CN104573690B - A kind of Technique for Hyper-spectral Images Classification for gypsum information extraction - Google Patents

A kind of Technique for Hyper-spectral Images Classification for gypsum information extraction Download PDF

Info

Publication number
CN104573690B
CN104573690B CN201310502783.9A CN201310502783A CN104573690B CN 104573690 B CN104573690 B CN 104573690B CN 201310502783 A CN201310502783 A CN 201310502783A CN 104573690 B CN104573690 B CN 104573690B
Authority
CN
China
Prior art keywords
gypsum
wave band
information extraction
hyper
image
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201310502783.9A
Other languages
Chinese (zh)
Other versions
CN104573690A (en
Inventor
杨燕杰
赵英俊
张宏光
陆冬华
冯平
王奉宝
田丰
伊丕源
张东辉
周家晶
张铁岭
范光
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Research Institute of Uranium Geology
Original Assignee
Beijing Research Institute of Uranium Geology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Research Institute of Uranium Geology filed Critical Beijing Research Institute of Uranium Geology
Priority to CN201310502783.9A priority Critical patent/CN104573690B/en
Publication of CN104573690A publication Critical patent/CN104573690A/en
Application granted granted Critical
Publication of CN104573690B publication Critical patent/CN104573690B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02—Preprocessing

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Image Analysis (AREA)
  • Investigating Or Analysing Materials By Optical Means (AREA)

Abstract

The invention belongs to Technique for Hyper-spectral Images Classification, more particularly to a kind of Technique for Hyper-spectral Images Classification for gypsum information extraction.It includes：Step 1, pretreatment；Step 2：Sampling, to wave band in 1145nm, 1175nm, 1325nm, 1445nm, 1460nm, 1700nm, 1745nm, 1820nm, 1940nm, 2060nm, 2210nm, 2330nm, 2405nm image sampling；Step 3：Judge, and step 4：Calculate.The present invention effect of the method is that：13 wave bands are only used, relative to 101 wave bands of Hyperspectral imaging SASI all bands, it is necessary to which the data volume of processing reduces 87%, and due to being the automatic onestep extraction of computer, reduce the operating procedures such as the selection of principal component transform, end member wave spectrum, arithmetic speed can improve more than 7 times.Due to eliminating most of wave band little to information extraction relation, the interference of other materials or noise to its spectrum is reduced, improves the precision of information extraction.There is preferable function and significance to the rapid extraction of gypsum information in airborne-remote sensing.

