CN109978855A - A kind of method for detecting change of remote sensing image and device - Google Patents
A kind of method for detecting change of remote sensing image and device Download PDFInfo
- Publication number
- CN109978855A CN109978855A CN201910232115.6A CN201910232115A CN109978855A CN 109978855 A CN109978855 A CN 109978855A CN 201910232115 A CN201910232115 A CN 201910232115A CN 109978855 A CN109978855 A CN 109978855A
- Authority
- CN
- China
- Prior art keywords
- remote sensing
- sensing images
- difference image
- energy coefficient
- frequency domain
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 81
- 230000008859 change Effects 0.000 title claims abstract description 69
- 238000012360 testing method Methods 0.000 claims abstract description 49
- 230000009466 transformation Effects 0.000 claims description 48
- 238000001514 detection method Methods 0.000 claims description 38
- 238000010586 diagram Methods 0.000 claims description 25
- 238000005070 sampling Methods 0.000 claims description 22
- 238000010008 shearing Methods 0.000 claims description 19
- 230000003044 adaptive effect Effects 0.000 claims description 8
- 238000013528 artificial neural network Methods 0.000 claims description 8
- 238000006243 chemical reaction Methods 0.000 claims description 7
- 238000004891 communication Methods 0.000 claims description 3
- 235000007926 Craterellus fallax Nutrition 0.000 abstract description 4
- 240000007175 Datura inoxia Species 0.000 abstract 1
- 230000008569 process Effects 0.000 description 14
- 238000012545 processing Methods 0.000 description 12
- 238000004422 calculation algorithm Methods 0.000 description 11
- 238000004590 computer program Methods 0.000 description 9
- 230000000694 effects Effects 0.000 description 7
- 238000001914 filtration Methods 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000000007 visual effect Effects 0.000 description 4
- 244000086443 Craterellus fallax Species 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000036039 immunity Effects 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 238000000844 transformation Methods 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20172—Image enhancement details
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
Abstract
The embodiment of the present invention provides a kind of method for detecting change of remote sensing image and device, the described method includes: the first remote sensing images of acquisition and the second remote sensing images, first remote sensing images and second remote sensing images are the remote sensing images of the same area shot in different time；According to first remote sensing images and second remote sensing images, difference image is generated；Based on the frequency domain information of the difference image, the noise in the difference image is removed；Two classification are carried out to the difference image after removal noise, generate variation testing result, the variation testing result is used to indicate the difference between first remote sensing images and second remote sensing images.Due to pixel in difference image information in frequency domain compared with horn of plenty, using the frequency domain information of difference image, the noise in difference image can be removed, enabling difference image really indicates the variation of remote sensing images, the accuracy of difference image classification is improved, and then improves the accuracy of variation testing result.
Description
Technical field
The present invention relates to digital image processing techniques field more particularly to a kind of method for detecting change of remote sensing image and dress
It sets.
Background technique
Remote Sensing Imagery Change Detection refers to that the two width remote sensing images obtained to the same area different time are analyzed, to obtain
The process for taking the region earth's surface variation characteristic is one of the main direction of development of current Remote Sensing Data Processing technology.
Currently, the two width remote sensing images obtained to the same area different time, which are changed detection, usually first generates difference
Image (Difference Image, DI) then carries out two classification to difference image, i.e., to each pixel in difference image
Classify, be divided into variation class or do not change class and obtain the variation of two width remote sensing images further according to the division result of each pixel
Testing result.
But there is also noises due to there are noise, making in the difference image generated in remote sensing images, so that difference
Image can not really represent the variation of remote sensing images, reduce the accuracy of difference image classification, and then variation is caused to be examined
Result is surveyed to be inaccurate.
Summary of the invention
In view of the above problems, the purpose of the embodiment of the present invention is that providing a kind of method for detecting change of remote sensing image and device,
It is intended to remove the noise in difference image, difference image is enable really to represent the variation of remote sensing images, improve difference diagram
As the accuracy of classification, and then improve the accuracy of variation testing result.
In a first aspect, the embodiment of the present invention provides a kind of method for detecting change of remote sensing image, which comprises obtain the
One remote sensing images and the second remote sensing images, first remote sensing images and second remote sensing images are to shoot in different time
The remote sensing images of the same area；According to first remote sensing images and second remote sensing images, difference image is generated；Based on institute
The frequency domain information for stating difference image removes the noise in the difference image；Two points are carried out to the difference image after removal noise
Class, generates variation testing result, and the variation testing result is used to indicate first remote sensing images and the second remote sensing figure
Difference as between.
Second aspect, the embodiment of the present invention provide a kind of Remote Sensing Imagery Change Detection device, and described device includes: reception mould
Block is configured as obtaining the first remote sensing images and the second remote sensing images, first remote sensing images and second remote sensing images
Remote sensing images for the same area shot in different time；Image generation module is configured as according to the first remote sensing figure
Picture and second remote sensing images generate difference image；Module is denoised, the frequency domain letter based on the difference image is configured as
Breath, removes the noise in the difference image；Detection module is configured as carrying out two points to the difference image after removal noise
Class, generates variation testing result, and the variation testing result is used to indicate first remote sensing images and the second remote sensing figure
Difference as between.
The third aspect, the embodiment of the present invention provide a kind of electronic equipment, and the electronic equipment includes: at least one processing
Device；And at least one processor, the bus being connected to the processor；Wherein, the processor, memory pass through described total
Line completes mutual communication；The processor is used to call program instruction in the memory, with execute said one or
Method in multiple technical solutions.
Fourth aspect, the embodiment of the present invention provide a kind of computer readable storage medium, and the storage medium includes storage
Program, wherein equipment where controlling the storage medium in described program operation executes said one or multiple technical sides
Method in case.
Method for detecting change of remote sensing image and device provided in an embodiment of the present invention are obtaining the first remote sensing images and second
After remote sensing images, firstly, generating difference image according to the first remote sensing images and the second remote sensing images；Then, it is based on difference image
Frequency domain information, remove difference image in noise；Finally, carrying out two classification to the difference image after removal noise, generates and become
Change testing result.Due to pixel in difference image information frequency domain can using the frequency domain information of difference image compared with horn of plenty
The noise in difference image is removed, enabling difference image really indicates the variation of remote sensing images, improves difference image classification
Accuracy, and then improve variation testing result accuracy.
