CN104008552A - Time sequence SAR image cultivated land extraction method based on dynamic time warp - Google Patents
Time sequence SAR image cultivated land extraction method based on dynamic time warp Download PDFInfo
- Publication number
- CN104008552A CN104008552A CN201410266813.5A CN201410266813A CN104008552A CN 104008552 A CN104008552 A CN 104008552A CN 201410266813 A CN201410266813 A CN 201410266813A CN 104008552 A CN104008552 A CN 104008552A
- Authority
- CN
- China
- Prior art keywords
- pixel
- sar image
- image
- time
- sequential
- 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.)
- Granted
Links
- 238000000605 extraction Methods 0.000 title claims abstract description 32
- 230000011218 segmentation Effects 0.000 claims abstract description 27
- 238000001914 filtration Methods 0.000 claims abstract description 12
- 238000012549 training Methods 0.000 claims description 22
- 238000012937 correction Methods 0.000 claims description 11
- 238000004364 calculation method Methods 0.000 claims description 9
- 230000006870 function Effects 0.000 claims description 8
- 239000000284 extract Substances 0.000 claims description 6
- 238000003384 imaging method Methods 0.000 claims description 6
- 238000011160 research Methods 0.000 claims description 5
- 230000019771 cognition Effects 0.000 claims description 2
- 238000000034 method Methods 0.000 abstract description 20
- 238000004519 manufacturing process Methods 0.000 abstract description 4
- 238000005259 measurement Methods 0.000 abstract description 3
- 238000005192 partition Methods 0.000 abstract 1
- 239000011159 matrix material Substances 0.000 description 15
- 238000005452 bending Methods 0.000 description 12
- 241000196324 Embryophyta Species 0.000 description 11
- 238000005516 engineering process Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 3
- 230000001932 seasonal effect Effects 0.000 description 3
- 230000002123 temporal effect Effects 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 230000007812 deficiency Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000013519 translation Methods 0.000 description 2
- 241000132092 Aster Species 0.000 description 1
- 101100423891 Caenorhabditis elegans qars-1 gene Proteins 0.000 description 1
- 241001269238 Data Species 0.000 description 1
- 240000007594 Oryza sativa Species 0.000 description 1
- 235000007164 Oryza sativa Nutrition 0.000 description 1
- 240000008042 Zea mays Species 0.000 description 1
- 235000005824 Zea mays ssp. parviglumis Nutrition 0.000 description 1
- 235000002017 Zea mays subsp mays Nutrition 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 235000013339 cereals Nutrition 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 235000005822 corn Nutrition 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000012854 evaluation process Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 230000012010 growth Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000010287 polarization Effects 0.000 description 1
- 230000001850 reproductive effect Effects 0.000 description 1
- 235000009566 rice Nutrition 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 239000002689 soil Substances 0.000 description 1
- 238000009331 sowing Methods 0.000 description 1
- 238000003971 tillage Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000009105 vegetative growth Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Landscapes
- Image Analysis (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention relates to a time sequence SAR image cultivated land extraction method based on dynamic time warp. The method includes the steps that a time sequence SAR image is established; a cultivated land reference time sequence is extracted; the dynamic time warp distance between a time sequence of image elements to be classified and the cultivated land reference time sequence is calculated; result threshold segmentation is calculated, and image elements to be classified are classified into cultivated land image elements and uncultivated land image elements; spatial filtering is conducted on a segmentation result, the isolated cultivated land image elements are filtered out, gaps between contiguous cultivated land are filled up, and then the final extraction result of the cultivated land is acquired. The specific 'time axis warp' phenomenon of the cultivated land time sequence is taken into consideration, the dynamic time warp (DTW) distance serves as a similarity measurement standard, partition of the cultivated land image elements and the uncultivated land image elements is achieved, the problem that a traditional method cannot adapt to time sequence similarity measurement with time shaft distortion is solved, and extraction accuracy of the cultivated land is improved. The method is high in adaptability, extraction accuracy can reach more than 82%, and the requirements of actual production can be met.
Description
Technical field
The present invention relates to a kind of remote sensing image information extracting method, particularly a kind of sequential SAR image plant extraction method based on dynamic time warping.
Background technology
Arable land is the pacing items of grain-production, is the material base of guaranteeing grain security, grasps real-time farmland information significant to the formulation of national agricultural policy.Remote sensing technology, through the development of decades, has formed the earth observation network of " star---sky---ground " one, and has produced magnanimity heap volume data, therefore uses remote sensing technology to carry out the rapid extraction field of farmland information on a large scale and has original advantage.
The crop of arable land surface attachment has the phenology feature fairly obvious with seasonal variations, different growth periods (as sowing time, vegetative growth phase, reproductive stage and maturity stage etc.) has different performances on remote sensing image, therefore uses the remote sensing image of single timing node often can not effectively extract farmland information.Time series remote sensing image (hereinafter to be referred as sequential remote sensing image) is satellite sensor carries out observation station's acquisition sequence of values to areal different time, by building sequential remote sensing image, can identify the temporal change characteristic of observation value, and utilize this feature to carry out effective extraction of farmland information.But be subject to the impact of the factors such as observation condition is changeable, crops phenology differs greatly, still there are many problems in the work that present stage utilizes the inverting of sequential remote sensing image to plough, roughly can be summarized as two aspects:
1) deficiency of the remotely-sensed data that uses.The vegetation index such as NDVI, EVI product (Sakamoto, 2005 that the current great majority of the sequential remotely-sensed data for the inverting of ploughing build for image datas such as SPOT/VEGETATION, NOAA/AVHRR, MODIS; Wardlow, 2008; Dheeravath, 2010), they have higher temporal resolution, but spatial resolution very low (the highest only 250m), compound pixel problem is very serious; Meanwhile, being often subject to cloud layer and blocking in the optical observation process of satellite sensor, there is much noise in the time series obtaining; In addition, NDVI itself can be subject to the impact that the factor such as Soil Background and water cut is larger, and easily reach capacity in territory, high vegetation-covered area (Baush, 1993; Tang Yi, 2006).Synthetic-aperture radar (SAR) is active sensor, and the microwave of transmitting can penetrate cloud layer, can not be subject to weather effect to complete earth observation task by round-the-clock.Along with the development of technology, the spatial resolution of SAR sensor and temporal resolution have had great lifting, the ENVISAT ASAR data space resolution of European Space Agency reaches as high as 10m, China HJ-1C star SAR data space resolution reaches as high as 4m, and the SAR data space resolution of German TerraSAR-X, Canadian RADARSAT-2 reaches as high as 1m; Meanwhile, the characteristic that SAR can side-looking imaging has ensured the high time resolution of image, and many satellite sensors can provide the SAR image data (TerraSAR-X can realize and heavily visiting for 2 days the area in the whole world 95%) that the time interval is less than 10 days.Although but SAR image range of application is more and more wider, the rare research that is built time series and carry out farmland information extraction.
2) deficiency of the method that adopts.Utilizing sequential remote sensing image to extract the gordian technique of ploughing is Time Series Similarity tolerance, i.e. the similarity degree of qualitative assessment pixel time series to be sorted and arable land reference time array.Ploughed in large region, the crops kind of its plantation is various, the start time in phenological period of variety classes crops differs, and continues length difference, causes arable land time series on time shaft, to have certain translation and bending (Calendar Shift).And current method of carrying out remote sensing Time Series Similarity tolerance can be divided three classes substantially: Furthest Neighbor (Euclidean distance, mahalanobis distance etc.), algebraic approach (make poor method, make commercial law etc.) and correlation method (Pearson correlation coefficient, Spearman related coefficient etc.), they all cannot be applicable to the time series of " time shaft displacement ".Dynamic time warping (Dynamic time warping, DTW) be a time series similarity algorithm conventional in computer science, it is Automatic-searching generation translation and bending timing node to a certain extent, and the time shaft of sequence is carried out to dynamic bending to realize Adaptive matching, thereby calculate dynamic bending distance, as two seasonal effect in time series similarities, extremely be suitable for from mechanism the seasonal effect in time series similarity analysis of ploughing, but research accordingly or application are actually rare.
Summary of the invention
The technical problem to be solved in the present invention is: overcome prior art shortcoming, propose a kind of sequential SAR image plant extraction method based on dynamic time warping.
The sequential SAR image plant extraction method that the present invention is based on dynamic time warping, comprises the following steps:
The first step, sequential SAR image build---the original SAR image data in the research period of study area is carried out to pre-service and registration, and according to time order and function sequential build sequential SAR image;
Second step, arable land reference time array extract---and the some arable lands pixel in selection study area is as training sample, for each the width image in described sequential SAR image, plough in calculation training sample respectively average and the standard deviation of backscattering coefficient of pixel, if described standard deviation sum is less than the reliability thresholds of setting, utilize described average to build arable land reference time array according to the imaging time of corresponding image; Otherwise, reselect the training sample duplicate step of laying equal stress on, until standard deviation is less than reliability thresholds;
The 3rd step, time series similarity calculate---for each pixel to be sorted, build its backscattering coefficient time series, and calculate the dynamic time warping distance between described backscattering coefficient time series and arable land reference time array, as the similarity between two time serieses;
The 4th step, result of calculation Threshold segmentation---for similarity result of calculation, use predefined segmentation threshold to carry out binary segmentation, the correspondence pixel to be sorted that dynamic time warping distance is greater than to segmentation threshold classifies as bare place pixel, all the other are classified as arable land pixel, obtain Threshold segmentation result;
The 5th step, segmentation result airspace filter---usage space operation operator carries out airspace filter to described Threshold segmentation result, and the arable land pixel that filtering is isolated is filled up the gap between ploughing in flakes, obtains the final extraction result of ploughing.
The sequential SAR image plant extraction method that the present invention is based on dynamic time warping, also has following improvement:
1, in the described first step, pre-service at least comprises orbit parameter renewal, radiant correction, topographic correction, geometry correction and noise-removed filtering.
2, in the described first step, the registration error of image is not more than 0.5 pixel.
3, in described second step, the arable land pixel of composing training sample extracts by artificial cognition, and the arable land pixel quantity in training sample is not less than 9.
4, the reliability thresholds Th1 in second step calculates acquisition by formula Th1=N × c × R, wherein, N is the image quantity in sequential SAR image, c is the coefficient of ratio, span is [0.3,0.5] maximal value of backscattering coefficient average of pixel of ploughing in the training sample that, R is all images and minimum value poor.
5,, in described the 3rd step, utilize the backscattering coefficient of pixel to be sorted in each width image to build the backscattering coefficient time series of pixel to be sorted according to the imaging time of corresponding image.
6,, in described the 4th step, segmentation threshold Th2 is by formula Th2=c
1× Σ calculates acquisition, wherein, and c
1for adjustment factor, its span is that 0.3～1, Σ is the summation of all standard deviations in second step.
7, in described the 5th step, spatial operator is the one in mode filter operator, mathematical morphology open-close operation, medium filtering operator, and the size size of spatial operator obtains according to following formula:
Wherein, re is the spatial resolution of SAR image, and ps is the Pixel size of SAR image, and int ' () is round function.
Beneficial effect of the present invention is as follows:
1), the present invention uses high-resolution synthetic-aperture radar (SAR) backscattering coefficient to build time series, it is even higher that the spatial resolution of plant extraction result can reach 30 meters (as satellite sensors such as ENVISAT ASAR, ERS-1/2SAR), and the products such as the more traditional MODIS/NDVI time series of the discrimination degree in space (250 meters of high spatial resolutions) have been improved more than 8 times.
2), consider arable land time series distinctive " time shaft bending " phenomenon, the present invention uses dynamic time warping distance (DTW) as similarity measurement standard, thereby realize the division of arable land pixel and bare place pixel, solve the Time Series Similarity that classic method cannot the distortion of adaptation time axle and measured this problem, improved theoretically the extraction accuracy of ploughing.
3), the present invention is directed to training sample and proposed reliability thresholds, guarantee the reliability of training sample, and cut apart for dynamic distance the choosing method that has proposed segmentation threshold, guarantee the precision of final plant extraction.
4), the inventive method strong adaptability, be proven, use the inventive method extract plough, extraction accuracy can reach more than 82%, can meet the needs of actual production, compares traditional plant extraction method and has stronger practicality.
Brief description of the drawings
Below in conjunction with accompanying drawing, the present invention is further illustrated.
Fig. 1 is test block bitmap area in the embodiment of the present invention.
Fig. 2 is the reference time array schematic diagram of ploughing in the embodiment of the present invention.
Fig. 3-a is " time shaft bending " time series schematic diagram.
Fig. 3-b is dynamic bending path schematic diagram in the embodiment of the present invention.
Fig. 4 is study area plant extraction result figure in the embodiment of the present invention (dark expression ploughs).
Embodiment
Elaborate with reference to the accompanying drawings the present invention below, make technology path of the present invention and operation steps more clear.
It is study area that the present embodiment is selected west area, Jiawang District, Xuzhou City of Jiangsu Province, and longitude scope is 117.28 °～117.56 ° of east longitudes, and latitude scope is 34.30 °～34.54 ° of north latitude, area 614.57km
2(seeing Fig. 1).This genus warm temperate zone, study area half moistening monsoon climate, makes a clear distinction between the four seasons, 14.2 DEG C of average temperatures of the whole year, average rainfall 802.4mm, in study area tillage method for yielding two crops a year, plant May to October one season rice or summer corn, May October to next year Winter Wheat Planted.
Experimental data is the totally 11 scape ENVISAT ASAR image model data in year Dec in Dec, 2009 to 2010 that cover this region, and L1 level product, has passed through system geometry correction, 30 meters of spatial resolutions, 20.25 meters of Pixel sizes, VV polarization, every scape image capturing time is in table 1.
Table 1 is tested the ENVISAT ASAR image capturing time used
The sequential SAR image plant extraction method of the embodiment of the present invention based on dynamic time warping, comprises the following steps:
The first step, sequential SAR image build---the original SAR image data in the research period of study area is carried out to pre-service and registration, and according to time order and function sequential build sequential SAR image.
First use the NEST software that European Space Agency provides successively original SAR image data to be carried out to the pretreatment operation such as orbit parameter renewal, radiant correction, topographic correction, geometry correction and noise-removed filtering, wherein topographic correction uses Range-Doppler model and ASTER GDEM data, and noise-removed filtering adopts the Enhanced Lee filter operator of 3*3 size.Then use local auto-adaptive method for registering (Cheng, L.et al., Generation of Pixel-Level SAR Image Time Series Using a Locally Adaptive Matching Technique, 2014) all images are carried out to registration, the registration error of image is not more than 0.5 pixel, finally, build sequential SAR image according to the sequencing of image capturing time time.
Second step, arable land reference time array extract---and the some arable lands pixel in selection study area is as training sample, for each the width image in described sequential SAR image, plough in calculation training sample respectively average and the standard deviation of backscattering coefficient of pixel, if described standard deviation sum is less than the reliability thresholds of setting, utilize described average to build arable land reference time array according to the imaging time of corresponding image; Otherwise, reselect the training sample duplicate step of laying equal stress on, until standard deviation is less than reliability thresholds.
In this step, open the sequential SAR image building in ENVI, contrast Google Earth history image is chosen arable land training sample in sequential SAR image, according to the reliability thresholds Th1 setting, training sample is carried out to reliability evaluation.
Reliability thresholds Th1 calculates acquisition by formula Th1=N × c × R, wherein, N is the image quantity in sequential SAR image, N=11 in this example, c is the coefficient of ratio, and span is [0.3,0.5], in this example, get the poor of the maximal value of backscattering coefficient average of the pixel of ploughing in the training sample that 0.3, R is all images and minimum value.
To the training sample arable land pixel backscattering coefficient standard deviation summation of all images, if when standard deviation sum is greater than reliability thresholds Th1, need reselect sample and repeat above-mentioned reliability evaluation process, until standard deviation sum is not more than Th1, each timing node average is arranged sequentially, obtain the reference time array of ploughing.
The present embodiment has been chosen 1000 of study areas arable land pixel (in order to ensure extraction accuracy, minimum can not lower than 9), the arable land time series S=[-9.762 obtaining altogether,-13.898,-11.319 ,-12.304 ,-15.811,-14.235,-11.285 ,-8.761 ,-8.753,-9.585 ,-11.945] (referring to Fig. 2).The backscattering coefficient standard deviation of every width image training sample sorts according to time order and function, STD=[3.498,2.144,3.169,2.640,2.939,2.813,2.592,2.716,3.021,3.134,3.165].
The 3rd step, time series similarity calculate---for each pixel to be sorted, build its backscattering coefficient time series Q, and calculate the dynamic time warping distance between described backscattering coefficient time series Q and arable land reference time array S, as the similarity between two time serieses.
The present embodiment uses Microsoft Visual Studio2010 and C# programming, calculates the dynamic bending distance of pixel backscattering coefficient time series Q to be sorted and arable land reference time array S.
The similar time series of time shaft bending (is shown in that the dynamic time warping distance between Fig. 3-a) calculates, belongs to existing ripe algorithm, provide a kind of algorithm in this example, dynamic time warping distance is described.But in the present invention, related dynamic bending distance is not limited to this algorithm acquisition.
Computing time, sequence Q (supposed Q=q
1, q
2..., q
n) (suppose S=s with time series S
1, s
2..., s
m) dynamic bending distance (subscript n in this example=m), concrete steps are as follows:
A1), calculate the Distance matrix D of Q and S, the element d (i, j) of Distance matrix D=| q
i-s
j|, that is:
A2), the definition minimum cost Matrix C identical with Distance matrix D dimension, value c (the i of element in Matrix C, j) be illustrated in Distance matrix D and locate from initial point (1,1) the minimal path cost of locating through a crooked route in-position (i, j);
A3), according to the value of each element in Distance matrix D compute matrix C, wherein last element c (n, m) is the dynamic bending distance of time series Q and time series S.
Above-mentioned steps a2) in definition crooked route be the ordered set P=p of one group of continuous element in Distance matrix D
1, p
2, p
3..., p
k(see Fig. 3-b), and need meet the following conditions:
I) boundary condition: p
1=d (1,1), p
k=d (n, m), the two ends element of the clinodiagonal that the start-stop element in this path is distance matrix;
II) continuity: if there is p
k=d (i, j) and p
k-1=d (i', j'), must have | i-i'|≤1 and | j-j'|≤1.Be that in path, adjacent element must be continuously;
III) monotonicity: if there is p
k=d (i, j) and p
k-1=d (i', j'), must have i-i'>=0 and j-j'>=0.Even path P is by position (i, j), it must at least pass through (i-1, j), (i, and (i-1 j-1), j-1) in three positions, a rear position in guarantee path remains on line direction on the basis of last position and column direction does not subtract;
IV) boundedness: max (m, n)≤k≤m+n-1.This path the matrix element number of process there is upper and lower bound.
Above-mentioned steps a3) in to calculate the method for each element value in minimum cost Matrix C as follows:
B1), origin position in Distance matrix D (1,1) is located to element value and is assigned to the element of origin position in minimum cost Matrix C, i.e. c (1,1)=d (1,1);
B2), calculate the first row element value, c (1, j)=c (1, j-1)+d (1, j), wherein j=2,3 ..., m;
B3), calculate first row element value, c (i, 1)=c (i-1,1)+d (i, 1), wherein i=2,3 ..., n;
B4), all the other element values of compute matrix, computing formula is as follows:
c(i,j)＝min(c(i-1,j),c(i,j-1),c(i-1,j-1))+d(i,j)
In formula, c (i, j) represents the element of position (i, j), and min () is minimum value value function, i=2, and 3 ..., n, j=2,3 ..., m.
The 4th step, result of calculation Threshold segmentation---for similarity result of calculation, use predefined segmentation threshold to carry out binary segmentation, the correspondence pixel to be sorted that dynamic time warping distance is greater than to segmentation threshold classifies as bare place pixel, all the other are classified as arable land pixel, obtain Threshold segmentation result.
In the present embodiment, segmentation threshold Th2 is by formula Th2=c
1× Σ calculates acquisition, wherein, and c
1for adjustment factor, its span is 0.3～1, c in this example
1get the summation that 0.3, Σ is all standard deviations in second step, Σ=31.832 in this example.So calculating segmentation threshold Th2 is 9.55.Finally the segmentation result of acquisition can represent with bianry image, and for example, 1 representative is ploughed, and 0 represents bare place.
The 5th step, segmentation result airspace filter---usage space operation operator carries out airspace filter to described Threshold segmentation result, and the arable land pixel that filtering is isolated is filled up the gap between ploughing in flakes, obtains the final extraction result in arable land, sees Fig. 4.In Fig. 4, dark pixel is the arable land of extracting.
In the 4th step, use mode filter operator (Marjority) to carry out the airspace filter (can use in addition mathematical morphology open-close operation, medium filtering operator etc.) of segmentation result, the size size of spatial operator obtains according to following formula:
Wherein, re is the spatial resolution of SAR image, is that 30, ps is the Pixel size of SAR image herein, is that 20.25, int ' () is round function herein.Finally calculate spatial filtering operator and be of a size of 3*3.
Utilize ground truth to the extract precision test result of ploughing to be: producer's precision 84.37%, user's precision 82.19%, shows that the inventive method effectively carries out the farmland information of higher spatial resolution and extract, and precision is higher, meets need of production.
In addition to the implementation, the present invention can also have other embodiments.All employings are equal to the technical scheme of replacement or equivalent transformation formation, all drop on the protection domain of requirement of the present invention.
Claims (8)
1. the sequential SAR image plant extraction method based on dynamic time warping, comprises the following steps:
The first step, sequential SAR image build---the original SAR image data in the research period of study area is carried out to pre-service and registration, and according to time order and function sequential build sequential SAR image;
Second step, arable land reference time array extract---and the some arable lands pixel in selection study area is as training sample, for each the width image in described sequential SAR image, plough in calculation training sample respectively average and the standard deviation of backscattering coefficient of pixel, if described standard deviation sum is less than the reliability thresholds of setting, utilize described average to build arable land reference time array according to the imaging time of corresponding image; Otherwise, reselect the training sample duplicate step of laying equal stress on, until standard deviation is less than reliability thresholds;
The 3rd step, time series similarity calculate---for each pixel to be sorted, build its backscattering coefficient time series, and calculate the dynamic time warping distance between described backscattering coefficient time series and arable land reference time array, as the similarity between two time serieses;
The 4th step, result of calculation Threshold segmentation---for similarity result of calculation, use default segmentation threshold to carry out binary segmentation, the correspondence pixel to be sorted that dynamic time warping distance is greater than to segmentation threshold classifies as bare place pixel, and all the other are classified as arable land pixel, obtains Threshold segmentation result;
The 5th step, segmentation result airspace filter---usage space operation operator carries out airspace filter to described Threshold segmentation result, and the arable land pixel that filtering is isolated is filled up the gap between ploughing in flakes, obtains the final extraction result of ploughing.
2. the sequential SAR image plant extraction method based on dynamic time warping according to claim 1, is characterized in that: in the described first step, pre-service at least comprises orbit parameter renewal, radiant correction, topographic correction, geometry correction and noise-removed filtering.
3. the sequential SAR image plant extraction method based on dynamic time warping according to claim 1, is characterized in that: in the described first step, the registration error of image is not more than 0.5 pixel.
4. the sequential SAR image plant extraction method based on dynamic time warping according to claim 1, it is characterized in that: in described second step, the arable land pixel of composing training sample extracts by artificial cognition, and the arable land pixel quantity in training sample is not less than 9.
5. the sequential SAR image plant extraction method based on dynamic time warping according to claim 1, it is characterized in that: the reliability thresholds Th1 in second step calculates acquisition by formula Th1=N × c × R, wherein, N is the image quantity in sequential SAR image, c is the coefficient of ratio, span is [0.3,0.5], the maximal value of the backscattering coefficient average of the pixel of ploughing in the training sample that R is all images and minimum value poor.
6. the sequential SAR image plant extraction method based on dynamic time warping according to claim 1, it is characterized in that: in described the 3rd step, utilize the backscattering coefficient of pixel to be sorted in each width image to build the backscattering coefficient time series of pixel to be sorted according to the imaging time of corresponding image.
7. the sequential SAR image plant extraction method based on dynamic time warping according to claim 1, is characterized in that: in described the 4th step, segmentation threshold Th2 is by formula Th2=c
1× Σ calculates acquisition, wherein, and c
1for adjustment factor, its span is that 0.3～1, Σ is the summation of all standard deviations in second step.
8. the sequential SAR image plant extraction method based on dynamic time warping according to claim 1, it is characterized in that: in described the 5th step, spatial operator is the one in mode filter operator, mathematical morphology open-close operation, medium filtering operator, and the size size of spatial operator obtains according to following formula:
Wherein, re is the spatial resolution of SAR image, and ps is the Pixel size of SAR image, and int ' () is round function.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410266813.5A CN104008552B (en) | 2014-06-16 | 2014-06-16 | Time sequence SAR image cultivated land extraction method based on dynamic time warp |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410266813.5A CN104008552B (en) | 2014-06-16 | 2014-06-16 | Time sequence SAR image cultivated land extraction method based on dynamic time warp |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104008552A true CN104008552A (en) | 2014-08-27 |
| CN104008552B CN104008552B (en) | 2017-01-25 |
Family
ID=51369192
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410266813.5A Expired - Fee Related CN104008552B (en) | 2014-06-16 | 2014-06-16 | Time sequence SAR image cultivated land extraction method based on dynamic time warp |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104008552B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106446555A (en) * | 2016-09-29 | 2017-02-22 | 福州大学 | Vegetation change occurrence time detection method based on time series similarity |
| CN107064932A (en) * | 2017-02-28 | 2017-08-18 | 华南农业大学 | A kind of construction land change detection method based on time series SAR remote sensing images |
| CN109471106A (en) * | 2018-11-22 | 2019-03-15 | 上海海事大学 | In conjunction with the SAR internal wave of ocean fringe counting method method of clustering and frontier tracing method |
| CN109840553A (en) * | 2019-01-17 | 2019-06-04 | 苏州中科天启遥感科技有限公司 | The extracting method and system, storage medium, electronic equipment for agrotype of ploughing |
| CN110147843A (en) * | 2019-05-22 | 2019-08-20 | 哈尔滨工程大学 | Voice Time Series Similar measure based on metric learning |
| CN111666914A (en) * | 2020-06-15 | 2020-09-15 | 中国科学院地理科学与资源研究所 | Cultivated land identification method, system, equipment and storage medium based on distance between curves |
| CN114202691A (en) * | 2021-12-09 | 2022-03-18 | 广东省科学院广州地理研究所 | Rice identification method, device and equipment based on remote sensing image |
| CN114627148A (en) * | 2021-07-05 | 2022-06-14 | 广东省科学院广州地理研究所 | Coastal zone aquaculture water body object extraction method and device based on microwave remote sensing |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090208117A1 (en) * | 2008-02-20 | 2009-08-20 | Juliet Eichen | Fast block matching in digital images |
| CN102073869A (en) * | 2010-12-27 | 2011-05-25 | 中国农业大学 | Point location data and remote sensing image data-based regional farmland quality monitoring method |
| CN102645350A (en) * | 2012-03-08 | 2012-08-22 | 浙江大学 | Soil-sampling method based on high-resolution satellite remote sensing data |
| CN103440489A (en) * | 2013-09-16 | 2013-12-11 | 南京大学 | Water body extraction method based on pixel-level SAR (synthetic aperture radar) image time sequence similarity analysis |
| CN103440490A (en) * | 2013-09-16 | 2013-12-11 | 南京大学 | Urban impervious surface extraction method under SAR (Synthetic Aperture Radar) image space-time similarity analysis |
-
2014
- 2014-06-16 CN CN201410266813.5A patent/CN104008552B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090208117A1 (en) * | 2008-02-20 | 2009-08-20 | Juliet Eichen | Fast block matching in digital images |
| CN102073869A (en) * | 2010-12-27 | 2011-05-25 | 中国农业大学 | Point location data and remote sensing image data-based regional farmland quality monitoring method |
| CN102645350A (en) * | 2012-03-08 | 2012-08-22 | 浙江大学 | Soil-sampling method based on high-resolution satellite remote sensing data |
| CN103440489A (en) * | 2013-09-16 | 2013-12-11 | 南京大学 | Water body extraction method based on pixel-level SAR (synthetic aperture radar) image time sequence similarity analysis |
| CN103440490A (en) * | 2013-09-16 | 2013-12-11 | 南京大学 | Urban impervious surface extraction method under SAR (Synthetic Aperture Radar) image space-time similarity analysis |
Non-Patent Citations (3)
| Title |
|---|
| 徐超 等: "基于多时相TM图像的耕地信息提取", 《国土资源遥感》 * |
| 李向军: "耕地变化的影像比例化分析", 《遥感信息》 * |
| 王亚飞 等: "像素级SAR影像时间序列的建模方法研究", 《地理与地理信息科学》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106446555A (en) * | 2016-09-29 | 2017-02-22 | 福州大学 | Vegetation change occurrence time detection method based on time series similarity |
| CN106446555B (en) * | 2016-09-29 | 2019-01-22 | 福州大学 | Coupling relationship time of origin detection method based on sequential similarity |
| CN107064932A (en) * | 2017-02-28 | 2017-08-18 | 华南农业大学 | A kind of construction land change detection method based on time series SAR remote sensing images |
| CN109471106A (en) * | 2018-11-22 | 2019-03-15 | 上海海事大学 | In conjunction with the SAR internal wave of ocean fringe counting method method of clustering and frontier tracing method |
| CN109840553A (en) * | 2019-01-17 | 2019-06-04 | 苏州中科天启遥感科技有限公司 | The extracting method and system, storage medium, electronic equipment for agrotype of ploughing |
| CN110147843A (en) * | 2019-05-22 | 2019-08-20 | 哈尔滨工程大学 | Voice Time Series Similar measure based on metric learning |
| CN111666914A (en) * | 2020-06-15 | 2020-09-15 | 中国科学院地理科学与资源研究所 | Cultivated land identification method, system, equipment and storage medium based on distance between curves |
| CN111666914B (en) * | 2020-06-15 | 2022-05-10 | 中国科学院地理科学与资源研究所 | Cultivated land identification method, system, equipment and storage medium based on distance between curves |
| CN114627148A (en) * | 2021-07-05 | 2022-06-14 | 广东省科学院广州地理研究所 | Coastal zone aquaculture water body object extraction method and device based on microwave remote sensing |
| CN114627148B (en) * | 2021-07-05 | 2024-01-02 | 广东省科学院广州地理研究所 | Method and device for extracting coastal zone aquaculture water body object based on microwave remote sensing |
| CN114202691A (en) * | 2021-12-09 | 2022-03-18 | 广东省科学院广州地理研究所 | Rice identification method, device and equipment based on remote sensing image |
| CN114202691B (en) * | 2021-12-09 | 2022-10-04 | 广东省科学院广州地理研究所 | Rice identification method, device and equipment based on remote sensing image |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104008552B (en) | 2017-01-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104008552B (en) | Time sequence SAR image cultivated land extraction method based on dynamic time warp | |
| Clauss et al. | Estimating rice production in the Mekong Delta, Vietnam, utilizing time series of Sentinel-1 SAR data | |
| Kolassa et al. | Estimating surface soil moisture from SMAP observations using a Neural Network technique | |
| Sun et al. | Reconstructing daily clear-sky land surface temperature for cloudy regions from MODIS data | |
| Omran | Detection of land-use and surface temperature change at different resolutions | |
| Dall'Amico et al. | First results of SMOS soil moisture validation in the upper Danube catchment | |
| de Wit et al. | Estimating regional winter wheat yield with WOFOST through the assimilation of green area index retrieved from MODIS observations | |
| Joshi et al. | Vegetation cover mapping in India using multi-temporal IRS Wide Field Sensor (WiFS) data | |
| CN106908415B (en) | A kind of big region crops time of infertility Soil Moisture Monitoring method based on amendment NDVI time series | |
| Pasolli et al. | Estimation of soil moisture in mountain areas using SVR technique applied to multiscale active radar images at C-band | |
| Dall'Olmo et al. | Monitoring phenological cycles of desert ecosystems using NDVI and LST data derived from NOAA-AVHRR imagery | |
| Meier et al. | How do sea-ice concentrations from operational data compare with passive microwave estimates? Implications for improved model evaluations and forecasting | |
| CN114387516B (en) | Single-season rice SAR (synthetic aperture radar) identification method for small and medium-sized fields in complex terrain environment | |
| Torbick et al. | Integrating SAR and optical imagery for regional mapping of paddy rice attributes in the Poyang Lake Watershed, China | |
| Hereher | The status of Egypt’s agricultural lands using MODIS Aqua data | |
| CN113361191A (en) | Pixel scale winter wheat yield remote sensing estimation method based on multi-scenario simulation | |
| Setiyono et al. | Synthetic Aperture Radar (SAR)-based paddy rice monitoring system: Development and application in key rice producing areas in Tropical Asia | |
| Zhu et al. | Agricultural irrigation impacts on land surface characteristics detected from satellite data products in Jilin Province, China | |
| Liu et al. | Optimal MODIS data processing for accurate multi-year paddy rice area mapping in China | |
| Li et al. | Mapping rice cropping systems using Landsat-derived renormalized index of normalized difference vegetation index (RNDVI) in the Poyang Lake Region, China | |
| CN112418016A (en) | SAR-based irrigation information extraction method and device | |
| Usowicz et al. | Comparison of surface soil moisture from SMOS satellite and ground measurements | |
| Kadhim | Monitoring land cover change using remote sensing and GIS techniques: a case study of Al-Dalmaj marsh, Iraq | |
| Xie | Combining CERES-Wheat model, Sentinel-2 data, and deep learning method for winter wheat yield estimation | |
| Tran et al. | Monitoring vegetation cover changes by sentinel-1 radar images using random forest classification method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB03 | Change of inventor or designer information |
Inventor after: Zhong Lishan Inventor after: Pan Hang Inventor after: Zhu Mingyuan Inventor after: Li Manchun Inventor after: Cheng Liang Inventor after: Wang Jiechen Inventor after: Wang Yafei Inventor after: Jiang Penghui Inventor after: Sun Yuefan Inventor after: Ma Lei Inventor before: Zhong Lishan Inventor before: Pan Hang Inventor before: Zhu Mingyuan Inventor before: Li Manchun Inventor before: Cheng Liang Inventor before: Wang Jiechen Inventor before: Wang Yafei Inventor before: Jiang Penghui Inventor before: Sun Yuefan Inventor before: Ma Lei |
|
| COR | Change of bibliographic data | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170125 Termination date: 20170616 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |