CN105719236B - The method for generating complete high-resolution vegetation cover degree image seasonal variations sequence - Google Patents
The method for generating complete high-resolution vegetation cover degree image seasonal variations sequence Download PDFInfo
- Publication number
- CN105719236B CN105719236B CN201610024214.1A CN201610024214A CN105719236B CN 105719236 B CN105719236 B CN 105719236B CN 201610024214 A CN201610024214 A CN 201610024214A CN 105719236 B CN105719236 B CN 105719236B
- Authority
- CN
- China
- Prior art keywords
- cover degree
- vegetation cover
- resolution
- image
- period
- 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
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformation in the plane of the image
- G06T3/40—Scaling the whole image or part thereof
- G06T3/4053—Super resolution, i.e. output image resolution higher than sensor resolution
- G06T3/4076—Super resolution, i.e. output image resolution higher than sensor resolution by iteratively correcting the provisional high resolution image using the original low-resolution image
Abstract
The present invention relates to a kind of methods generating complete high-resolution vegetation cover degree image seasonal variations sequence, include the following steps：Based on high-precision land-use map, the pure pixel on high-resolution vegetation cover degree image using the search of window sliding method suitable for low-resolution image establishes pure pixel point data base；Based on the low resolution vegetation cover degree image of 24 periods half a month, the vegetation cover degree seasonal variations curve library of pure pixel is established；Preliminary NO emissions reduction processing is carried out to low resolution vegetation cover degree image, obtains the low resolution vegetation cover degree NO emissions reduction result images consistent with the resolution ratio of high-resolution vegetation cover degree image；Fusion interpolation generates the complete high-resolution vegetation cover degree image seasonal variations sequence for including 24 periods half a month.The present invention has merged Resolutions remote sensing cover degree, and the seasonal variations of different vegetation are reflected by NO emissions reduction image season distribution sequence, while supplementing the spatial variations situation of sub-pixed mapping scale by high-resolution vegetation cover degree image again.
Description
Technical field
The present invention relates to a kind of methods obtaining vegetation cover degree, and in particular to a kind of to generate complete high-resolution vegetation lid
Spend the method for image seasonal variations sequence.
Background technology
Vegetation cover degree refers to that phytobiocoenose is overall or the ratio between the planimetric area of the aerial part of each individual and sample area
Percentage, it reflect vegetation dense degree and plant carry out photosynthesis area size.
Monthly with 15 days for boundary, two half a months are divided into, wherein 1-15 days is the one before half a month, and 16 are rear a to the end of month
Half a month, such whole year are divided into 24 half a months.There are vegetation cover degree in annual 24 half a months, each half a month, can reflect vegetation lid
The season distribution of degree.
Currently, measuring vegetation cover degree by remote sensing mode, two kinds of vegetation cover degree image can be generally obtained：(1)
Low resolution vegetation cover degree image, satellite returns to that the period is short although resolution ratio is low, and the complete season of such as 24 periods half a month becomes
Change vegetation cover degree image；(2) high-resolution vegetation cover degree image, satellite long, the seasonal variations that return to the period although high resolution
Sequence is imperfect, and typically only only there are one the vegetation cover degree images of period or a few period.
Invention content
Complete high-resolution vegetation cover degree image seasonal variations sequential images are generated the object of the present invention is to provide a kind of
Or the method for image, merge two kinds of remote sensing cover degrees of high-resolution and low resolution.
It is described the present invention provides a kind of method generating complete high-resolution vegetation cover degree image seasonal variations sequence
Method merges the high-resolution vegetation cover degree figure of the low resolution vegetation cover degree image and at least one period of 24 periods half a month
Picture, wherein described method includes following steps：
Step 1: being based on high-precision land-use map, searched using window sliding method on high-resolution vegetation cover degree image
Rope is suitable for the pure pixel of low-resolution image, establishes pure pixel point data base；
Step 2: the low resolution vegetation cover degree image based on 24 periods half a month, establishes the vegetation cover degree season of pure pixel
Save change curve library；
Step 3: carrying out preliminary NO emissions reduction processing to low resolution vegetation cover degree image, obtain and high-resolution vegetation lid
Spend the consistent low resolution vegetation cover degree NO emissions reduction result images of the resolution ratio of image；And
Step 4: low resolution vegetation cover degree NO emissions reduction result images and existing 1 or the high score of a few period
Resolution vegetation cover degree image carries out fusion interpolation, generates the complete high-resolution vegetation cover degree image for including 24 periods half a month
Seasonal variations sequence.
According to the method for the present invention it may be preferred that in the step 3, with closest same land use pattern
The vegetation cover degree of pure pixel carries out assignment to the vegetation cover degree of sub-pixed mapping in low resolution vegetation cover degree image, to obtain low point
Resolution vegetation cover degree NO emissions reduction result images, formula are as follows：
In formula, fL’(j, t) indicates some mixed pixel of t period low resolution vegetation cover degree NO emissions reduction result images
In, the vegetation cover degree image of jth kind land use pattern sub-pixed mapping；fp LIt is closest with the mixed pixel that (j, t) refers to the t periods
The vegetation cover degree image of the pure pixel of jth kind land use pattern, can from pure pixel vegetation cover degree seasonal variations curve library root
It searches for obtain according to the nearest principle of space length；fL(t) indicate that the mixed pixel is in the vegetation lid of t periods on low-resolution image
Spend image；Any mixed pixel is made of N number of sub-pixed mapping, contains m kind land use patterns, the Asia of same land use pattern
Pixel is likely to have multiple, rjIndicate that the sub-pixed mapping of jth kind land use pattern accounts for the area ratio of mixed pixel, fp L(i,t)
Refer to the vegetation cover degree image of t periods and i-th kind of closest pure pixel of land use pattern of the mixed pixel, it can be from pure picture
It searches for obtain according to the nearest principle of space length in first vegetation cover degree seasonal variations curve library.
In the method for the invention it may be preferred that the fusion interpolation of the step 4 in the following ways：
When on the basis of period half a month where the high-resolution vegetation cover degree image of existing 1 or a few period
Section,
(1) for only there are one the fusion interpolations when benchmark period, using such as following formula A：
fH(t)=fH(k)+(fL(t)-fL(k))
In formula A, the benchmark period is k (1≤k≤24), and t is the period (1≤t≤24, and t ≠ k) for waiting for interpolation；fH(t) table
Show the high-resolution vegetation cover degree image of t periods；fH(k) the high-resolution vegetation cover degree image in benchmark period k, f are indicatedL
(k) the low resolution vegetation cover degree NO emissions reduction result images in benchmark period k, f are indicatedL(t) the low resolution in the t periods is indicated
Rate vegetation cover degree NO emissions reduction result images；
(2) for there are two or fusion interpolation when more than two benchmark period,
A. added time section to be inserted is between two benchmark periods, using such as following formula B：
fH(t)=α { fH(k)+(fL(t)-fL(k))}+β{fH(q)+(fL(t)-fL(q))}
In formula B, first benchmark period serial number k, second benchmark period serial number q (q>K), t be when interpolation when
Section (k<t<q)；fH(t) the high-resolution vegetation cover degree image of t periods is indicated；fH(k)、fH(q) it is illustrated respectively in the benchmark period
K, the high-resolution vegetation cover degree image of q；fL(t) the low resolution vegetation cover degree NO emissions reduction result images in the t periods are indicated；
fL(k)、fL(q) the low resolution vegetation cover degree NO emissions reduction result images of benchmark period k, q are illustrated respectively in；α, β are respectively weight
Coefficient, size are in inverse relation with the time gap to two benchmark periods；D is adjustable weight parameter, and D values are bigger,
Interpolation result is also bigger by the effect of the nearlyr benchmark period image of distance；
B. added time section to be inserted is before the smallest sequence number benchmark period, using such as following formula C：
fH(t)=fH(k)+(fL(t)-fL(k))
In formula C：The benchmark period is k (1<k<24), t is the period (1≤t for waiting for interpolation<k)；fH(t) the t periods are indicated
High-resolution vegetation cover degree image；fH(k) the high-resolution vegetation cover degree image in benchmark period k, f are indicatedL(k) it indicates in base
The low resolution vegetation cover degree NO emissions reduction result images of punctual section k, fL(t) the low resolution vegetation cover degree in the t periods is indicated
NO emissions reduction result images；
C. added time section to be inserted is after the maximum serial number benchmark period, using such as following formula D：
fH(t)=fH(q)+(fL(t)-fL(q))
In formula D：If the benchmark period is q (1<q<24), t is the period (q for waiting for interpolation<t≤24)；fH(t) the t periods are indicated
High-resolution vegetation cover degree image；fH(q) the high-resolution vegetation cover degree image in benchmark period q, f are indicatedL(q) it indicates
The low resolution vegetation cover degree NO emissions reduction result images of benchmark period q, fL(t) the low resolution vegetation lid in the t periods is indicated
Spend NO emissions reduction result images.
The present invention utilizes the complete vegetation cover degree image season distribution sequence of low resolution, interpolation to lack the height of period half a month
Resolution ratio vegetation cover degree image forms 24 half a month vegetation cover degree images of high-resolution complete season distribution sequence.High score
Resolution vegetation cover degree image is merged with the NO emissions reduction result images of low resolution vegetation cover degree image, passes through NO emissions reduction image season
Distribution series reflect the seasonal variations of different vegetation, while supplementing sub-pixed mapping scale by high-resolution vegetation cover degree image again
Spatial variability information.
Description of the drawings
Fig. 1 is high-resolution vegetation cover degree image and low resolution vegetation cover degree image；
Fig. 2 is the mixed pixel of low resolution vegetation cover degree image and pure pixel schematic diagram；
Fig. 3 is one of the NO emissions reduction schematic diagram of low resolution vegetation cover degree image；
Fig. 4 is the a-quadrant NO emissions reduction schematic diagram in Fig. 3 of low resolution vegetation cover degree image；And
Fig. 5 is the B area NO emissions reduction schematic diagram in Fig. 4 of low resolution vegetation cover degree image.
Specific implementation mode
Below in conjunction with drawings and embodiments, the present invention is described in further detail.
One, data basis：
(1) the remote sensing cover degree data of low resolution：24 half a month low resolution vegetation cover degree image parameters, GRID or IMG
Format, cover degree value is between 0-100.
(2) high-resolution remote sensing cover degree data：K half a month high-resolution vegetation cover degree image parameter (1≤k<24),
GRID or IMG formats, cover degree value is between 0-100.
(3) land use data derives from high-definition picture, GRID or IMG formats.
Two, basic assumption
(1) for identical land use pattern, the vegetation cover degree season change of high-definition picture and low-resolution image
Change trend has similitude, i.e.,：
fH(m,j)-fH(m, i)=fL(m,j)-fL(m, i) formula (1)
In formula：M indicates land use pattern；J and i indicates segment number when different half a months；fHIt indicates with high-resolution
Vegetation cover degree image, fLIndicate the vegetation cover degree image of low resolution.
(2) in the preliminary NO emissions reduction step of low-resolution image, the pixel of low-resolution image is mostly mixed pixel, every
In a mixed pixel, the different sub-pixed mapping vegetation cover degrees of identical land use pattern are equal；
(3) in the same period, for each pixel of low-resolution image, sub-pixed mapping vegetation cover degree and with land use class
The difference of the vegetation cover degree of the closest pure pixel of type is equal, i.e.,：
In formula：fL’Indicate the sub-pixed mapping vegetation cover degree image of mixed pixel in low resolution vegetation cover degree image, fp LIt indicates
Pure pixel vegetation cover degree seasonal variations curve；T indicates that period half a month, i and j indicate the coding value of land use pattern；fL’(i,
T) the vegetation cover degree image of i-th kind of land use pattern of sub-pixed mapping in t period mixed pixels, f are indicatedp L(i, t) indicates the t periods
The vegetation cover degree image of the i-th kind land use pattern pure pixel nearest apart from the mixed pixel.
Three, the method that high-resolution vegetation cover degree seasonal variations sequence generates
Step 1: based on the land-use map that high-resolution vegetation cover degree image obtains, in high-resolution vegetation cover degree figure
The pure pixel of low-resolution image is applicable to using the search of window sliding method as on, establishes pure pixel point data base.
Pure pixel is defined as, if in mixed pixel certain land use pattern accounting some threshold percentage (such as
95%) more than, you can be referred to as pure pixel.Each grid is exactly a pixel.It, can be to land use pattern according to application demand
Part merger processing is carried out, the following table 1 gives a kind of merger processing mode.
1 land use pattern of table
Select a certain size rectangular window in land-use map and certain high-resolution vegetation cover degree image (such as summer period
Season) it is slided：(1) ratio for counting each land use pattern in the window in land-use map, counts in vegetation cover degree figure
The window mean absolute deviation；(2) if some land use pattern percentage is more than or equal to some percentage threshold (such as
95%) and mean absolute deviation is less than or equal to some threshold value (such as 2.5%), records the land use pattern and its position (window
Plane coordinates or column locations after the longitude and latitude at center, projection), land use pattern ratio, mean absolute deviation and window it is big
It is small, and think that the pixel on low-resolution image where the position is pure pixel.
The minimum value of rectangular window size be low-resolution image resolution ratio divided by high-definition picture resolution ratio simultaneously
It is rounded up to the rounding value (along the increased direction of absolute value), actual window size may be the integral multiple of minimum value.It is average exhausted
Each pixel value of window is referred to deviation (Mean Absolute Difference) and the absolute difference of pixel average value in window is asked
With, then be averaged:
In formula：MAD is mean absolute deviation；xiRefer to the value of i-th of pixel, n is pixel number；Expression is averaged；||
Expression takes absolute value.When actual implementation, the threshold value that pure pixel defines can be used as the input parameter of a user interface, and default value is
95%.The rectangular window size also input parameter as a user interface, according to low resolution and high-resolution relative size
It determines.The coefficient of variation threshold value also input parameter as a user interface, default value 2.5%.
If the pure pixel point that search result is found is too many, so that vector file may be larger, the reason is that finds is pure
Pixel point is much adjacent.At this point it is possible to be joined by adjusting the inputs such as pure pixel threshold value, mean absolute deviation, window size
Number so that pure pixel point number and distribution are more reasonable.
It searches for obtained pure pixel position and its relevant information constitutes pure pixel point parameter library.To pure pixel point parameter library into
Pedestrian's work is verified：Judge various land use patterns and whether have, the spatial distribution reasonability of pure pixel point, pure pixel point it is poly-
Collection property.According to verification as a result, the pure pixel point parameter library of human-edited, is increased or deleted to pure pixel point, and rejected
Abnormal point or points of contamination.
Step 2: the low resolution vegetation cover degree image based on 24 periods half a month, establishes the vegetation cover degree season of pure pixel
Save change curve library.
The pure pixel dot position information searched for according to front, the vegetation lid of each pure pixel is extracted from 24 periods half a month
Degree, it is preliminary to establish the vegetation cover degree seasonal variations library for including all pure pixels, the information that each pure pixel includes：The volume of pure pixel
Code, land use pattern, spatial position (being indicated with latitude and longitude coordinates or column locations), land use pattern ratio, variation lines
Number sizes, window size, the vegetation cover degree of 24 periods half a month.
The programming count point of quality examination is carried out to tentatively establishing the vegetation cover degree seasonal variations library comprising all pure pixels
Analysis：
(1) same land use pattern：The average vegetation cover degree of day part is calculated, the season for drawing average vegetation cover degree becomes
Change curve；
(2) mean absolute difference of each pure pixel point and average vegetation cover degree is calculated, calculation formula is similar with formula (3b), only
It is n=24 at this time；
(3) related coefficient of each pure pixel point vegetation cover degree seasonal variations and average vegetation cover degree seasonal variations is calculated.
Automatic statistical analysis for quality examination is as a result, carry out hand inspection, in conjunction with the former vegetation cover degree of pure pixel point
Seasonal variations data delete the larger curve of fractional error from the vegetation cover degree seasonal variations curve library of pure pixel.
Step 3: the preliminary NO emissions reduction based on low resolution vegetation cover degree image.
Preliminary NO emissions reduction processing is carried out to low resolution vegetation cover degree image, is obtained and high-resolution vegetation cover degree image
The consistent low resolution vegetation cover degree NO emissions reduction result images of resolution ratio.Such as low-resolution image resolution ratio is 270m, high score
The resolution ratio of resolution image is 30m, then a pixel of low-resolution image has corresponded to 81 pixels of high-definition picture,
Resolution ratio is also 30m after low-resolution image resolution ratio NO emissions reduction.
As shown in Fig. 2, the grid of low-resolution image be 250m × 250m, if the grid of high-definition picture be 25m ×
25m, then as shown in Figure 3-Figure 5, by the grid that the grid NO emissions reduction of each 250m × 250m is 100 25m × 25m.
One pixel of low-resolution image is generally mixed pixel, N number of pixel of corresponding high-definition picture, if certain is mixed
It closes pixel to be made of N number of sub-pixed mapping, the sub-pixed mapping vegetation cover degree in the mixed pixel with land use pattern is equal, then has：
In formula：
(1)fL(t) indicate that certain mixed pixel is in the vegetation cover degree image of t periods in lower resolution image；
(2) mixed pixel is made of N number of sub-pixed mapping, contains m kind land use patterns, rjIndicate jth kind soil class
The sub-pixed mapping of type accounts for the area ratio of mixed pixel；
(3)fL′(j, t) indicate t period NO emissions reduction result images some mixed pixel in, jth kind land type it is more
The vegetation cover degree image of a sub-pixed mapping is (it is assumed that the sub-pixed mapping vegetation cover degree of class is equal in the same manner in same mixed pixel；It is same mixed
It closes in pixel, the sub-pixed mapping of same ground class is likely to have multiple).
(4)fp LThe vegetation cover degree figure of (j, t) with referring to t periods and the closest jth kind of the mixed pixel pure pixel of class
Picture；
(5) d (t) is residual error, primarily determine not class variation everywhere (or according to the methods of investigation determine different land types residual errors it
Between proportionate relationship).
Formula (4) is substituted into formula (3), can be obtained：
Then：
Formula (6) is substituted into formula (4), can be obtained：
The preliminary NO emissions reduction of low resolution vegetation cover degree image, its essence is the pictures for assert low resolution vegetation cover degree image
First majority is mixed pixel, but there is also the pure pixels in part (as shown in Figure 2), are searched under the support of high-resolution land-use map
Rope obtains the pure pixel in low-resolution image, establishes pure pixel vegetation cover degree seasonal variations curve library；It is assumed that mixed pixel
In with the sub-pixed mapping vegetation cover degree phase of land use pattern when, with the pure pixel of closest same land use pattern
Vegetation cover degree carries out assignment to the vegetation cover degree of sub-pixed mapping, to obtain low resolution vegetation cover degree NO emissions reduction result images.
Step 4: the fusion interpolation of high-resolution vegetation cover degree image seasonal variations sequence.
It is raw using low resolution vegetation cover degree NO emissions reduction result images and existing high-resolution vegetation cover degree image co-registration
At the complete high-resolution vegetation cover degree image seasonal variations sequence for including 24 periods half a month, interpolation (I) is shown in Fig. 1
Schematic diagram, wherein H are high-resolution vegetation cover degree image, and L is low resolution vegetation cover degree image, and 1-24 is half a month vegetation cover degree
Image.
Period on the basis of period half a month where existing high-resolution vegetation cover degree image.
1. the fusion interpolation there are one when the benchmark period
The vegetation cover degree seasonal variations trend of high-definition picture and low-resolution image has similitude.In view of low resolution
Preliminary NO emissions reduction is low to (grid cell, raster cell size shows as equal) consistent with high-definition picture scale for rate vegetation cover degree image
Image in different resolution seems one-to-one with high resolution graphics, can be obtained by formula (1)：
fH(t)=fH(k)+(fL(t)-fL(k)) formula (8)
In formula：If the benchmark period is k (1≤k≤24), t is the period (1≤t≤24, and t ≠ k) for waiting for interpolation；fH(t) table
Show the high-resolution vegetation cover degree image of t periods；fH(k) the high-resolution vegetation cover degree image in the benchmark period (k) is indicated,
fL(k) the low resolution vegetation cover degree NO emissions reduction result images in the benchmark period (k), f are indicatedL(t) indicate low in the t periods
Resolution ratio vegetation cover degree NO emissions reduction result images.
2. fusion interpolation when two or more benchmark periods
(1) added time section to be inserted is between two benchmark periods
In two benchmark periods of arbitrary neighborhood, first benchmark period serial number k, second benchmark period serial number q
(q>k).If (q-k)=1, two benchmark periods are connected in meaning, and the high-resolution of added time to be inserted section is not present therebetween
Rate image；(if q-k)>1, then there is (q-k-1) a high-resolution vegetation map for waiting for interpolation between the two benchmark periods
Picture.Formula (9) interpolation high-resolution vegetation cover degree is pressed according to two benchmark periods, it is assumed that added time section to be inserted is by two benchmark respectively
The size of period image is in inverse relation with its time gap, then t (k<t<Q) the high-resolution vegetation cover degree of period is final
As a result be respectively interpolation result apart from inverse ratio weighted average：
fH(t)=α { fH(k)+(fL(t)-fL(k))}+β{fH(q)+(fL(t)-fL(q)) } formula (9)
In formula：α, β are respectively weight coefficient, and size is in inverse relation with the time gap to two benchmark periods；D is
Adjustable weight parameter, D values are bigger, and interpolation result is also bigger by the effect of the nearlyr benchmark period image of distance.
(2) added time section to be inserted is before the smallest sequence number benchmark period
If the benchmark period serial number k of smallest sequence number, if k=1, nothing waits for interpolation before the smallest sequence number benchmark period
Period image.For the period half a month t (t before k<K), the interpolation formula of corresponding high-resolution vegetation cover degree image is：
fH(t)=fH(k)+(fL(t)-fL(k)) formula (12)
In formula：If the benchmark period is k (1<k<24), t is the period (1≤t for waiting for interpolation<k)；fH(t) the t periods are indicated
High-resolution vegetation cover degree image；fH(k) the high-resolution vegetation cover degree image in the benchmark period (k), f are indicatedL(k) it indicates
The low resolution vegetation cover degree NO emissions reduction result images of benchmark period (k), fL(t) the low resolution vegetation in the t periods is indicated
Cover degree NO emissions reduction result images.
(3) added time section to be inserted is after the maximum serial number benchmark period
If the benchmark period serial number q of maximum serial number, if q=24, nothing waits for interpolation after the maximum serial number benchmark period
Period image.For in the subsequent period half a month t (q of q<T), the interpolation formula of corresponding high-resolution vegetation cover degree image
For：
fH(t)=fH(q)+(fL(t)-fL(q)) formula (13)
In formula：If the benchmark period is q (1<q<24), t is the period (q for waiting for interpolation<t≤24)；fH(t) the t periods are indicated
High-resolution vegetation cover degree image；fH(q) the high-resolution vegetation cover degree image in the benchmark period (q), f are indicatedL(q) it indicates
In the low resolution vegetation cover degree NO emissions reduction result images of benchmark period (q), fL(q) indicate that the low resolution in the t periods is planted
By cover degree NO emissions reduction result images.
Step 5: the statistical analysis of high-resolution vegetation cover degree image seasonal variations sequence.
(1) to low-resolution image (24 periods), high-definition picture (period or several periods), NO emissions reduction knot
Fruit image (24 periods), fusion interpolation results image are counted respectively, obtain the descriptive statistic amount of various images, and
Hierarchical statistics result.
(2) to low-resolution image (24 periods), high-definition picture (period or several periods), NO emissions reduction knot
Fruit image (24 periods), fusion interpolation results image, calculate the related coefficient between them.
(3) to low-resolution image (24 periods), high-definition picture (period or several periods), NO emissions reduction knot
Fruit image (24 periods), fusion interpolation results image, it is for statistical analysis respectively, obtain the various land use classes of various images
The descriptive statistic amount and hierarchical statistics result of type.
Embodiment of above is only used to illustrate the technical scheme of the present invention, rather than its limitations；Although with reference to aforementioned implementation
Invention is explained in detail for example, it will be understood by those of ordinary skill in the art that：It still can be to aforementioned implementation
Technical solution recorded in example is modified, or carries out equivalent replacement to which part or all technical features；And these are repaiied
Change or replaces, the range for various embodiments of the present invention technical solution that it does not separate the essence of the corresponding technical solution.
Claims (2)
1. a kind of method generating complete high-resolution vegetation cover degree image seasonal variations sequence, the method merge 24 half
Month low resolution vegetation cover degree image of period and the high-resolution vegetation cover degree image of at least one period, which is characterized in that institute
The method of stating includes the following steps：
Step 1: it is based on high-precision land-use map, it is suitable using the search of window sliding method on high-resolution vegetation cover degree image
For the pure pixel of low-resolution image, pure pixel point data base is established；
Step 2: the low resolution vegetation cover degree image based on 24 periods half a month, the vegetation cover degree season for establishing pure pixel becomes
Change curve library；
Step 3: carrying out preliminary NO emissions reduction processing to low resolution vegetation cover degree image, obtain and high-resolution vegetation cover degree figure
The consistent low resolution vegetation cover degree NO emissions reduction result images of the resolution ratio of picture；
Step 4: the high-resolution vegetation cover degree of low resolution vegetation cover degree NO emissions reduction result images and at least one period
Image carries out fusion interpolation, generates the complete high-resolution vegetation cover degree image seasonal variations sequence for including 24 periods half a month
Row；
Wherein, step 3 is specially with the vegetation cover degree of the closest pure pixel of same land use pattern to low resolution vegetation lid
The vegetation cover degree for spending sub-pixed mapping in image carries out assignment, to obtain low resolution vegetation cover degree NO emissions reduction result images, formula
It is as follows：
In formula, fL’(j, t) is indicated in some mixed pixel of t period low resolution vegetation cover degree NO emissions reduction result images, jth
The vegetation cover degree image of kind land use pattern sub-pixed mapping；fp L(j, t) refers to t periods and the closest jth kind of the mixed pixel
The vegetation cover degree image of the pure pixel of land use pattern, can be from pure pixel vegetation cover degree seasonal variations curve library according to space
It searches for obtain apart from nearest principle；fL(t) indicate that the mixed pixel is in the vegetation cover degree figure of t periods on low-resolution image
Picture；Any mixed pixel is made of N number of sub-pixed mapping, contains m kind land use patterns, the sub-pixed mapping of same land use pattern
There is multiple, rjIndicate that the sub-pixed mapping of jth kind land use pattern accounts for the area ratio of mixed pixel, fp L(i, t) refer to the t periods with
The vegetation cover degree image of i-th kind of closest pure pixel of land use pattern of the mixed pixel, can be from pure pixel vegetation cover degree
It searches for obtain according to the nearest principle of space length in seasonal variations curve library.
2. according to the method described in claim 1, it is characterized in that,
The fusion interpolation of the step 4 is in the following ways：
Period on the basis of period half a month where high-resolution vegetation cover degree image by least one period,
(1) for only there are one the fusion interpolations when benchmark period, using formula A：
fH(t)=fH(k)+(fL(t)-fL(k))
In formula A, the benchmark period is k, and 1≤k≤24, t are the period for waiting for interpolation, 1≤t≤24, and t ≠ k；fH(t) when indicating t
The high-resolution vegetation cover degree image of section；fH(k) the high-resolution vegetation cover degree image in benchmark period k, f are indicatedL(k) it indicates
In the low resolution vegetation cover degree NO emissions reduction result images of benchmark period k, fL(t) the low resolution vegetation in the t periods is indicated
Cover degree NO emissions reduction result images；
(2) for there are two or fusion interpolation when more than two benchmark period,
A. added time section to be inserted is between two benchmark periods, using formula B：
fH(t)=α { fH(k)+(fL(t)-fL(k))}+β{fH(q)+(fL(t)-fL(q))}
In formula B, first benchmark period serial number k, second benchmark period serial number q, q>K, t are the period for waiting for interpolation, k<t
<q；fH(t) the high-resolution vegetation cover degree image of t periods is indicated；fH(k)、fH(q) it is illustrated respectively in the height of benchmark period k, q
Resolution ratio vegetation cover degree image；fL(t) the low resolution vegetation cover degree NO emissions reduction result images in the t periods are indicated；fL(k)、fL
(q) the low resolution vegetation cover degree NO emissions reduction result images of benchmark period k, q are illustrated respectively in；α, β are respectively weight coefficient,
Size is in inverse relation with the time gap to two benchmark periods；D is adjustable weight parameter, and D values are bigger, interpolation result
Effect by the nearlyr benchmark period image of distance is also bigger；
B. added time section to be inserted is before the smallest sequence number benchmark period, using formula C：
fH(t)=fH(k)+(fL(t)-fL(k))
In formula C：The benchmark period be k, 1<k<24, t be the period for waiting for interpolation, 1≤t<k；fH(t) high-resolution of t periods is indicated
Vegetation cover degree image；fH(k) the high-resolution vegetation cover degree image in benchmark period k, f are indicatedL(k) it indicates benchmark period k's
Low resolution vegetation cover degree NO emissions reduction result images, fL(t) the low resolution vegetation cover degree NO emissions reduction result in the t periods is indicated
Image；
C. added time section to be inserted is after the maximum serial number benchmark period, using formula D：
fH(t)=fH(q)+(fL(t)-fL(q))
In formula D：If the benchmark period is q, 1<q<24, t be the period for waiting for interpolation, q<t≤24；fH(t) high score of t periods is indicated
Resolution vegetation cover degree image；fH(q) the high-resolution vegetation cover degree image in benchmark period q, f are indicatedL(q) it indicates in benchmark
The low resolution vegetation cover degree NO emissions reduction result images of section q, fL(q) indicate that ruler drops in the low resolution vegetation cover degree in the t periods
Spend result images.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610024214.1A CN105719236B (en) | 2016-01-14 | 2016-01-14 | The method for generating complete high-resolution vegetation cover degree image seasonal variations sequence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610024214.1A CN105719236B (en) | 2016-01-14 | 2016-01-14 | The method for generating complete high-resolution vegetation cover degree image seasonal variations sequence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105719236A CN105719236A (en) | 2016-06-29 |
| CN105719236B true CN105719236B (en) | 2018-08-14 |
Family
ID=56147839
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610024214.1A Active CN105719236B (en) | 2016-01-14 | 2016-01-14 | The method for generating complete high-resolution vegetation cover degree image seasonal variations sequence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105719236B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109741261A (en) * | 2019-01-03 | 2019-05-10 | 北京师范大学 | A kind of Remote Sensing temperature NO emissions reduction algorithm based on object-oriented window |
| CN109800315B (en) * | 2019-01-31 | 2022-11-29 | 交通运输部天津水运工程科学研究所 | Method for establishing photographic grassland coverage measurement gallery and using method |
| CN111767278B (en) * | 2020-07-09 | 2022-11-22 | 中国科学院地理科学与资源研究所 | Downscaling method and downscaling device for characteristic quantity of land ecosystem |
| CN114254802B (en) * | 2021-11-09 | 2023-09-08 | 北京师范大学 | Prediction method for vegetation coverage space-time change under climate change drive |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101488216A (en) * | 2008-01-18 | 2009-07-22 | 三洋电机株式会社 | Image processing device and method, and image sensing apparatus |
| CN101650422A (en) * | 2009-09-27 | 2010-02-17 | 北京师范大学 | Denoising method of remote sensing vegetation index time series data |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8538203B2 (en) * | 2007-07-24 | 2013-09-17 | Sharp Laboratories Of America, Inc. | Image upscaling technique |
-
2016
- 2016-01-14 CN CN201610024214.1A patent/CN105719236B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101488216A (en) * | 2008-01-18 | 2009-07-22 | 三洋电机株式会社 | Image processing device and method, and image sensing apparatus |
| CN101650422A (en) * | 2009-09-27 | 2010-02-17 | 北京师范大学 | Denoising method of remote sensing vegetation index time series data |
Non-Patent Citations (3)
| Title |
|---|
| 北京地区植被盖度提取及其分布变化研究;史世莲;《遥感技术与应用》;20141031;第29卷(第5期);866-872 * |
| 北京市典型植被的盖度季节变化监测及分析;章文波;《水利发展研究》;20131130;19-23 * |
| 植被覆盖度遥感估算方法研究进展;程红芳;《国土资源遥感》;20080315(第75期);13-18 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105719236A (en) | 2016-06-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Mertens et al. | A sub‐pixel mapping algorithm based on sub‐pixel/pixel spatial attraction models | |
| CN110163198B (en) | Table identification reconstruction method and device and storage medium | |
| CN105719236B (en) | The method for generating complete high-resolution vegetation cover degree image seasonal variations sequence | |
| CN108871286B (en) | Urban built-up area population density estimation method and system based on spatial big data cooperation | |
| CN107194872B (en) | Remote sensed image super-resolution reconstruction method based on perception of content deep learning network | |
| Johansen et al. | Automatic geographic object based mapping of streambed and riparian zone extent from LiDAR data in a temperate rural urban environment, Australia | |
| Lizarazo | Accuracy assessment of object-based image classification: another STEP | |
| DeWitt et al. | Creating high-resolution bare-earth digital elevation models (DEMs) from stereo imagery in an area of densely vegetated deciduous forest using combinations of procedures designed for lidar point cloud filtering | |
| CN106570538B (en) | Character image processing method and device | |
| CN108573222A (en) | The pedestrian image occlusion detection method for generating network is fought based on cycle | |
| CN102073867B (en) | Sorting method and device for remote sensing images | |
| CN108805825A (en) | A kind of reorientation image quality evaluating method | |
| CN104182757A (en) | Method of acquiring actual coverage area of measured target and device | |
| CN112990183A (en) | Method, system and device for extracting homonymous strokes of offline handwritten Chinese characters | |
| CN102208103A (en) | Method of image rapid fusion and evaluation | |
| Slingsby | Tilemaps for summarising multivariate geographical variation | |
| Ismail et al. | DEM derived from photogrammetric generated DSM using morphological filter | |
| CN108985154B (en) | Small-size ground object sub-pixel positioning method and system based on image concentration | |
| CN102706326B (en) | Processing method of light beam method aerial triangulation file data | |
| Aghighi et al. | Fully spatially adaptive smoothing parameter estimation for Markov random field super-resolution mapping of remotely sensed images | |
| CN109492067A (en) | Method and device for converting vector data into grid slices | |
| CN110309381A (en) | A kind of display methods of demographic data, device and equipment | |
| CN109697418A (en) | The post-processing approach that image is extracted for remote sensing image road network restored for scene | |
| Pahlavani et al. | 3D reconstruction of buildings from LiDAR data considering various types of roof structures | |
| CN114863274A (en) | Surface green net thatch cover extraction method based on deep learning |
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 |