Press, 1983.
Zhang Yujun. Digital image processing of airborne magnetic data in central Qaidam Basin. UAS: Overview of exploration geophysics in Tulsa and China. American Exploration Geophysical Society,1989,517-535
[5]M.P.Ekstrom. Digital image processing technology. America: Academic Press, 1984
Study on extracting copper mineralization and alteration information in Lancang-Lanping area by principal component analysis of remote sensing data
Li Changguo, Zhang
(MGMR, Aviation Geophysics and Remote Sensing Center, Beijing, 100083)
There are spectral anomalies in Lanping area of Lancang River (near infrared TM 5 1.55 μm- 1.75 μm high reflection). It provides a scientific basis for the experimental study of extracting TM remote sensing information related to copper mineralization and alteration by image processing technology. Principal component analysis is the best method. Through the comparison with the geological work in Hongtujian area and the theoretical calculation of image sampling, the abnormal geological properties are evaluated. Due to image preprocessing (geometric restoration, brightness scaling, multivariate statistics, optimal selection of TM channels, etc.). ) evenly throughout the area. Therefore, it is reasonable to think that the obtained treatment technology is also applicable to the whole area. This paper describes it with detailed tables (4), formulas (2), graphs (4) and color images (5).
Keywords principal component analysis, Lanping area of Lancang River, remote sensing information of copper mineralization and alteration
Remote sensing of land and resources, 1997, 1.