Description

A kind of Technique for Hyper-spectral Images Classification for gypsum information extraction
Technical field
The invention belongs to Technique for Hyper-spectral Images Classification, more particularly to a kind of EO-1 hyperion shadow for gypsum information extraction As processing method.
Background technology
The gypsum information extracting method of current target in hyperspectral remotely sensed image is mainly that the matching of spectrum all band or part connect The Spectral matching of continuous wave band, specific algorithm has spectral modeling, mixing demodulation filtering etc., because the material composition of earth's surface is seldom by list One mineral composition, these methods are easily influenceed in the process of information extraction by other ground-object spectrums or noise, extraction information essence Spend relatively low.Secondly existing spectrum extracting method manual steps are more, add artificial error in judgement.3rd is bloom Modal data wave band is more, and data volume is big, existing method processing time length, reduces the speed of data processing and using scale.Cause How this, reduce the influence, manual steps and processing data of other atural objects or noise during gypsum information extraction Amount, turn into one of forward position of current target in hyperspectral remotely sensed image processing.
The content of the invention
The technical problem to be solved in the present invention is to provide a kind of influence for reducing other atural objects or noise, reduce processing The Technique for Hyper-spectral Images Classification for gypsum information extraction of data volume.By specific to Hyperspectral imaging resampling, extraction Wave band, a series of judgements and calculating are carried out, calculate the Abundances of the different zones gypsum in image capturing range.
In order to solve the above technical problems, the present invention provides a kind of Hyperspectral imaging processing side for gypsum information extraction Method, include successively：
Step 1, pretreatment
Obtain Hyperspectral imaging；To pre-processing for Hyperspectral imaging, atmospheric correction is carried out, obtains ground surface reflectance Image data；
Step 2：Sampling
The image data that step 1 obtains is sampled, to wave band in 1145nm, 1175nm, 1325nm, 1445nm, 1460nm, 1700nm, 1745nm, 1820nm, 1940nm, 2060nm, 2210nm, 2330nm, 2405nm image sampling, and according to Secondary to be recorded as b1~b13, i.e. b1 is wave band 1145nm sampled data, and b2 is wave band 1175nm sampled data, by that analogy；
Step 3：Judge
By a series of following judgements of progress, and record result
a1=（B1 is more than b2）；
a2=（B3 is more than b4）；
a3=（B4 is less than b5）；
a4=（B6 is more than b7）；
a5=（B7 is less than b8）；
a6=（B8 is more than b9）；
a7=（B9 is less than b10）；
a8=（B10 is more than b11）；
a9=（B12 is more than b13）；
Above-mentioned judgement is for judging that the corresponding pixel of image is judged every time；
This step obtains a1~a9 after terminating, totally 9 matrixes；
Step 4：Calculate
Calculating b14 is carried out with following formula
b14=b1+b3+b5+b6+2*b8+2*b10+b12–b2-2*b4-2*b7-2*b9-b11-b13
Described * represents to be multiplied；
A0 is calculated with formula below
a0=a1*a2*a3*a4*a5*a6*a7*a8*a9*b11
Above-mentioned all calculating are that corresponding pixel calculates.
The present invention effect of the method is that：13 wave bands are only used, relative to Hyperspectral imaging SASI all bands 101 Wave band is, it is necessary to the data volume of processing reduces 87%, and due to being the automatic onestep extraction of computer, reduce principal component transform, The operating procedures such as the selection of end member wave spectrum, arithmetic speed can improve more than 7 times.It is most of to information extraction due to eliminating The little wave band of relation, the interference of other materials or noise to its spectrum is reduced, improve the precision of information extraction.To EO-1 hyperion The rapid extraction of gypsum information has preferable function and significance in image data.
Embodiment
The present invention comprises the steps successively：
Step 1, pretreatment
Obtain Hyperspectral imaging；To pre-processing for Hyperspectral imaging, atmospheric correction is carried out, obtains ground surface reflectance Image data；
Carry out atmospheric correction described in this step is processing method well known in the art.
Step 2：Sampling
The image data that step 1 obtains is sampled, to wave band in 1145nm, 1175nm, 1325nm, 1445nm, 1460nm, 1700nm, 1745nm, 1820nm, 1940nm, 2060nm, 2210nm, 2330nm, 2405nm image sampling, and according to Secondary to be recorded as b1~b10, i.e. b1 is wave band 1145nm sampled data, and b2 is wave band 1175nm sampled data, with such Push away.What sampling obtained every time is all a width gray-scale map, and the value of each pixel is its gray value in figure, i.e. b1 is a width gray scale Figure, b1 images（1,1）The value of point be gray value, and remaining is put, and the rest may be inferred, and remaining sample graph is also.
Step 3：Judge
By a series of following judgements of progress, and record result
a1=（B1 is more than b2）；
a2=（B3 is more than b4）；
a3=（B4 is less than b5）；
a4=（B6 is more than b7）；
a5=（B7 is less than b8）；
a6=（B8 is more than b9）；
a7=（B9 is less than b10）；
a8=（B10 is more than b11）；
a9=（B12 is more than b13）；
Above-mentioned judgement be for judging that the corresponding pixel of image is judged every time, with a1=（B1 is more than b2）Exemplified by, take Certain pixel of b1 images（Such as（1,1）Point）Gray value, pixel corresponding to b2 images（When b1 images take（1,1）Point, then b2 Image must also take（1,1）Point）Gray value, then according to judgment rule " b1 more than b2 " judge, when judged result is "Yes" When, record judged result is 1, and it is 0 otherwise to record result.Therefore when a1=（B1 is more than b2）Judgement is when finishing, obtained a1 be with The matrix of b1 matrix formed objects, wherein the value each put is the judged result obtained according to judgment rule（The value each put It is 0 or 1）.
It according to similar rule it is judged that also carry out.This step obtains a1~a9 after terminating, totally 9 matrixes.
Step 4：Calculate
Calculating b14 is carried out with following formula
b14=b1+b3+b5+b6+2*b8+2*b10+b12–b2-2*b4-2*b7-2*b9-b11-b13
Described * represents to be multiplied；
A0 is calculated with formula below
a0=a1*a2*a3*a4*a5*a6*a7*a8*a9*b11
Above-mentioned all calculating are that corresponding pixel calculates, i.e., are calculated using the corresponding pixel of different images.With b14=b1+b3 Exemplified by+b5+b6+2*b8+2*b10+b12-b2-2*b4-2*b7-2*b9-b11-b13 formula, when calculating point（X, y）When, take b1, B3, b5, b6, b8, b10, b12, b2, b4, b7, b9, b11, b13 point（X, y）Gray value participate in calculating, obtained result is B14 point（X, y）Value.In another example a0=a1*a2*a3*a4*a5*a6*a7*a8*a9*b11, when calculating point（X, y）When, take A1, a2, a3, a4, a5, a6, a7, a8, a9, b11 point（X, y）Value participate in calculate, obtained result is a0 point（X, y）'s Value.
The a0 being calculated is exactly the abundance figure of gypsum information, i.e., the numerical value in certain region is bigger in image represents the region stone The abundance of cream is higher.

Claims (1)

  1. A kind of 1. Technique for Hyper-spectral Images Classification for gypsum information extraction, it is characterised in that：Comprise the steps
    Step 1, pretreatment
    Obtain Hyperspectral imaging；To pre-processing for Hyperspectral imaging, atmospheric correction is carried out, obtains the image of ground surface reflectance Data；
    Step 2：Sampling
    The image data that step 1 obtains is sampled, to wave band in 1145nm, 1175nm, 1325nm, 1445nm, 1460nm, 1700nm, 1745nm, 1820nm, 1940nm, 2060nm, 2210nm, 2330nm, 2405nm image sampling, and according to Secondary to be recorded as b1~b13, what sampling obtained every time is all a width gray-scale map, and the value of each pixel is its gray value in figure, That is b1 is wave band 1145nm sampled data, and b2 is wave band 1175nm sampled data, by that analogy；
    Step 3：Judge
    A series of following judgements are carried out, and record result
    A1=(b1 is more than b2)；
    A2=(b3 is more than b4)；
    A3=(b4 is less than b5)；
    A4=(b6 is more than b7)；
    A5=(b7 is less than b8)；
    A6=(b8 is more than b9)；
    A7=(b9 is less than b10)；
    A8=(b10 is more than b11)；
    A9=(b12 is more than b13)；
    Above-mentioned judgement is that, when judged result is "Yes", record is sentenced for judging that the corresponding pixel of image is judged every time Disconnected result is 1, and it is 0 otherwise to record result,
    This step obtains a1~a9 after terminating, totally 9 matrixes；
    Step 4：Calculate and carry out calculating b14 with following formula
    B14=b1+b3+b5+b6+2*b8+2*b10+b12-b2-2*b4-2*b7-2*b9-b11-b13
    Described * represents to be multiplied；
    A0 is calculated with formula below
    A0=a1*a2*a3*a4*a5*a6*a7*a8*a9*b14
    A0 is the abundance figure of gypsum information；
    Above-mentioned all calculating are that corresponding pixel calculates, i.e., are calculated using the corresponding pixel of different images.
CN201310502783.9A 2013-10-23 2013-10-23 A kind of Technique for Hyper-spectral Images Classification for gypsum information extraction Active CN104573690B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201310502783.9A CN104573690B (en) 2013-10-23 2013-10-23 A kind of Technique for Hyper-spectral Images Classification for gypsum information extraction

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201310502783.9A CN104573690B (en) 2013-10-23 2013-10-23 A kind of Technique for Hyper-spectral Images Classification for gypsum information extraction

Publications (2)

Publication Number Publication Date
CN104573690A CN104573690A (en) 2015-04-29
CN104573690B true CN104573690B (en) 2018-03-16

Family

ID=53089712

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201310502783.9A Active CN104573690B (en) 2013-10-23 2013-10-23 A kind of Technique for Hyper-spectral Images Classification for gypsum information extraction

Country Status (1)

Country Link
CN (1) CN104573690B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106845326B (en) * 2015-12-04 2020-10-23 核工业北京地质研究院 Glacier identification method based on aviation hyperspectral remote sensing data
CN109738371A (en) * 2018-12-20 2019-05-10 核工业北京地质研究院 A kind of spectral manipulation method suitable for extracting ferrous ion abundance messages

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102385694A (en) * 2010-09-06 2012-03-21 邬明权 Hyperspectral identification method for land parcel-based crop variety
CN102636778A (en) * 2012-02-21 2012-08-15 核工业北京地质研究院 Information extracting method suitable for high-spectrum image

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4097873B2 (en) * 2000-03-06 2008-06-11 富士フイルム株式会社 Image compression method and image compression apparatus for multispectral image

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102385694A (en) * 2010-09-06 2012-03-21 邬明权 Hyperspectral identification method for land parcel-based crop variety
CN102636778A (en) * 2012-02-21 2012-08-15 核工业北京地质研究院 Information extracting method suitable for high-spectrum image

Also Published As

Publication number Publication date
CN104573690A (en) 2015-04-29

Similar Documents

Publication Publication Date Title
CN103903225B (en) A kind of Technique for Hyper-spectral Images Classification for dolomite information retrieval
CN103900964A (en) Hyperspectral image processing method used for extracting muscovite information
CN105787915A (en) Hyperspectral image processing method for extracting information of jarosite
CN110207592B (en) Building crack measuring method and device, computer equipment and storage medium
CN103900965A (en) Hyperspectral image processing method used for extracting calcite information
CN103714345B (en) A kind of method and system of binocular stereo vision detection finger fingertip locus
JP2009516172A5 (en)
WO2006102570A3 (en) System and method for vascular segmentation by monte-carlo sampling
CN110348425B (en) Method, device and equipment for removing shading and computer readable storage medium
CN103558190A (en) Atmospheric correction method for multi-spectral data of inland turbid water body based on green light wave band
CN104573690B (en) A kind of Technique for Hyper-spectral Images Classification for gypsum information extraction
CN103901497B (en) A kind of Technique for Hyper-spectral Images Classification for illite information extraction
CN107341807B (en) Method for extracting tobacco leaf color digital expression characteristic value
CN103902999B (en) A kind of Technique for Hyper-spectral Images Classification for montmorillonite information extraction
EP3067856A1 (en) Image processing apparatus and image processing method
CN103900967B (en) A kind of Technique for Hyper-spectral Images Classification for kaolin information extraction
CN103900966B (en) A kind of Technique for Hyper-spectral Images Classification for allochite information extraction
CN103902998A (en) High-spectral image processing method for chlorite information extraction
CN104574283B (en) A kind of Technique for Hyper-spectral Images Classification for pyrophyllite information extraction
CN108230365B (en) SAR image change detection method based on multi-source difference image content fusion
CN105405102A (en) High-spectral image processing method for gibbsite information extraction
Zhang et al. Analyzing the saturation of growing stem volume based on ZY-3 stereo and multispectral images in planted coniferous forest
Li et al. Comparison of different analytical edge spread function models for MTF calculation using curve-fitting
CN104732488B (en) A kind of Technique for Hyper-spectral Images Classification for actinolite information extraction
CN107886530A (en) A kind of improved image registration algorithm based on SIFT feature

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant