基于多尺度纹理分类及矿物识别的ASTER地质填图

唐淑兰

doi:10.19509/j.cnki.dzkq.2022.0077

基于多尺度纹理分类及矿物识别的ASTER地质填图

doi: 10.19509/j.cnki.dzkq.2022.0077

唐淑兰

基金项目:

中国地质调查局项目 DD20190364

中国地质调查局项目 202009000000180703

西安财经大学科学研究扶持计划项目 21FCJH008

详细信息

作者简介:
唐淑兰(1979—), 女, 讲师, 主要从事遥感影像模糊识别研究。E-mail: mzwsbjh@126.com

中图分类号: P627
计量
- 文章访问数: 758
- PDF下载量: 52
- 被引次数: 0
出版历程
- 收稿日期: 2021-12-11

Aster geological mapping based on multi-scale texture classification and mineral recognition

Tang Shulan

摘要

摘要:
遥感技术已经成为基础地质调查必不可少的手段。为提高地质填图效率及精度, 本研究提出了基于ASTER的岩性自动分类加主要和典型造岩矿物识别的填图方法。首先, 对ASTER数据进行主成分变换, 对选取的第一主分量采用Haar小波进行多尺度小波分解, 将小波系数的统计特征作为纹理特征, 构建纹理及光谱多维特征空间; 接着, 运用支持向量机(SVM)进行岩性分类; 同时, 根据光谱特征提取主要造岩矿物; 最后将主要造岩矿物叠加在分类结果上, 结合野外调查背景进行岩性填图。混淆矩阵结果显示光谱+小波纹理分类精度可以达到83.496 2%, 较光谱+灰度共生矩阵纹理分类精度提高了2.675 6%, 较光谱特征分类精度提高了6.318 9%。与最大似然法(MLC)分类相比, SVM分类精度提高了6.623 7%。矿物提取结果表明, 构造的提取指数可有效提取白云母、黑云母、方解石、角闪石等矿物。野外工作证明, 填图结果与野外调查结果的相关系数为0.7, 可见, 基于ASTER数据利用图像处理技术、机器学习算法及波段运算可作为植被覆盖较少地区有效的地质填图手段。
- 地质填图 /
- 小波变换 /
- 多尺度纹理 /
- 支持向量机 /
- 矿物识别 /
- ASTER
Abstract:
Remote sensing technology has become an indispensable means in geological survey. In order to improve the efficiency and accuracy of geological mapping, a method based on Aster automatic lithology classification combined with the identification of main rock forming minerals is proposed in this study. Firstly, the principal component transform of ASTER data is carried out, the first principal component is selected for multi-scale Haar wavelet decomposition, and the statistical characteristics of wavelet coefficients are taken as texture features to construct multi-dimensional feature space of texture and spectrum; Then, support vector machine is adopted to classify lithology; At the same time, the main rock forming minerals are extracted according to the spectral characteristics; Finally, the main rock forming minerals are superimposed on the classification results, and the lithology mapping is completed in combination with the field investigation background. The confusion matrix results show that the classification accuracy of spectrum- wavelet texture can reach 83.496 2%, which is 2.675 6% higher than that of spectrum-gray level co-occurrence matrix texture classification and 6.3189% higher than that of spectral feature classification. Compared with the maximum likelihood classification method, the classification accuracy of SVM is improved by 6.623 7%. The mineral extraction results indicate that the extraction index of structure can effectively extract muscovite, biotite, calcite, amphibole and other minerals. It can be seen that image processing technology, machine learning algorithm and band operation can be used as effective means of remote sensing mapping in areas with less vegetation coverage.
- geological mapping /
- wavelet transform /
- multi-scale texture /
- support vector machine /
- mineral recognition /
- ASTER

HTML全文

图 1 尾亚地区地质图

Figure 1. Sketch geological map of Weiya area

下载: 全尺寸图片幻灯片

图 2 本研究填图工作流程

Figure 2. Flow chart of geological mapping

下载: 全尺寸图片幻灯片

图 3 小波纹理特征

Figure 3. Wavelet texture features

下载: 全尺寸图片幻灯片

图 4 不同移动窗口的小波纹理分类精度

Figure 4. Wavelet texture classification accuracy of different moving windows

下载: 全尺寸图片幻灯片

图 5 各分解级数的小波纹理分类精度

Figure 5. Wavelet texture classification accuracy of each decomposition series

下载: 全尺寸图片幻灯片

图 6 不同特征组合分类结果

Figure 6. Classification results of different feature combinations

下载: 全尺寸图片幻灯片

图 7 方解石提取结果

Figure 7. Calcite extraction results

下载: 全尺寸图片幻灯片

图 8 矿物提取结果

Figure 8. Extraction results for various minerals

下载: 全尺寸图片幻灯片

图 9 遥感填图结果

Q.第四系; C₁.下石炭统; Pt₂.中元古界; Pt₁.下元古界; T₂ξ.中三叠石正长岩; T₂ηγ.中三叠世二长花岗岩; T₂ξο.中三叠世石英正长岩; T₂γδ .中三叠世花岗闪长岩; T₂υ.中三叠世辉长岩; C₂ηγ.晚石炭世二长花岗岩; C₂γδ.晚石炭世花岗闪长岩; Pt₃gn.新元古代片麻岩

Figure 9. Remote sensing mapping results

下载: 全尺寸图片幻灯片

表 1 矿物的吸收谱带与ASTER波段的对应关系

Table 1. Correspondence relation between absorption bands of minerals and ASTER bands

矿物	ASTER波段
矿物	1	2	3	4	5	6	7	8	9	10	11	12	13	14
黑云母											强吸	高反	弱反	高反
石英										强吸	高反	强吸
方解石												弱吸	强吸	高反
正长石										弱反	强吸	高反
角闪石						弱反		弱吸		强吸		强反
辉石	高反		强吸
白云母					高反	强吸	高反

下载: 导出CSV

表 2 矿物提取指数

Table 2. Extraction index of minerals

矿物	矿物提取指数
黑云母	[b12/(b11+b13)]×(b14/b11)
石英	[b11/(b10+b12)]×(b13/b12)
方解石	b14/(b13+b12)
正长石	(b12+b10)/b11
角闪石	(b6/b8)×(b12/b10)
辉石	b1/b3
白云母、绢云母	(b5+b7)/b6

下载: 导出CSV

表 3 岩性分类所用样本数

Table 3. Sample number for lithological classification

岩性	Q	C₁	Pt₂	Pt₁	T₂ξ	T₂ ηγ	T₂γδ	T₂ξo	T₂υ	C₂ηγ	C₂γδ	Pt₃ gn
训练样本	1 706	1 737	1 469	951	1 019	1 564	1 469	316	795	1 844	3 639	770
检验样本	2 460	2 239	2 625	2 248	1 020	1 755	2 191	1 724	439	2 007	1 148	812
注：岩性同图 1

下载: 导出CSV

表 4 不同特征的分类精度

Table 4. Classification accuracy of different features

分类组合	特征维数	总体精度/%	Kappa系数
SF	9	77.177 3	0.747 7
G-TF	8	61.825 0	0.577 4
W-TF	8	62.037 9	0.580 1
SF-G-TF	17	80.820 6	0.787 8
SF-W-TF	17	83.496 2	0.830 7
注：SF.光谱；G-TF.GLCM纹理；W-TF.小波纹理；SF-G-TF.光谱+GLCM纹理；SF-W-TF.光谱+小波纹理；下同

下载: 导出CSV

表 5 不同分类算法的精度比较

Table 5. Accuracy evaluation of different classification algorithms

特征	SVM法		MLC法
特征	O_A/%	Kappa系数	O_A/%	Kappa系数
SF	77.177 3	0.747 7	75.248 1	0.720 7
G-TF	61.825 0	0.577 4	56.023 8	0.516 1
W-TF	62.037 9	0.580 1	59.067 2	0.547 8
SF-G-TF	80.820 6	0.787 8	76.035 4	0.735 2
SF-W-TF	83.496 2	0.830 7	76.872 5	0.744 2

下载: 导出CSV

表 6 不同岩性的分类精度

Table 6. Classification accuracy of different lithologies

岩性	SF		G-TF		W-TF		SF-G-TF		SF-W-TF
岩性	P_A/%	U_A/%	P_A/%	U_A/%	P_A/%	U_A/%	P_A/%	U_A/%	P_A/%	U_A/%
Q	78.01	65.92	81.95	69.04	78.86	73.07	88.46	76.86	88.98	83.84
C₁	97.14	92.63	97.23	80.66	87.05	76.07	97.28	93.20	96.69	97.17
Pt₂	91.73	85.91	75.43	57.42	68.91	72.56	94.10	86.03	94.93	80.88
Pt₁	74.02	85.82	19.17	47.68	58.76	51.70	73.13	84.22	83.67	81.75
T₂ξ	84.02	67.16	92.94	66.39	89.12	62.18	94.41	76.25	93.43	69.97
T₂ηγ	66.04	63.58	62.22	54.22	70.26	50.47	73.39	63.08	73.68	71.20
T₂γδ	95.42	94.81	87.53	96.05	91.71	81.41	95.30	97.10	94.84	92.37
T₂ξo	52.85	76.07	43.96	80.75	54.21	73.68	51.03	73.44	55.13	97.19
T₂υ	68.69	96.23	40.76	74.73	48.74	57.42	73.71	96.02	76.49	92.34
C₂ηγ	58.10	78.78	38.57	37.96	8.32	26.63	60.24	80.49	63.38	92.31
C₂γδ	74.74	46.91	66.55	38.31	47.91	50.60	77.7	59.71	85.89	68.14
Pt₃gn	44.70	55.42	0.12	0.46	7.02	8.65	49.51	57.68	58.25	77.54
注：岩性代号说明同图 9; P_A.生产者精度；U_A.用户精度

下载: 导出CSV

参考文献(17)

[1]	Hosseinjani Z M, Tangestani M H, Roldan F V, et al. Sub-pixel mineral mapping of a porphyry copper belt using EO-1 hyperion data[J]. Advances in Space Research, 2014, 53(3): 440-451. doi: 10.1016/j.asr.2013.11.029
[2]	Hisham G, Habes G. Detection of gossan zones in Aridregions using Landsat 8 OLI data: Implication for mineral exploration in the eastern Arabian shield, Saudi Arabia[J]. Natural Resources Research, 2018, 27(1): 109-124. doi: 10.1007/s11053-017-9341-8
[3]	崔静月, 董玉森, 岳文丽, 等. 基于遥感蚀变信息提取巴林格撞击坑周边铁陨石[J]. 地质科技通报, 2021, 40(1): 209-216. https://www.cnki.com.cn/Article/CJFDTOTAL-DZKQ202101023.htm Cui J Y, Dong Y S, Yue W L, et al. Extraction of iron meteorites from the Barringer Meteor Crater based on remote sensing alteration information[J]. Bulletin of Geological Science and Technology, 2021, 40(1): 209-216(in Chinese with English abstract). https://www.cnki.com.cn/Article/CJFDTOTAL-DZKQ202101023.htm
[4]	王俊虎, 武鼎, 张杰林, 等. 基于多源遥感数据的纳米比亚欢乐谷地区千岁兰断裂带识别及新发现[J]. 地质科技通报, 2020, 39(5): 183-190. https://www.cnki.com.cn/Article/CJFDTOTAL-DZKQ202005023.htm Wang J H, Wu D, Zhang J L, et al. Identification and new discovery of Qiansxuilan fault belt in Gaudeanmus area, Namibia based in the multi-source remote sensing data[J]. Bulletin of Geological Science and Technology, 2020, 39(5): 183-190(in Chinese with English abstract). https://www.cnki.com.cn/Article/CJFDTOTAL-DZKQ202005023.htm
[5]	Wang R, Lin J Y, Zhao B, et al. Integrated approach for lithological classification using ASTER imagery in a shallowly covered region: The eastern Yanshan Mountain of China[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2018, 11(12): 4791-4807. doi: 10.1109/JSTARS.2018.2879493
[6]	Aboelkair H, Ninomiya Y, Watanabe Y, et al. Processing and interpretation of ASTER TIR data for mapping of rare-metal-enrichedalbite granitoids in the Central Eastern Desert of Egypt[J]. Journal of African Earth Sciences, 2010, 58(1): 141-151. doi: 10.1016/j.jafrearsci.2010.01.007
[7]	He J, Harris J R, Sawada M, et al. Comparison of classification algorithms using Landsat-7 and Landsat-8 data for mapping lithology in Canada′s Arctic[J]. International Journal of Remote Sensing, 2015, 36(8): 2252-2276. doi: 10.1080/01431161.2015.1035410
[8]	Du P J, Tan K, Xing X S. A novel binary tree support vector machine for hyperspectral remote sensing image classification[J]. Optics Communications, 2012, 285: 3054-3060. doi: 10.1016/j.optcom.2012.02.092
[9]	Mryka H B. Practical guidelines for choosing GLCM textures to use in landscape classification tasks over a range of moderate spatial scales[J]. International Journal of Remote Sensing, 2017, 38: 1312-1338. doi: 10.1080/01431161.2016.1278314
[10]	Seyedmohammadi J, Navidi M N, Esmaeelnejad L. Geospatial modeling of surface soil texture of agricultural land using fuzzy logic, geostatistics and GIS techniques[J]. Communications in Soil Science and Plant Analysis, 2019, 50(12): 1452-1464. doi: 10.1080/00103624.2019.1626870
[11]	Li Q Y, Huang X, Wen D W, et al. Integrating multiple textural features for remote sensing image change detection[J]. Photogrammetric Engineering and Remote Sensing, 2017, 83(2): 109-121. doi: 10.14358/PERS.83.2.109
[12]	Mars J C, Rowan L C. ASTER spectral analysis and lithologic mapping of the Khanneshin carbonatite volcano, Afghanistan[J]. Geosphere, 2011, 7(1): 276-289. doi: 10.1130/GES00630.1
[13]	Ye M, Routsos D. Wavelet-based color texture retrieval using the independent component color space[J]. IEEE International Conference Image Processing, 2008, 15: 165-168.
[14]	Regniers O, Bombrun L, Guyon D, et al. Wavelet-based texture features for the classification of age classes in a maritime pine forest[J]. IEEE Geoscience and Remote Sensing Letters, 2015, 12(3): 621-625. doi: 10.1109/LGRS.2014.2353656
[15]	Wan S A, Chang S S. Crop classification with World View-2 imagery using Support Vector Machine comparing texture analysis approaches and grey relational analysis in Jianan Plain, Taiwan[J]. International Journal of Remote Sensing, 2018, 40(21): 8076-8092.
[16]	Ma D, Lai H C. Remote sensing image matching based improved ORB in NSCT domain[J]. Journal of The Indian Society of Remote Sensing, 2019, 47(5): 801-807. doi: 10.1007/s12524-019-00958-y
[17]	Moonon A U, Hu J W, Li S T. Remote sensing image fusion method based on nonsubsampled shearlet transform and sparse representation[J]. Sensing and Imaging, 2015, 16(1): 18-23.