Detailed description of the invention
By reading the following detailed description of the preferred embodiment, various other advantages and benefits are common for this field
Technical staff will become clear.The drawings are only for the purpose of illustrating a preferred embodiment, and is not considered as to the present invention
Limitation.And throughout the drawings, the same reference numbers will be used to refer to the same parts.In the accompanying drawings:
Fig. 1 is the flow diagram one of the method for detecting change of remote sensing image in the embodiment of the present invention；
Fig. 2 is the flow diagram two of the method for detecting change of remote sensing image in the embodiment of the present invention；
Fig. 3 is the remote sensing images and its variation detection figure in the Ottawa area in the embodiment of the present invention；
Fig. 4 is the remote sensing images and its variation detection figure in Vietnam, the area, Red River in the embodiment of the present invention；
Fig. 5 is PCANet, NR-ELM, FDA-RMG, NSST-APCNN in the embodiment of the present invention relative to speckle noise
Noiseproof feature comparison diagram；
Fig. 6 is PCANet, NR-ELM, FDA-RMG, NSST-APCNN in the embodiment of the present invention relative to Gaussian noise
Noiseproof feature comparison diagram；
Fig. 7 is the structural schematic diagram of the Remote Sensing Imagery Change Detection device in the embodiment of the present invention；
Fig. 8 is the structural schematic diagram of the electronic equipment in the embodiment of the present invention.
Specific embodiment
The exemplary embodiment that the present invention will be described in more detail below with reference to accompanying drawings.Although showing the present invention in attached drawing
Exemplary embodiment, it being understood, however, that may be realized in various forms the present invention without should be by embodiments set forth here
It is limited.It is to be able to thoroughly understand the present invention on the contrary, providing these embodiments, and can be by the scope of the present invention
It is fully disclosed to those skilled in the art.
The embodiment of the invention provides a kind of method for detecting change of remote sensing image and devices, in practical applications, work as needs
A certain region is detected in the earth's surface situation of change of different time, can first obtain the region in the remote sensing figure of two different times
Picture, then by the method for detecting change of remote sensing image and device, the region is obtained in the change of the remote sensing images of two different times
Change testing result, and then obtains earth's surface situation of change of the region within above-mentioned two time.
Next, method for detecting change of remote sensing image provided in an embodiment of the present invention is described in detail.
Fig. 1 is the flow diagram one of the method for detecting change of remote sensing image in the embodiment of the present invention, shown in Figure 1,
This method may include:
S101: the first remote sensing images and the second remote sensing images are obtained.
Herein, the first remote sensing images and the second remote sensing images are the remote sensing figure of the same area shot in different time
Picture.Such as: the first remote sensing images are the remote sensing images of region A shoot on January 2nd, 2018, the second remote sensing images for
The remote sensing images of the region A of shooting on March 19th, 2019.
It should be noted that the first remote sensing images and the second remote sensing images are to have already passed through geometric correction and geometry is matched
Quasi- image.
S102: according to the first remote sensing images and the second remote sensing images, difference image is generated.
In the specific implementation process, it can be the second remote sensing of sum of the grayscale values for first obtaining all pixels in the first remote sensing images
The gray value of all pixels in image, then will own in the gray value of all pixels in the first remote sensing images and the second remote sensing images
The gray value of pixel takes logarithm operation according to spatial position corresponding relationship, obtains difference image.Herein, spatial position is corresponding closes
System is the corresponding relationship of the pixel in the first remote sensing images and the pixel in the second remote sensing images, wherein in the first remote sensing images
Some pixel in the first remote sensing images position and the second remote sensing images in some pixel in the second remote sensing images
Position is identical.
Illustratively, it is assumed that all there was only 4 pixels in the first remote sensing images and the second remote sensing images, in the first remote sensing images
The gray value of all pixels is from left to right followed successively by G from top to bottom1、G2、G3、G4, the gray scale of all pixels in the second remote sensing images
Value is from left to right followed successively by G from top to bottom5、G6、G7、G8, then, the gray value of obtained difference image from left to right on to
Under be followed successively by
S103: the frequency domain information based on difference image removes the noise in difference image.
It herein, cannot be well to difference image due to being not easy to identify noise from image in airspace
It is denoised, and in a frequency domain, the information of pixel can be identified well from difference image and be made an uproar compared with horn of plenty in image
Sound, and then the noise in difference image is removed, enabling difference image really indicates the variation of remote sensing images, mentions for two classification
For more accurate data, the final accuracy for improving variation testing result.
S104: two classification are carried out to the difference image after removal noise, generate variation testing result.
Wherein, variation testing result is used to indicate the difference between the first remote sensing images and the second remote sensing images.
In the specific implementation process, two classification can be carried out to the difference image after removal noise using two sorting algorithms,
Each pixel in difference image after removal noise is divided into variation class or does not change class.Herein, two classification of use
Algorithm is the prior art, and details are not described herein.By remove noise after difference image in each pixel be divided into variation class or
Person does not change after class, can will change the corresponding pixel of class labeled as 1, will not change the corresponding pixel of class and be labeled as 0, can also
To set 255 for the gray value for changing the corresponding pixel of class, 0 is set by the gray value for not changing the corresponding pixel of class, that
, as soon as variation testing result is exactly width black white image, white region represents the first remote sensing images and the second remote sensing in image
The region of variation is generated between image, the region of black just represents in image does not have between the first remote sensing images and the second remote sensing images
There is the region for generating and changing.It is, of course, also possible to which the corresponding pixel of class pixel corresponding with class is not changed will be changed using other sides
Formula distinguishes, it is not limited here.
As shown in the above, method for detecting change of remote sensing image provided in an embodiment of the present invention is obtaining the first remote sensing
After image and the second remote sensing images, firstly, generating difference image according to the first remote sensing images and the second remote sensing images；Then, base
In the frequency domain information of difference image, the noise in difference image is removed；Finally, carrying out two points to the difference image after removal noise
Class generates variation testing result.Due to pixel in difference image information in frequency domain compared with horn of plenty, utilize the frequency domain of difference image
Information can remove the noise in difference image, and enabling difference image really indicates the variation of remote sensing images, improve difference
The accuracy of image classification, and then improve the accuracy of variation testing result.
Based on previous embodiment, as the refinement and extension of method shown in Fig. 1, the embodiment of the invention also provides a kind of distant
Feel image change detection method.Fig. 2 is the flow diagram two of the method for detecting change of remote sensing image in the embodiment of the present invention, ginseng
As shown in Figure 2, this method may include:
S201: the first remote sensing images and the second remote sensing images are obtained.
S202: denoising the first remote sensing images and the second remote sensing images, the first remote sensing images after being denoised and
The second remote sensing images after denoising.
It in the specific implementation process, can be by the neighborhood information of each pixel in the first remote sensing images to the first remote sensing figure
Each pixel is handled as in, to achieve the purpose that denoise the first remote sensing images.Specifically, being for size
The neighborhood information of the first remote sensing images midpoint pixel (i, j) of H × W can be obtained by following steps:
U=max (i-h, 1)
D=min (i+h, H)
L=max (j-w, 1)
R=min (j+w, W)
N=I (u:d, l:r)
X (i, j)=mean (N (:))
Wherein, [1, H] i ∈, j ∈ [1, W], h, w are Size of Neighborhood parameters, and N is a neighborhood for pixel (i, j), X (i, j)
It is a neighborhood information for pixel (i, j).
Then, the information of pixel (i, j) can be redefined according to a neighborhood information for pixel (i, j), specifically
, it can be using the mean value of the neighborhood information of pixel (i, j) or intermediate value as the information of point pixel (i, j), it is, of course, also possible to right
The neighborhood information of point pixel (i, j) does other processing, will treated information of the result as point pixel (i, j), do not do herein
It limits.
Herein, since the neighborhood information that each pixel in image is utilized denoises image, it can reduce noise
Interference to difference image, and then interference of the noise to variation testing result is reduced, variation testing result can be efficiently reduced
Number of false alarms.
Likewise, the method denoised to the second remote sensing images and the method phase denoised to the first remote sensing images
Together, details are not described herein.
In practical applications, the first remote sensing images and the second remote sensing images can be carried out respectively using Mean Filtering Algorithm
Denoising；The first remote sensing images and the second remote sensing images can also be denoised respectively using median filtering algorithm；It can also be both
The first remote sensing images and the second remote sensing images are denoised respectively using Mean Filtering Algorithm, and use median filtering algorithm point
It is other that first remote sensing images and the second remote sensing images are denoised, then therefrom select the good filtering algorithm difference of denoising effect
First remote sensing images and the second remote sensing images are denoised, it is, of course, also possible to using other methods respectively to the first remote sensing
Image and the second remote sensing images are denoised, it is not limited here.
S203: according to the first remote sensing images after denoising and the second remote sensing images after denoising, difference image is generated.
S204: frequency-domain transform is carried out to difference image, obtains multiple energy coefficient collection.
In the specific implementation process, wave conversion (Non-subsampled shearlet can be sheared using non-lower sampling
Transform, NSST) difference image transformed into frequency domain, since non-lower sampling shearing wave conversion is in the rotation, flat for carrying out image
Have the characteristics that Scale invariant when moving transformation, therefore, difference image can be made to transform to using non-lower sampling shearing wave conversion
Each Scale invariant is kept when frequency domain, keeps the accuracy of the frequency domain information of each scale of difference image, meanwhile, non-lower sampling shearing wave
The complexity of transformation is low, can save image processing time.
After difference image is transformed to frequency domain, so that it may obtain multiple energy coefficient collection, multiple energy coefficienies here
The scale of collection is different, and the energy coefficient collection of each scale can indicate the difference image, only carries out to difference image
The scale of expression is different, and it includes multiple energy coefficienies that each energy coefficient, which is concentrated, and the direction of multiple energy coefficienies is different,
Herein, the energy coefficient in frequency domain can analogize to the gray value of the pixel in airspace.
Illustratively, it is assumed that the size of difference image is 256 × 256, after difference image is transformed to frequency domain, so that it may
Obtain 128 × 128,64 × 64,32 × 32,16 × 16,8 × 8,4 × 4,2 × 2 these scales energy coefficient collection, be in scale
128 × 128 energy coefficient is concentrated there are 128 × 128 energy coefficienies, this 128 × 128 energy coefficienies respectively represent difference
The information of different location in image, it is believed that be that difference image is divided into 128 × 128 fritters, each fritter has an energy
Coefficient of discharge, the energy coefficient represent the information of the difference image at the tile position.Scale is respectively 64 × 64,32 × 32,16
× 16,8 × 8,4 × 4,2 × 2 energy coefficient collection is similar with the energy coefficient collection that scale is 128 × 128, and details are not described herein.
S205: whole energy coefficienies that multiple energy coefficienies are concentrated are arranged according to order of magnitude, and will be less than
The energy coefficient of preset value is set as 0.
Since in a frequency domain, the energy coefficient of noise is smaller, therefore, by the way that the lesser energy coefficient of energy coefficient is arranged
It is 0, it will be able to remove the partial noise in difference image, difference image is enable more really to represent the first remote sensing images
And the second variation between remote sensing images, then two classification are carried out to difference image, it can be improved the accuracy of variation testing result.
Illustratively, it is assumed that whole energy coefficienies that multiple energy coefficienies are concentrated are 4,9,8,7,6,1,3,10,2,5, will
These energy coefficienies are arranged as 10,9,8,7,6,5,4,3,2,1 according to the sequence of absolute value from big to small, due to energy in frequency domain
The lesser point of coefficient is significantly therefore noise or the lesser point of variation can all set the energy coefficient less than 1.5
It is set to 0, or can will come 10% last energy coefficient and be set as 0, finally obtains 10,9,8,7,6,5,4,3,2,0.
Data in above-mentioned example are only for example, and are not the restriction to the embodiment of the present invention.
Herein it should be noted that preset value can be configured according to actual detection demand, if necessary to detect
The larger of preset value setting then can be detected small variation if necessary, then can set preset value by biggish variation
That sets is smaller.In practical applications, 0 generally is set in 10% last energy coefficient by arrangement heel row from big to small,
In this way, global noise reduction can either be carried out, but can the point as much as possible that will change in difference image remain.
S206: noise estimation is carried out to each energy coefficient collection, obtains the noise variance of each energy coefficient collection.
Herein, the energy coefficient greater than preset value that each energy coefficient is concentrated remains unchanged, less than the energy of preset value
Coefficient of discharge has been arranged to 0.
S207: it is based on noise variance, the corresponding energy coefficient collection of noise variance is denoised.
It in the specific implementation process, can be based on the transformation of the afterbody in wavelet transformation, to each energy coefficient collection
Noise estimation is carried out, obtains the noise variance σ of each energy coefficient collection, noise variance σ here can indicate corresponding energy
Noise intensity in coefficient set.Due to the noise variance difference of the energy coefficient collection of different scale, according to each scale
The noise variance of energy coefficient collection denoises the energy coefficient collection of the corresponding scale of noise variance, can give for change and be mistakened as
The point making noise and deleting avoids the case where imposing uniformity without examining individual cases, that is, avoids the point really changed using same noise variance as making an uproar
Sound and the case where delete, the number of missed police of variation testing result can be effectively reduced, and then improve the standard of variation testing result
True property.
Illustratively, it is assumed that the noise variance in energy coefficient collection A is 2, and the noise variance in energy coefficient collection B is 0.2.
If it is 2 that global noise-removed threshold value, which is arranged, using the global denoising mode of clean cut, then, for energy coefficient collection A's
It is just very good to denoise effect, and it is just excessively poor for the denoising effect of energy coefficient collection B, because of the denoising mode handle of clean cut
It was that the point of signal is erroneously interpreted as noise and is removed originally in energy coefficient collection B, increased so as to cause false detection rate.If adopted
With the local denoising mode of " treating with a certain discrimination ", i.e. the noise-removed threshold value of setting energy coefficient collection A is 2, and setting energy coefficient collection B's goes
Threshold value of making an uproar is 0.2, in this way, not only the denoising effect of energy coefficient collection A is very good, and the denoising effect of energy coefficient collection B
It is very good, so that the denoising effect of entire energy coefficient collection is all very good.
S208: frequency domain inverse transformation is carried out to multiple energy coefficient collection after denoising, the difference diagram after obtaining frequency domain inverse transformation
Picture.
In the specific implementation process, the inverse transformation of non-lower sampling shearing wave can be used multiple energy coefficient collection after denoising
Airspace is transformed to, the difference image after obtaining frequency domain inverse transformation.Non-lower sampling shearing wave inverse transformation is that non-lower sampling shearing wave becomes
The inverse process changed, the inverse transformation of non-lower sampling shearing wave is in the spy for carrying out also having Scale invariant when the rotation of image, translation transformation
Therefore point can make difference image keep each Scale invariant when changing back to airspace using the inverse transformation of non-lower sampling shearing wave, protect
The accuracy of the spatial information (si) of each scale of difference image is held, meanwhile, the complexity of non-lower sampling shearing wave inverse transformation is low, can
Save image processing time.
It should be noted that inverse for the frequency domain carried out in the frequency-domain transform and step S208 that are carried out in step S204
Transformation can be using non-lower sampling shearing wave conversion in step S204 and be sheared in step S208 using non-lower sampling
Wave inverse transformation is also possible in step S204 using non-lower sampling shearing wave conversion and in step S208 except using under non-
Other frequency domain inverse transformations outside shearing wave inverse transformation are sampled, are also possible in step S204 using except non-lower sampling shearing wave becomes
It changes outer other frequency-domain transforms and uses the inverse transformation of non-lower sampling shearing wave in step S208, can also be in step S204
It is middle using except non-lower sampling shearing wave transformation in addition to other frequency-domain transforms and in step S208 using except non-lower sampling shear
Other frequency domain inverse transformations outside wave inverse transformation, it is not limited here.
S209: two classification are carried out to the difference image after frequency domain inverse transformation, generate variation testing result.
In the specific implementation process, adaptive Pulse Coupled Neural Network (Adaptive Pulse can be passed through
Coupled Neural Network, APCNN), two classification are carried out to the difference image after removal noise, generate variation detection knot
Fruit.Since adaptive Pulse Coupled Neural Network can automatically determine threshold according to the characteristics of image when carrying out two classification to image
Therefore value carries out two classification to the difference image after removal noise using adaptive Pulse Coupled Neural Network, can be improved pair
Difference image carries out the accuracy of two classification, meanwhile, adaptive Pulse Coupled Neural Network also has the function of denoising, additionally it is possible to
Difference image is denoised again, the noiseproof feature of improvement method, improves the accuracy for carrying out two classification to difference image, this
Outside, the complexity of adaptive Pulse Coupled Neural Network is low, additionally it is possible to save image processing time.
So far, according to the variation testing result of generation, it will be able to know between the first remote sensing images and the second remote sensing images
Difference, and then learn some region of earth's surface situation of change.
The detection effect of the method for detecting change of remote sensing image is illustrated with specific example below.
Example one: Fig. 3 is the remote sensing images and its variation detection figure in the Ottawa area in the embodiment of the present invention, referring to figure
Shown in 3,3a is the remote sensing images in the Ottawa area that in May, 1997 obtains, and 3b is in the Ottawa that in August, 1997 obtains
The remote sensing images in area, 3c are to be become according to obtained from the actual change situation of Ottawa earth's surface from May, 1997 in August, 1997
Change testing result, i.e., with reference to figure, 3d is according to scholars such as Gao Feng in article " Automatic Change Detection
The variation inspection based on PCANet proposed in Synthetic Aperture Radar Images Based on PCANet "
Survey method, abbreviation PCANet, and the variation testing result between obtained 3a and 3b, 3e are according to scholars such as Gao Feng in text
Chapter " Change detection from synthetic aperture radar images based on
It is proposed in neighborhood-based ratio and extreme learning machine " based on neighborhood logarithm ratio
With the change detection algorithm of extreme learning machine, abbreviation NR-ELM, and the variation testing result between obtained 3a and 3b, 3f are root
According to scholars such as Gao Feng in article " Synthetic aperture radar image change detection based
It is proposed on frequency-domain analysis and random multigraphs " based on frequency-domain analysis and with
The change detection algorithm of the more figures of machine, abbreviation FDA-RMG, and the variation testing result between obtained 3a and 3b, 3g are according to this
The method for detecting change of remote sensing image that inventive embodiments provide, abbreviation NSST-APCNN, and the variation between obtained 3a and 3b
Testing result.
In visual effect, 3d, 3e, 3f, 3g and 3c are compared, wherein white area represents the region of variation, black
The region that color Regional Representative does not change, by comparing it is found that the testing result of 3g is slightly better than 3d and 3f, better than 3e is more,
That is using method for detecting change of remote sensing image provided in an embodiment of the present invention, can more accurate detection go out it is two distant
Feel the variation between image.
In objective indicator, 3d, 3e, 3f, 3g are compared, referring to table 1, table 1 is the objective finger of 3d, 3e, 3f, 3g
Mark, wherein every objective indicator of 3d, 3e, 3f, 3g are obtained on the basis of 3c.
Table 1
Wherein, Kappa coefficient is used for consistency check, recall rate, that is, recall ratio, and F1 is to be used to measure two points in statistics
A kind of index of class model accuracy has combined the accuracy rate and recall rate of disaggregated model, it is accurate to can be regarded as model
A kind of weighted average of rate and recall rate.
By 3g and 3d, 3e, 3f being compared it is recognized that while the false-alarm pixel number of 3g increased compared to 3e, 3f, but
It is that the false dismissal pixel number of 3g and total erroneous pixel number are all reduced compared to 3d, 3e, 3f, the correct classification percentage of 3g,
Kappa coefficient, recall rate, F1 increased compared to 3d, 3e, 3f, and the runing time of 3g subtracts significantly compared to 3d, 3e, 3f
It is few, that is to say, that 3g is compared to 3d, 3e, 3f, and the accuracy for changing testing result increases, and runing time is reduced.
It can be seen that no matter in visual effect, or in objective indicator, using remote sensing provided in an embodiment of the present invention
Image change detection method can improve the accuracy of Remote Sensing Imagery Change Detection result, and reduce runing time.
Example two: Fig. 4 is the remote sensing images and its variation detection figure in Vietnam, the area, Red River in the embodiment of the present invention, referring to
Shown in Fig. 4,4a is the remote sensing images in Vietnam, the area, Red River obtained on August 24th, 1996, and 4b is to obtain on August 14th, 1999
The remote sensing images in Vietnam, the area, Red River taken, 4c are from August on August 14th, 24,1 1996 according to Vietnam's Red River earth's surface
Actual change situation obtained from change testing result, i.e., with reference to figure, 4d is according to scholars such as Gao Feng in article
“Automatic Change Detection in Synthetic Aperture Radar Images Based on
The change detecting method based on PCANet proposed in PCANet ", abbreviation PCANet, and the variation inspection between obtained 4a and 4b
It surveys as a result, 4e is according to scholars such as Gao Feng in article " Change detection from synthetic aperture
radar images based on neighborhood-based ratio and extreme learning machine”
The change detection algorithm based on neighborhood logarithm ratio and extreme learning machine of middle proposition, abbreviation NR-ELM, and obtained 4a and 4b it
Between variation testing result, 4f be according to scholars such as Gao Feng in article " Synthetic aperture radar image
In change detection based on frequency-domain analysis and random multigraphs "
The change detection algorithm based on frequency-domain analysis and random more figures proposed, abbreviation FDA-RMG, and the change between obtained 4a and 4b
Change testing result, 4g is and to obtain according to method for detecting change of remote sensing image provided in an embodiment of the present invention, abbreviation NSST-APCNN
The variation testing result between 4a and 4b arrived.
In visual effect, 4d, 4e, 4f, 4g and 4c are compared, wherein white area represents the region of variation, black
The region that color Regional Representative does not change, by comparing it is found that the testing result of 4g is slightly better than 4e and 4f, better than 4d is more,
That is using method for detecting change of remote sensing image provided in an embodiment of the present invention, can more accurate detection go out it is two distant
Feel the variation between image.
In objective indicator, 4d, 4e, 4f, 4g are compared, referring to table 2, table 2 is the objective finger of 4d, 4e, 4f, 4g
Mark, wherein every objective indicator of 4d, 4e, 4f, 4g are obtained on the basis of 4c.
Table 2
Wherein, Kappa coefficient is used for consistency check, recall rate, that is, recall ratio, and F1 is to be used to measure two points in statistics
A kind of index of class model accuracy has combined the accuracy rate and recall rate of disaggregated model, it is accurate to can be regarded as model
A kind of weighted average of rate and recall rate.
By 4g and 4d, 4e, 4f being compared it is recognized that while the false-alarm pixel number of 4g increased compared to 4d, 4e, 4f,
But the false dismissal pixel number of 4g and total erroneous pixel number are all reduced compared to 4d, 4e, 4f, the correct classification percentage of 4g,
Kappa coefficient, recall rate, F1 increased compared to 4d, 4e, 4f, and the runing time of 4g subtracts significantly compared to 4d, 4e, 4f
It is few, that is to say, that 4g is compared to 4d, 4e, 4f, and the accuracy for changing testing result increases, and runing time is reduced.
It can be seen that no matter in visual effect, or in objective indicator, using remote sensing provided in an embodiment of the present invention
Image change detection method can improve the accuracy of Remote Sensing Imagery Change Detection result, and reduce runing time.
Next, further illustrating the noiseproof feature of the method for detecting change of remote sensing image, herein, noiseproof feature refers to
Noise front and back is added into remote sensing images leads to the changed degree of change detection result, and variation degree is smaller, i.e., closer
1, illustrate that noiseproof feature is better.
Fig. 5 is PCANet, NR-ELM, FDA-RMG, NSST-APCNN in the embodiment of the present invention relative to speckle noise
Noiseproof feature comparison diagram, shown in Figure 5,5a is that speckle noise is added in the remote sensing images in Ottawa area to obtain noise immunity
Energy comparison diagram, 5b are that speckle noise is added in the remote sensing images in Vietnam, area, Red River to obtain noiseproof feature comparison diagram, addition
The range of speckle noise is that PSNR ∈ [26,51] dB, PSNR is Y-PSNR, i.e. Peak Signal to Noise
Ratio, τ are noiseproof feature, no matter in 5a or in 5b, it can be seen that the noiseproof feature ratio of NSST-APCNN
PCANet, FDA-RMG are slightly higher, than NR-ELM high.
Fig. 6 is PCANet, NR-ELM, FDA-RMG, NSST-APCNN in the embodiment of the present invention relative to Gaussian noise
Noiseproof feature comparison diagram, shown in Figure 6,6a is that Gaussian noise is added in the remote sensing images in Ottawa area to obtain noise immunity
Energy comparison diagram, 6b are that Gaussian noise is added in the remote sensing images in Vietnam, area, Red River to obtain noiseproof feature comparison diagram, addition
The range of Gaussian noise is that PSNR ∈ [35,50] dB, PSNR is Y-PSNR, i.e. Peak Signal to Noise
Ratio, τ are noiseproof feature, no matter in 6a or in 6b, it can be seen that the noiseproof feature ratio of NSST-APCNN
PCANet, FDA-RMG are slightly higher, than NR-ELM high.
It can be seen from the above, either speckle noise or Gaussian noise, in the range of PSNR ∈ [35,50] dB,
The noiseproof feature of NSST-APCNN is all more slightly higher than PCANet, FDA-RMG, than NR-ELM high.
Based on the same inventive concept, as an implementation of the above method, the embodiment of the invention also provides a kind of remote sensing figures
As change detecting device.Fig. 7 is the structural schematic diagram of the Remote Sensing Imagery Change Detection device in the embodiment of the present invention, referring to Fig. 7
Shown, which may include: receiving module 701, be configured as obtaining the first remote sensing images and the second remote sensing images, and first
Remote sensing images and the second remote sensing images are the remote sensing images of the same area shot in different time；Image generation module 702, quilt
It is configured to generate difference image according to the first remote sensing images and the second remote sensing images；Module 703 is denoised, is configured as based on poor
The frequency domain information of partial image removes the noise in difference image；Detection module 704 is configured as to the difference after removal noise
Image carries out two classification, generates variation testing result, and variation testing result is used to indicate the first remote sensing images and the second remote sensing figure
Difference as between.
Based on previous embodiment, module is denoised, is configured as carrying out frequency-domain transform to difference image, obtains multiple energy systems
The scale of manifold, multiple energy coefficient collection is different, and it includes multiple energy coefficienies, multiple energy systems that each energy coefficient, which is concentrated,
Several directions is different；Whole energy coefficienies that multiple energy coefficienies are concentrated are arranged according to order of magnitude, and will
Energy coefficient less than preset value is set as 0；Frequency is carried out with the energy coefficient for being arranged to 0 to the energy coefficient for being greater than preset value
Domain inverse transformation, the difference image after obtaining frequency domain inverse transformation；Detection module is configured as to the difference image after frequency domain inverse transformation
Two classification are carried out, variation testing result is generated.
Based on previous embodiment, module is denoised, is additionally configured to carry out noise estimation to each energy coefficient collection, is obtained every
The noise variance of a energy coefficient collection；Based on noise variance, the corresponding energy coefficient collection of noise variance is denoised；To denoising
Multiple energy coefficient collection afterwards carry out frequency domain inverse transformation, the difference image after obtaining frequency domain inverse transformation.
Based on previous embodiment, module is denoised, is configured as shearing wave conversion based on non-lower sampling, difference image is carried out
Frequency-domain transform obtains multiple energy coefficient collection.
Based on previous embodiment, module is denoised, is configured as based on the inverse transformation of non-lower sampling shearing wave, to greater than preset value
Energy coefficient and be arranged to 0 energy coefficient carry out frequency domain inverse transformation, the difference image after obtaining frequency domain inverse transformation.
Based on previous embodiment, detection module is configured as based on adaptive Pulse Coupled Neural Network, to removal noise
Difference image afterwards carries out two classification, generates variation testing result.
Based on previous embodiment, which can also include: preprocessing module；Preprocessing module is configured as to first
Remote sensing images and the second remote sensing images are denoised, the first remote sensing images after being denoised and the second remote sensing figure after denoising
Picture；Image generation module is configured as according to the first remote sensing images after denoising and the second remote sensing images after denoising, and it is poor to generate
Partial image.
It need to be noted that: the description of apparatus above embodiment, be with the description of above method embodiment it is similar,
With the similar beneficial effect of same embodiment of the method.For undisclosed technical detail in apparatus of the present invention embodiment, please refer to
The description of embodiment of the present invention method and understand.
Based on the same inventive concept, the embodiment of the invention also provides a kind of electronic equipment.Fig. 8 is in the embodiment of the present invention
Electronic equipment structural schematic diagram, shown in Figure 8, which may include: at least one processor 801；With
And at least one processor 802, the bus 803 being connect with processor 801；Wherein, processor 801, memory 802 pass through bus
803 complete mutual communication；Processor 801 is used to call program instruction in memory 802, to execute said one or more
Method in a embodiment.
It need to be noted that: the description of the above electronic equipment embodiment, the description with above method embodiment are classes
As, there is with embodiment of the method similar beneficial effect.For not draped over one's shoulders in the embodiment of the electronic equipment of the embodiment of the present invention
The technical detail of dew please refers to the description of embodiment of the present invention method and understands.
Based on the same inventive concept, the embodiment of the invention also provides a kind of computer readable storage medium, above-mentioned calculating
Machine readable storage medium storing program for executing includes the program of storage, wherein in program operation, equipment where control storage medium executes above-mentioned one
Method in a or multiple embodiments.
It need to be noted that: the description of the above computer readable storage medium embodiment, with above method embodiment
Description be it is similar, have with embodiment of the method similar beneficial effect.Computer-readable for the embodiment of the present invention is deposited
Undisclosed technical detail in the embodiment of storage media, please refers to the description of embodiment of the present invention method and understands.
It should be understood by those skilled in the art that, embodiments herein can provide as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the application
Apply the form of example.Moreover, it wherein includes the computer of computer usable program code that the application, which can be used in one or more,
The computer program implemented in usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) produces
The form of product.
The application is referring to method, the process of equipment (system) and computer program product according to the embodiment of the present application
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
In a typical configuration, calculating equipment includes one or more processors (CPU), input/output interface, net
Network interface and memory.
Memory may include the non-volatile memory in computer-readable medium, random access memory (RAM) and/
Or the forms such as Nonvolatile memory, such as read-only memory (ROM) or flash memory (flash RAM).Memory is computer-readable Jie
The example of matter.
Computer-readable medium includes permanent and non-permanent, removable and non-removable media can be by any method
Or technology come realize information store.Information can be computer readable instructions, data structure, the module of program or other data.
The example of the storage medium of computer includes, but are not limited to phase change memory (PRAM), static random access memory (SRAM), moves
State random access memory (DRAM), other kinds of random access memory (RAM), read-only memory (ROM), electric erasable
Programmable read only memory (EEPROM), flash memory or other memory techniques, read-only disc read only memory (CD-ROM) (CD-ROM),
Digital versatile disc (DVD) or other optical storage, magnetic cassettes, tape magnetic disk storage or other magnetic storage devices
Or any other non-transmission medium, can be used for storage can be accessed by a computing device information.As defined in this article, it calculates
Machine readable medium does not include temporary computer readable media (transitory media), such as the data-signal and carrier wave of modulation.
It should also be noted that, the terms "include", "comprise" or its any other variant are intended to nonexcludability
It include so that the process, method, commodity or the equipment that include a series of elements not only include those elements, but also to wrap
Include other elements that are not explicitly listed, or further include for this process, method, commodity or equipment intrinsic want
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including element
There is also other identical elements in process, method, commodity or equipment.
It will be understood by those skilled in the art that embodiments herein can provide as method, system or computer program product.
Therefore, complete hardware embodiment, complete software embodiment or embodiment combining software and hardware aspects can be used in the application
Form.It is deposited moreover, the application can be used to can be used in the computer that one or more wherein includes computer usable program code
The shape for the computer program product implemented on storage media (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.)
Formula.
The above is only embodiments herein, are not intended to limit this application.To those skilled in the art,
Various changes and changes are possible in this application.It is all within the spirit and principles of the present application made by any modification, equivalent replacement,
Improve etc., it should be included within the scope of the claims of this application.
Claims (10)
1. a kind of method for detecting change of remote sensing image, which is characterized in that the described method includes:
The first remote sensing images and the second remote sensing images are obtained, first remote sensing images and second remote sensing images are in difference
The remote sensing images of the same area of time shooting；
According to first remote sensing images and second remote sensing images, difference image is generated；
Based on the frequency domain information of the difference image, the noise in the difference image is removed；
Two classification are carried out to the difference image after removal noise, generate variation testing result, the variation testing result is for referring to
Show the difference between first remote sensing images and second remote sensing images.
2. the method according to claim 1, wherein the frequency domain information based on the difference image, removal
Noise in the difference image, comprising:
Frequency-domain transform is carried out to the difference image, obtains multiple energy coefficient collection, the scale of the multiple energy coefficient collection is each
Not identical, it includes multiple energy coefficienies that each energy coefficient, which is concentrated, and the direction of the multiple energy coefficient is different；
Whole energy coefficienies that the multiple energy coefficient is concentrated are arranged according to order of magnitude, and will be less than preset value
Energy coefficient be set as 0；
Frequency domain inverse transformation is carried out with the energy coefficient for being arranged to 0 to the energy coefficient for being greater than the preset value, it is inverse to obtain frequency domain
Transformed difference image；
Difference image after described pair of removal noise carries out two classification, generates variation testing result, comprising:
Two classification are carried out to the difference image after frequency domain inverse transformation, generate variation testing result.
3. according to the method described in claim 2, it is characterized in that, described pair is greater than the energy coefficient of the preset value and is set
The energy coefficient for being set to 0 carries out frequency domain inverse transformation, before the difference image after obtaining frequency domain inverse transformation, the method also includes:
Noise estimation is carried out to each energy coefficient collection, obtains the noise variance of each energy coefficient collection；
Based on the noise variance, the corresponding energy coefficient collection of the noise variance is denoised；
The described pair of energy coefficient for being greater than the preset value carries out frequency domain inverse transformation with the energy coefficient for being arranged to 0, obtains frequency
Difference image after the inverse transformation of domain, comprising:
Frequency domain inverse transformation is carried out to multiple energy coefficient collection after denoising, the difference image after obtaining frequency domain inverse transformation.
4. according to the method in claim 2 or 3, which is characterized in that it is described that frequency-domain transform is carried out to the difference image, it obtains
To multiple energy coefficient collection, comprising:
Wave conversion is sheared based on non-lower sampling, frequency-domain transform is carried out to the difference image, obtains multiple energy coefficient collection.
5. according to the method described in claim 4, it is characterized in that, described pair is greater than the energy coefficient of the preset value and is set
The energy coefficient for being set to 0 carries out frequency domain inverse transformation, the difference image after obtaining frequency domain inverse transformation, comprising:
Based on non-lower sampling shearing wave inverse transformation, to being greater than the energy coefficient of the preset value and be arranged to 0 energy coefficient
Carry out frequency domain inverse transformation, the difference image after obtaining frequency domain inverse transformation.
6. according to the method in claim 2 or 3, which is characterized in that described pair be greater than the preset value energy coefficient and
The energy coefficient for being arranged to 0 carries out frequency domain inverse transformation, the difference image after obtaining frequency domain inverse transformation, comprising:
Based on non-lower sampling shearing wave inverse transformation, to being greater than the energy coefficient of the preset value and be arranged to 0 energy coefficient
Carry out frequency domain inverse transformation, the difference image after obtaining frequency domain inverse transformation.
7. according to the method in any one of claims 1 to 3, which is characterized in that the difference diagram after described pair of removal noise
As carrying out two classification, variation testing result is generated, comprising:
Based on adaptive Pulse Coupled Neural Network, two classification are carried out to the difference image after removal noise, generate variation detection
As a result.
8. according to the method in any one of claims 1 to 3, which is characterized in that described according to first remote sensing images
With second remote sensing images, generate difference image before, the method also includes:
First remote sensing images and second remote sensing images are denoised, the first remote sensing images after being denoised and are gone
The second remote sensing images after making an uproar；
It is described according to first remote sensing images and second remote sensing images, generate difference image, comprising:
According to the first remote sensing images after denoising and the second remote sensing images after denoising, difference image is generated.
9. a kind of Remote Sensing Imagery Change Detection device, which is characterized in that described device includes:
Receiving module, is configured as obtaining the first remote sensing images and the second remote sensing images, first remote sensing images and described the
Two remote sensing images are the remote sensing images of the same area shot in different time；
Image generation module is configured as generating difference image according to first remote sensing images and second remote sensing images；
Module is denoised, the frequency domain information based on the difference image is configured as, removes the noise in the difference image；
Detection module is configured as carrying out two classification to the difference image after removal noise, generates variation testing result, the change
Change testing result and is used to indicate the difference between first remote sensing images and second remote sensing images.
10. a kind of electronic equipment, which is characterized in that the electronic equipment includes:
At least one processor；
And at least one processor, the bus being connected to the processor；
Wherein, the processor, memory complete mutual communication by the bus；The processor is described for calling
Program instruction in memory, to execute such as method described in any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910232115.6A CN109978855A (en) | 2019-03-26 | 2019-03-26 | A kind of method for detecting change of remote sensing image and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910232115.6A CN109978855A (en) | 2019-03-26 | 2019-03-26 | A kind of method for detecting change of remote sensing image and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109978855A true CN109978855A (en) | 2019-07-05 |
Family
ID=67080659
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910232115.6A Pending CN109978855A (en) | 2019-03-26 | 2019-03-26 | A kind of method for detecting change of remote sensing image and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109978855A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112329677A (en) * | 2020-11-12 | 2021-02-05 | 北京环境特性研究所 | Remote sensing image river target detection method and device based on feature fusion |
| CN113515971A (en) * | 2020-04-09 | 2021-10-19 | 阿里巴巴集团控股有限公司 | Data processing method and system, network system and training method and device thereof |
| CN114049568A (en) * | 2021-11-29 | 2022-02-15 | 中国平安财产保险股份有限公司 | Object shape change detection method, device, equipment and medium based on image comparison |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107103595A (en) * | 2017-05-27 | 2017-08-29 | 新疆大学 | Method, device, storage medium and the equipment of detection image change |
| CN107301641A (en) * | 2017-06-21 | 2017-10-27 | 新疆大学 | A kind of detection method and device of Remote Sensing Imagery Change |
| CN107590788A (en) * | 2017-09-12 | 2018-01-16 | 新疆大学 | A kind of remote sensing image processing method |
| GB201801838D0 (en) * | 2018-02-05 | 2018-03-21 | Sony Corp | Data encoding and decoding |
-
2019
- 2019-03-26 CN CN201910232115.6A patent/CN109978855A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107103595A (en) * | 2017-05-27 | 2017-08-29 | 新疆大学 | Method, device, storage medium and the equipment of detection image change |
| CN107301641A (en) * | 2017-06-21 | 2017-10-27 | 新疆大学 | A kind of detection method and device of Remote Sensing Imagery Change |
| CN107590788A (en) * | 2017-09-12 | 2018-01-16 | 新疆大学 | A kind of remote sensing image processing method |
| GB201801838D0 (en) * | 2018-02-05 | 2018-03-21 | Sony Corp | Data encoding and decoding |
Non-Patent Citations (4)
| Title |
|---|
| PENGYUN CHEN 等: "Unsupervised Change Detection of SAR Images Based on an Improved NSST Algorithm", 《JOURNAL OF THE INDIAN SOCIETY OF REMOTE SENSING》 * |
| ZHOU WENYAN 等: "SAR image change detection based on equal weight image fusion and adaptive threshold in the NSST domain", 《EUROPEAN JOURNAL OF REMOTE SENSING》 * |
| 周文艳: "基于差异图融合的SAR遥感图像变化检测研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 王鸿闯: "基于NSCT域的自适应阈值图像去噪算法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113515971A (en) * | 2020-04-09 | 2021-10-19 | 阿里巴巴集团控股有限公司 | Data processing method and system, network system and training method and device thereof |
| CN112329677A (en) * | 2020-11-12 | 2021-02-05 | 北京环境特性研究所 | Remote sensing image river target detection method and device based on feature fusion |
| CN112329677B (en) * | 2020-11-12 | 2024-02-02 | 北京环境特性研究所 | Remote sensing image river channel target detection method and device based on feature fusion |
| CN114049568A (en) * | 2021-11-29 | 2022-02-15 | 中国平安财产保险股份有限公司 | Object shape change detection method, device, equipment and medium based on image comparison |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110472627B (en) | End-to-end SAR image recognition method, device and storage medium | |
| CN110598558B (en) | Crowd density estimation method, device, electronic equipment and medium | |
| Chen et al. | Ship detection from coastal surveillance videos via an ensemble Canny-Gaussian-morphology framework | |
| CN110781756A (en) | Urban road extraction method and device based on remote sensing image | |
| CN109284673B (en) | Object tracking method and device, electronic equipment and storage medium | |
| CN109978855A (en) | A kind of method for detecting change of remote sensing image and device | |
| CN110415208A (en) | A kind of adaptive targets detection method and its device, equipment, storage medium | |
| CN110400294B (en) | