CN104036507A - Light satellite based urban change area and evolution type rapid extraction method - Google Patents

Abstract

The present invention provides a method for quickly extracting urban changing areas and evolution types based on lighting satellites, which acquires lighting satellite images of different years and performs data preprocessing, selects images of three years, and sets the colors of different urban evolution types, according to Color discrimination realizes rapid extraction of different urban evolution areas, constructs pseudo-spectral curves in different evolution type areas and establishes urban evolution type spectral library, realizes rapid extraction of urban change areas and evolution types, and draws urban evolution type maps. The invention realizes the rapid extraction of urban change areas in a large range and the discrimination of individual urban evolution types, and has the advantages of fast and efficient; it can provide scientific basis for the macro-planning of urbanization by the government, the market layout of consulting companies, and multinational companies, and has great advantages Good market prospect.

Description

Rapid extraction method of urban change area and evolution type based on lighting satellite

technical field

The invention relates to a method for quickly extracting urban change areas and evolution types based on lighting satellites, and belongs to the field of remote sensing image processing and pattern recognition.

Background technique

The current analysis method for urban change areas and evolution types is basically a complex process of applying conventional Landsat, first extracting cities, and then detecting changes. Urban research based on lighting satellite images is also limited to the simple analysis of urban expansion, or the extraction of urban change areas through some change detection methods. The algorithm is complex and inefficient, and there is no research on the rapid extraction of urban change areas and the identification of individual urban evolution types. analyze.

Contents of the invention

In order to solve the deficiencies of the prior art, the present invention provides a method for quickly extracting urban change areas and evolution types based on lighting satellites. By means of false color synthesis, extraction of pseudo spectral curves, and construction of urban evolution type spectral libraries, large-scale The rapid extraction of urban change areas within the scope and the identification of individual urban evolution types have the advantages of fast and efficient.

The technical solution adopted by the present invention to solve the technical problem is: a method for quickly extracting urban change areas and evolution types based on lighting satellites is provided, which specifically includes the following steps:

(1) Obtain satellite images of lights in different years in the same detection area; the years mentioned include at least Y ₁ , Y ₂ and Y ₃ ;

(2) Perform data preprocessing on lighting satellite images:

(201) Radiation calibration: performing relative radiation correction on urban lighting satellite images;

(202) negative value removal: the image after step (201) radiation calibration is carried out to negative value processing, and the value less than 0 is discarded;

(203) Image registration: registering the year data according to the geographic coordinates of the lighting satellite images after the negative value processing in step (202);

(204) Band synthesis: for each image of different years, it is regarded as a band, and the images are synthesized according to the time sequence to obtain a scene of multi-band images;

(3) Set the color of the evolution type, and use different colors to represent different evolution types;

(4) False color synthesis: select images of Y ₁ year, Y ₂ year and Y ₃ year from the city lights satellite images processed in step (204), wherein, Y ₁ <Y ₂ <Y ₃ ; for 3 The images of the year are displayed in false color synthesis, where the image of the year Y ₁ is set as the red channel, the image of the year Y ₂ is set as the green channel, and the image of the year Y ₂ is set as the blue channel;

(5) According to the discrimination standard established in step (3), the evolution type of the false-color composite image in step (4) is preliminarily judged by visual interpretation, so as to divide the detection area into undeveloped area, stable area, developing area and recession district;

(6) Construct a pseudo-spectral curve: select a pixel in each evolution type area in the detection area, extract the DN value of the pixel over the years, construct a DN value change curve with time as the horizontal axis and DN value as the vertical axis, and obtain each Pseudo-spectral curves of evolving types;

(7) Establish evolution type spectral library: according to the colors of different evolution types set in step (3), select a single pixel and/or more than 2 pixels in each evolution type area, and calculate the average value for each evolution type area , and construct pseudo-spectral curves of each evolution type, and establish an evolution-type spectral library composed of pseudo-spectral curves;

(8) Rapid extraction of urban change areas: According to the evolution type spectral library established in step (7), the spectral matching algorithm is used to quickly identify the evolution type, assign corresponding colors to different evolution types, and generate an evolution type map, so that Areas of urban change and types of evolution are shown in the figure.

Further, in step (201), the relative radiation correction is performed using the following formula:

y＝c ₀ +c ₁ x+c ₂ x ² ... (1)

Among them, c ₀ , c ₁ and c ₂ are relative radiation correction parameters, which are provided by the remote sensor manufacturer or user unit, x represents the DN value, and y represents the DN value after relative radiation correction.

In step (202), utilize following formula to remove negative value:

y _DN>0 ＝(x _radiation >0)×x _radiation ......(2)

Among them, y _DN>0 means the image after removing negative values, and x _radiation means the image after relative radiation correction.

Further, the criteria for judging urban evolution types described in step (3) are: use black to represent evolution type 1; use white to represent evolution type 2; use blue, cyan and green to represent evolution type 3; use yellow, red and pink to represent evolution Type 4.

Further, in step (8), the following steps are used to quickly identify the evolution type:

(801) Obtaining the synthetic image of the lighting satellite for the evolution type discrimination;

(802) Obtain the size of the image, set the number of columns and the number of rows of its pixel to be nSample and nLine respectively, the number of bands is nBand, define the independent variable i=0, j=0 that the row and column change;

(803) row number judgment, when row number is less than nLine, enter step 804, otherwise enter step 810;

(804) row number judgment, when row number is less than nSample, enter step 805, otherwise enter step 806;

(805) column number is cleared j=0, row number i increases by 1, returns to step 803;

(806) Obtain the pseudo-spectral curve at the image (i, j);

(807) For the pseudo-spectral curve obtained in step 806, use a spectral matching algorithm to realize matching with the pseudo-spectrum in the evolution type spectral library;

(808) Discriminate the spectral matching result of step 807, obtain the evolution type at the position of step 806 (i, j), and assign the type to the corresponding color;

(809) Column number is accumulated by 1;

(810) After traversing all the rows and columns, end the rapid identification to generate a spectral matching discriminant evolution type map based on the evolution type spectral library.

The beneficial effect that the present invention has based on its technical scheme is:

(1) The present invention avoids the complex process of "extraction first, then change detection", which can save a lot of time and meet the requirements of efficient data processing;

(2) The present invention realizes the rapid extraction of urban change areas in a large range and the discrimination of individual urban evolution types through the means of false color synthesis, extraction of pseudo spectral curves, and construction of urban evolution type spectral libraries, and has the advantages of fast and efficient;

(3) The present invention can provide a scientific basis for the macro-planning of urbanization by the government, the market layout of consulting companies and multinational companies, etc., and has a good market prospect.

Description of drawings

Figure 1 is a flow chart of the rapid extraction method of urban change areas and evolution types based on lighting satellites.

Figure 2 is a flow chart of data preprocessing.

Figure 3 is a schematic diagram of urban evolution types defined by the three primary colors, where A, B, C, D, E, F, and G represent the white stable area, blue development area, cyan development area, green development area, yellow recession area, respectively. Red recession zone and pink recession zone.

Figure 4 is a pseudo-spectral curve in the urban evolution type spectral library.

Figure 5 is a flow chart for rapid identification of urban evolution types.

Figure 6 is a map of urban evolution types.

Detailed ways

The present invention will be further described below in conjunction with drawings and embodiments.

The present invention provides a method for quickly extracting urban change areas and evolution types based on lighting satellites, referring to Figure 1, which specifically includes the following steps:

(1) Obtain satellite images of lights in different years in the same detection area; the years mentioned include at least Y ₁ , Y ₂ and Y ₃ ;

(2) Referring to Figure 2, data preprocessing is performed on the lighting satellite image through the following steps:

(201) Use the following formula to perform relative radiation correction on urban lighting satellite images:

y＝c ₀ +c ₁ x+c ₂ x ² ... (1)

Among them, c ₀ , c ₁ and c ₂ are relative radiation correction parameters, which are provided by the remote sensor manufacturer or user unit, x represents the DN value, and y represents the DN value after relative radiation correction;

(202) Utilize the following formula to carry out de-negative value processing to the image after step (201) radiation calibration:

y _DN>0 ＝(x _radiation >0)×x _radiation ......(2)

Among them, y _DN>0 means the image after removing the negative value, and x _radiation means the image after relative radiation correction;

(203) Image registration: registering the year data according to the geographic coordinates of the lighting satellite images after the negative value processing in step (202);

(204) Band synthesis: for each image of different years, it is regarded as a band, and the images are synthesized according to the time sequence to obtain a scene of multi-band images;

(3) With reference to Figure 3, the color of the evolution type is set, and the three primary colors are used to represent different evolution types, wherein:

Evolution type 1 is an undeveloped area, which is set to black, indicating that the DN value of the light satellite is close to 0, and the urban area has not undergone any changes for many years;

Evolution type 2 is a stable area, set to white, indicating that the DN values of the light satellites in the early, middle, and late stages are all close to 60, and there are no major changes. Here, it represents the urban area that existed in Y ₁ , Y ₂ , and Y ₃ years ;

Evolution type 3 is the development zone, set with the following three colors:

Blue means that there is no change in the period from the early to the middle period, and the DN value of the urban area is increasing continuously from the middle to the late period, which means that there is basically no change from Y ₁ to Y ₂ , and it expands into a city from Y ₂ to Y ₃ Area;

Cyan indicates that it is in a state of development in the early to mid-term, and in a stable state in the mid-to-late period. Here, it represents an urban area that expanded from Y ₁ to Y ₂ , but remained stable in Y ₂ to Y ₃ ;

Green indicates urban areas that expand from the early to mid-term and decline in the mid-to-late period. Here, it is the urban area that expands from Y ₁ to Y ₂ , and the urban area that gradually declines after Y ₃ years.

Evolution type 4 represents the recession zone, set with the following three colors:

Yellow indicates that the area is in a stable state from the early to mid-term, and gradually disappears after the mid-term. Here, it means that the urban area is in a stable state from Y ₁ to Y ₂ , and the urban area begins to die from Y ₂ to Y ₃ ;

Red indicates the urban area that has been in decline from the early stage to the late stage, here it is the urban area that existed in Y ₁ year, and gradually died out after Y ₁ to Y ₂ years;

Pink means that it is in a state of decline from the early to mid-term, and it is in a state of development from the mid-term to the late stage. Here it means that the urban area begins to die from Y ₁ to Y ₂ , and the urban area is in a developing state from Y ₂ to Y ₃ ;

(4) False color synthesis: select images of Y ₁ year, Y ₂ year and Y ₃ year from the city lights satellite images processed in step (204), wherein, Y ₁ <Y ₂ <Y ₃ ; for 3 The images of the year are displayed in false color synthesis, among which, the image of the year Y ₁ is set as the red channel, the image of the year Y ₂ is set as the green channel, and the image of the year Y ₂ is set as the blue channel;

(5) According to the colors of different evolution types set in step (3), the false color composite image in step (4) is visually interpreted to preliminarily determine the evolution type, so as to divide the detection area into evolution type 1, evolution type 2, Evolution type 3 and evolution type 4, that is, undeveloped area, stable area, development area and recession area;

(6) Construct a pseudo-spectral curve: select a pixel in each evolution type area in the detection area, extract the DN value of the pixel over the years, construct a DN value change curve with time as the horizontal axis and DN value as the vertical axis, and obtain each Pseudo-spectral curves of evolving types;

(7) Establish evolution type spectral library: according to the colors of different evolution types set in step (3), select a single pixel and/or more than 2 pixels in each evolution type area, and calculate the average value for each evolution type area , and construct the pseudo-spectral curves of each evolution type, and establish the evolution-type spectral library composed of pseudo-spectral curves; the pseudo-spectral curves in the urban evolution type spectral library are shown in Figure 4;

(8) Rapid extraction of urban change areas: According to the evolution type spectral library established in step (7), the spectral matching algorithm is used to quickly identify the evolution type, assign corresponding colors to different evolution types, and generate an evolution type map, so that The area of urban change and the type of evolution are shown in the figure; referring to Figure 5, the specific steps are as follows:

(801) Obtaining the synthetic image of the lighting satellite for the evolution type discrimination;

(802) Obtain the size of the image, set the number of columns and the number of rows of its pixel to be nSample and nLine respectively, the number of bands is nBand, define the independent variable i=0, j=0 that the row and column change;

(803) row number judgment, when row number is less than nLine, enter step 804, otherwise enter step 810;

(804) row number judgment, when row number is less than nSample, enter step 805, otherwise enter step 806;

(805) column number is cleared j=0, row number i increases by 1, returns to step 803;

(806) Obtain the pseudo-spectral curve at the image (i, j) by using a spectral matching algorithm according to the evolution type spectral library;

(807) For the pseudo-spectral curve obtained in step 806, use a spectral matching algorithm to realize matching with the pseudo-spectrum in the evolution type spectral library;

(808) Discriminate the spectral matching result of step 807, obtain the evolution type at the position of step 806 (i, j), and assign the type to the corresponding color;

(809) Column number is accumulated by 1;

(810) After traversing all the rows and columns, end the rapid identification to generate a spectral matching discriminant evolution type map based on the evolution type spectral library, the effect is shown in FIG. 6 .

Claims (5)

1. the city region of variation based on light satellite and differentiation type rapid extracting method, is characterized in that specifically comprising the following steps:

(1) obtain the light satellite image of different year in same detection region; The described time at least comprises Y ₁time, Y ₂time and Y ₃time;

(2) by following steps, light satellite image is carried out to data pre-service:

(201) radiation calibration: urban lighting satellite image is carried out to relative radiant correction;

(202) go negative value: the image after step (201) radiation calibration is gone to negative value processing, the value that is less than 0 is given up;

(203) Image registration: will go negative value light satellite image after treatment according to geographic coordinate through step (202), to year piece of data carry out registration;

(204) wave band is synthetic: the image of each different year is considered as a wave band, according to time order and function order, image is carried out to wave band and synthesizes, and obtains a scape multiband image;

(3) arrange and develop the color of type, represent different types of evolution with different colours;

(4) false colored synthetic: from choose Y through step (204) urban lighting satellite image after treatment ₁time, Y ₂time and Y ₃the image in time, wherein, Y ₁<Y ₂<Y ₃; The image in 3 times is carried out to false colored compound display, wherein, Y ₁time image is set to red channel, Y ₂time image is set to green channel, Y ₂time image is set to blue channel;

(5) color of the different types of evolution arranging according to step (3), develops type to the False color comp osite image of step (4) by visual interpretation preliminary judgement, surveyed area is divided into the region of different types of evolution;

(6) build the pseudo-curve of spectrum: choose respectively in surveyed area each pixel that develops type area, extract this pixel DN value over the years, taking the time as transverse axis, DN value is that the longitudinal axis is constructed DN value change curve, obtains the pseudo-curve of spectrum of each differentiation type;

(7) set up and develop type library of spectra: the color of the different types of evolution arranging according to step (3), choose respectively the single pixel of each differentiation type area and/or more than 2 pixel, average for each differentiation type area respectively, and construct the pseudo-curve of spectrum of each differentiation type, set up the differentiation type library of spectra being formed by the pseudo-curve of spectrum;

(8) rapid extraction of city region of variation: the differentiation type library of spectra of setting up according to step (7), adopt Spectral matching algorithm, identify fast developing type, for different types of evolution is given corresponding color, generate and develop type map, thereby city region of variation and differentiation type are shown in the drawings.

2. the city region of variation based on light satellite according to claim 1 and differentiation type rapid extracting method, is characterized in that: in step (201), utilize following formula to carry out relative radiant correction:

y＝c ₀+c ₁x+c ₂x ²……(1)

Wherein, c ₀, c ₁and c ₂be relative radiant correction parameter, provided by remote sensor production unit or Subscriber Unit, x represents DN value, and y represents the DN value after relative radiant correction.

3. the city region of variation based on light satellite according to claim 1 and differentiation type rapid extracting method, is characterized in that: in step (202), utilize following formula to go negative value:

y _DN>0＝(x _radiation>0)×x _radiation……(2)

Wherein, y _dN>0represent the image after the negative value of place to go, x _radiationrepresent the image after relative radiant correction.

4. the city region of variation based on light satellite according to claim 1 and differentiation type rapid extracting method, is characterized in that: step (3) is described represents that with different colours different types of evolution is specially:

Represent to develop Class1 with black; Represent to develop type 2 by white; Represent to develop type 3 by blue, cyan and green; Represent to develop type 4 by yellow, redness and pink colour.

5. the city region of variation based on light satellite according to claim 1 and differentiation type rapid extracting method, is characterized in that: in step (8), utilize following steps to identify fast developing type:

(801) obtain the light satellite resultant image that will develop type identification;

(802) obtain the size of image, columns and the line number of establishing its pixel are respectively nSample and nLine, and wave band number is nBand, the independent variable i=0 that definition ranks change, j=0;

(803) line number judgement, line number, in the time being less than nLine, enters step 804, otherwise enters step 810;

(804) columns judgement, columns, in the time being less than nSample, enters step 805, otherwise enters step 806;

(805) columns zero clearing j=0, line number i increases by 1, returns to step 803;

(806) obtain the pseudo-curve of spectrum that image (i, j) is located;

(807) pseudo-curve of spectrum step 806 being obtained, utilizes Spectral matching algorithm to realize and the mating of pseudo-spectrum in differentiation type library of spectra;

(808) step 807 Spectral matching result is differentiated, obtained the differentiation type of step 806 (i, j) position, and give corresponding color by the type;

(809) columns cumulative 1;

(810) traveled through all row and columns, finished quick identification and differentiate differentiation type map with the Spectral matching generating based on developing type library of spectra.