CN103218806A - Method for detecting submerged mangrove forest distribution in high tide by facing object classification method and based on remote sensing satellite image - Google Patents

Abstract

The invention discloses a method for detecting submerged mangrove forest distribution in high tide by facing an object classification method and based on a remote sensing satellite image and relates to a method of utilizing remote sensing technology to detect mangrove forests below a water surface in the high tide. The method solves the problem that existing remote sensing technology can not accurately detect and obtain the distribution of the mangrove forests submerged below the water surface in the high tide. The method comprises the following steps of utilizing a land resource satellite to obtain a Landsat TM image, and obtaining the Landsat TM image after rectification; conducting multi-scale segmentation on the image after rectification, obtaining a series of segmentation units, and confirming segmentation units of a mangrove forest object and segmentation units of a pure water body object; obtaining spectrum response curve graphs of the mangrove forest object and the pure water body object; building a mangrove forest index; according to the mangrove forest index MVI, judging the Landsat TM image after rectification in the step one to distinguish the mangrove forest object and the pure water body object; and extracting the submerged mangrove forest object, and obtaining distribution information of the submerged mangrove forest object. The method can be widely applied to judgment of the mangrove forest distribution situation in the high tide.

Description

A kind ofly survey high water time based on remote sensing satellite image and object-oriented classification and be submerged the method that mangrove distributes

Technical field

The present invention relates to a kind of method of utilizing remote sensing technology to survey the following mangrove of the high water time water surface.

Background technology

Mangrove has great economy, society, the ecological value, and it can windproof bank protection, and firm shore line purifies water, for the local resident provides important forest product and community service.Keeping sea life diversity, Ecological Environment and Developing Ecological Tourism, carry out scientific research, keeping aspects such as the coastal zone ecologic equilibrium and mitigation and play a part particular importance, is that the Nature is vouchsafed human treasure.Therefore, it is significant to wet land protection management of effective reinforcement mangrove and ecological recovery to make the mangrove thematic maps fast and accurately.

In recent years, remote sensing technology has become the effective ways of making the mangrove thematic maps.But because mangrove grows in area, tropical and subtropical zone seashore mesolittoral zone, periodic tidewater submergence is the essential condition that mangrove grows, and the mangrove tree crown is exposed to the water surface or all is submerged in the bottom at the incoming tide.Existing remote sensing technology can't accurately be surveyed and obtain the mangrove that high water time is submerged under the water surface and distribute.

Summary of the invention

The present invention can't accurately survey the problem that the acquisition high water time is submerged in the mangrove distribution under the water surface in order to solve existing remote sensing technology, surveys the method that high water time is submerged the mangrove distribution thereby provide a kind of based on remote sensing satellite image and object-oriented classification.

A kind ofly survey high water time based on remote sensing satellite image and object-oriented classification and be submerged the method that mangrove distributes, it comprises the steps:

Step 1: utilize road resource satellite to obtain the LandsatTM image, utilize the collinearity equation model that the LandsatTM image is carried out orthorectify, again the LandsatTM image behind the orthorectify is carried out geometric exact correction, the LandsatTM image behind the acquisition registration;

Described geometric exact correction is a two-wire interpolation method geometric exact correction;

Step 2: the LandsatTM image behind the registration that obtains in the step 1 is carried out multi-scale division, obtain a series of cutting units, and determine the cutting unit of mangrove object and the cutting unit of pure water body object respectively according to known mangrove object and pure water body object;

Step 3: according to the cutting unit of mangrove object and the cutting unit of pure water body object, obtain the spectral response curve figure of mangrove object and pure water body object, described spectral response curve figure is the curve map of being made up of each wave band value of LandsatTM image, and one of them wave band is a figure layer of LandsatTM image;

Step 4: the brightness value of image that obtains the brightness value of image of mangrove object and pure water body object according to spectral response curve figure has the wave band of notable difference, i.e. the 4th wave band, described notable difference are that the difference of brightness value of image of the brightness value of image of mangrove object and pure water body object is greater than 20;

Step 5: utilize the 5th wave band strong characteristics that absorb of moisture to the LandsatTM image, and in the step 4 brightness value of image of the 4th wave band of mangrove object greater than the characteristics of the brightness value of image of the 4th wave band of pure water body object, set up the mangrove index, formula is as follows, wherein B4 is the brightness value of image of object the 4th wave band, and B5 is the brightness value of image of the 5th wave band;

$\mathrm{MVI}=(\frac{B4}{B5}-1)*100$

Step 6: divide mangrove object and pure water body object:, be pure water body object smaller or equal to 90 o'clock these objects when MVI this object greater than 90 time is the mangrove object according to the LandsatTM image area behind the mangrove index M VI determining step one described registration;

Step 7: the result according to step 6 is distinguished, extract covered mangrove object, obtain to be submerged the distributed intelligence of mangrove object.

The present invention has realized that remote sensing technology is accurately surveyed and has obtained high water time and be submerged in the problem that the mangrove under the water surface distributes.The present invention at first utilizes object-oriented method that the LandsatTM image is cut apart, and pixel is independently merged the object that becomes homogeneity.Fully analyze the spectral response curve of covered mangrove object and peripheral pure water body object then.The spectral response curve of finding these two kinds of objects only there are differences at the 4th wave band, this species diversity ubiquity, but be not the obvious of ten minutes.Therefore the present invention sets up a kind of new index (mangrove index), amplifies this species diversity.Determine index threshold at last, extract covered mangrove.Through check, covered mangrove all can accurately extract according to the mangrove index more than 90%.The present invention has overcome periodic tidewater and has flooded the difficulty of bringing to the mangrove remote Sensing Interpretation, has solved covered mangrove is divided into water body by mistake problem.The mangrove index that the present invention sets up extracts effective and rapid for covered mangrove, improved the accuracy and confidence of mangrove remote Sensing Interpretation, has repeatability and robustness, and the mangrove remote sensing mapping is had and important meaning.The method of extracting mangrove among the present invention has overcome the difficult problem of remote Sensing Interpretation mangrove.

Description of drawings

Fig. 1 is a process flow diagram of the present invention;

Fig. 2 is the described spectral response curve figure of step 2 of the present invention.

Embodiment

Embodiment one, this embodiment is described in conjunction with Fig. 1 and Fig. 2.A kind ofly survey high water time based on remote sensing satellite image and object-oriented classification and be submerged the method that mangrove distributes, it comprises the steps:

Step 1: utilize road resource satellite to obtain the LandsatTM image, utilize the collinearity equation model that the LandsatTM image is carried out orthorectify, again the LandsatTM image behind the orthorectify is carried out geometric exact correction, the LandsatTM image behind the acquisition registration;

Described geometric exact correction is a two-wire interpolation method geometric exact correction;

Step 2: the LandsatTM image behind the registration that obtains in the step 1 is carried out multi-scale division, obtain a series of cutting units, and determine the cutting unit of mangrove object and the cutting unit of pure water body object respectively according to known mangrove object and pure water body object;

Step 3: according to the cutting unit of mangrove object and the cutting unit of pure water body object, obtain the spectral response curve figure of mangrove object and pure water body object, described spectral response curve figure is the curve map of being made up of each wave band value of LandsatTM image, and one of them wave band is a figure layer of LandsatTM image;

Step 4: the brightness value of image that obtains the brightness value of image of mangrove object and pure water body object according to spectral response curve figure has the wave band of notable difference, i.e. the 4th wave band, described notable difference are that the difference of brightness value of image of the brightness value of image of mangrove object and pure water body object is greater than 20;

Step 5: utilize the 5th wave band strong characteristics that absorb of moisture to the LandsatTM image, and in the step 4 brightness value of image of the 4th wave band of mangrove object greater than the characteristics of the brightness value of image of the 4th wave band of pure water body object, set up the mangrove index, formula is as follows, wherein B4 is the brightness value of image of object the 4th wave band, and B5 is the brightness value of image of the 5th wave band;

$\mathrm{MVI}=(\frac{B4}{B5}-1)*100$

Step 6: divide mangrove object and pure water body object:, be pure water body object smaller or equal to 90 o'clock these objects when MVI this object greater than 90 time is the mangrove object according to the LandsatTM image area behind the mangrove index M VI determining step one described registration;

Step 7: the result according to step 6 is distinguished, extract covered mangrove object, obtain to be submerged the distributed intelligence of mangrove object.

Road resource satellite (Landsat) has been a series of road resource satellites that NASA (NASA) launches since 1972.The sensor that Landsat5 carries is thematic map instrument (TM), contain 7 wave bands (0.45～0.53 μ m, 0.52～0.60 μ m, 0.63～0.69 μ m, 0.76～0.90 μ m, 1.55～1.75 μ m, 10.40～12.50 μ m, 2.08～2.35 μ m), orbit altitude 705km, spatial resolution 30m heavily visits 16 days cycles, has been proved the mangrove spatial information that is very suitable for the coastland and has extracted research.

What embodiment two, this embodiment and embodiment one were different is described step 2: the process of the LandsatTM image behind the registration that obtains in the step 1 being carried out multi-scale division is:

Step 21:, obtain R with the nonoverlapping quartern of LandsatTM image behind the registration _i, i=1 wherein, 2,3,4;

Step 22: the consistance of the inside pixel color harmony texture of subregions such as judgement, when consistance below 85%, then divide;

Described fission process is with the nonoverlapping quartern in these subregions;

Step 23: judge the consistance of the inside pixel color harmony texture of subregion such as adjacent, when consistance more than or equal to 85%, then carry out merging process;

Described fission process and merging process all continue to proceed to can not divide with merge etc. the subregion, promptly the cutting unit internal consistency is greater than 85%, and with the consistance of adjacent cutting unit on every side smaller or equal to 85%.

The condition of division is the consistance decision of object inside, and the condition of merging is the consistance decision between the adjacent object.That is to say that cutting unit inside is very even, and the cutting unit that this cutting unit is adjacent is inconsistent.

Because necessarily there are discernible mangrove object and pure water body object in cutting unit inside, then can judge that this cutting unit is mangrove cutting unit or pure water body cutting unit according to consistance.

The principle of described multi-scale division is: less according to the mangrove plaque area, and color uniqueness, out-of-shape, texture is level and smooth, and the characteristics of compacting are as shown in table 1, the yardstick of cutting apart that limits is 10 (less, be fit to the little atural object of area), color index (0.8 is bigger, is fit to the atural object of color uniqueness), (0.2 is less for shape index, be fit to atural object in irregular shape), smoothness (0.6 is fit to the level and smooth atural object of texture), degree of compacting (0.4 is fit to the atural object that inner vein comparatively compacts).

Table 1

Cut apart yardstick	Color factor	Form factor	Smoothness	Degree of compacting
					10	0.8	0.2	0.6	0.4

Technical solution of the present invention is not limited to above cited concrete remotely-sensed data, also comprises each period, the remote sensing images of various tidal level.

Specific embodiment:

Step 1: download intermediate resolution remote sensing images LandsatTM data, orbit number is P124R45, and the time is on October 30th, 2006, and tidal level is 187cm, and a big chunk mangrove is flooded by tidewater in the image.Use the collinearity equation model intermediate resolution remote sensing images LandsatTM data are carried out orthorectify, utilize 1: 50000 terrain data then, in ERDAS software, choose ground control point, the image behind the orthorectify is carried out geometric exact correction, the LandsatTM image behind the acquisition registration;

Step 2: the LandsatTM image behind the registration that obtains in the step 1 is carried out the multilayer multi-scale division, obtain a series of cutting units, each cutting unit is made up of the pixel that adjacent on the space, homogeney reach more than 85%, with each cutting unit as an object;

Step 3: utilize the spectral response curve of the direct extraction step two resulting covered mangrove objects of object-oriented classification software and the spectral response curve of peripheral pure water body object;

Step 4: analyze the difference of the spectral response curve of covered mangrove object and peripheral pure water body, find that the spectral response curve of these two kinds of objects only has notable difference at the 4th wave band (B4), all the other wave bands are similar substantially.Difference shows as: the reflectivity of covered its 4th wave band of mangrove object is a little more than the reflectivity of pure water body the 4th wave band;

Step 5: utilize LandsatTM the 5th wave band (B5) that water body is changed responsive characteristics, and the 4th wave band can be distinguished the characteristics of water body and covered mangrove object in the step 4, set up index, amplify the SPECTRAL DIVERSITY of the 4th wave band, the index of setting up is called mangrove index (MVI), formula is as follows, and wherein B4 is the value of object the 4th wave band, and B5 is the value of the 5th wave band;

$\mathrm{MVI}=(\frac{B4}{B5}-1)*100$

Step 6: determine the threshold value of mangrove index, distinguish pure water in the face of resembling and be submerged the mangrove object.Find by a series of test, can well distinguish covered mangrove object and peripheral pure water body object when the MVI threshold value is set to 90;

Step 7: utilize the threshold value of determining in object-oriented software and the step 6, extract covered mangrove object.

Claims (2)

1. survey high water time based on remote sensing satellite image and object-oriented classification and be submerged the method that mangrove distributes for one kind, it is characterized in that it comprises the steps:

Step 1: utilize road resource satellite to obtain the LandsatTM image, utilize the collinearity equation model that the LandsatTM image is carried out orthorectify, again the LandsatTM image behind the orthorectify is carried out geometric exact correction, the LandsatTM image behind the acquisition registration;

Described geometric exact correction is a two-wire interpolation method geometric exact correction;

Step 2: the LandsatTM image behind the registration that obtains in the step 1 is carried out multi-scale division, obtain a series of cutting units, and determine the cutting unit of mangrove object and the cutting unit of pure water body object respectively according to known mangrove object and pure water body object;

Step 3: according to the cutting unit of mangrove object and the cutting unit of pure water body object, obtain the spectral response curve figure of mangrove object and pure water body object, described spectral response curve figure is the curve map of being made up of each wave band value of LandsatTM image, and one of them wave band is a figure layer of LandsatTM image;

Step 4: the brightness value of image that obtains the brightness value of image of mangrove object and pure water body object according to spectral response curve figure has the wave band of notable difference, i.e. the 4th wave band, described notable difference are that the difference of brightness value of image of the brightness value of image of mangrove object and pure water body object is greater than 20;

Step 5: utilize the 5th wave band strong characteristics that absorb of moisture to the LandsatTM image, and in the step 4 brightness value of image of the 4th wave band of mangrove object greater than the characteristics of the brightness value of image of the 4th wave band of pure water body object, set up the mangrove index, formula is as follows, wherein B4 is the brightness value of image of object the 4th wave band, and B5 is the brightness value of image of the 5th wave band;

$\mathrm{MVI}=(\frac{B4}{B5}-1)*100$

Step 6: divide mangrove object and pure water body object:, be pure water body object smaller or equal to 90 o'clock these objects when MVI this object greater than 90 time is the mangrove object according to the LandsatTM image area behind the mangrove index M VI determining step one described registration;

Step 7: the result according to step 6 is distinguished, extract covered mangrove object, obtain to be submerged the distributed intelligence of mangrove object.

2. according to claim 1ly a kind ofly survey high water time based on remote sensing satellite image and object-oriented classification and be submerged the method that mangrove distributes, it is characterized in that described step 2: the process of the LandsatTM image behind the registration that obtains in the step 1 being carried out multi-scale division is:

Step 21:, obtain R with the nonoverlapping quartern of LandsatTM image behind the registration _i, i=1 wherein, 2,3,4;

Step 22: the consistance of the inside pixel color harmony texture of subregions such as judgement, when consistance below 85%, then divide;

Described fission process is with the nonoverlapping quartern in these subregions;

Step 23: judge the consistance of the inside pixel color harmony texture of subregion such as adjacent, when consistance more than or equal to 85%, then carry out merging process;

Described fission process and merging process all continue to proceed to can not divide with merge etc. the subregion, promptly the cutting unit internal consistency is greater than 85%, and with the consistance of adjacent cutting unit on every side smaller or equal to 85%.

Images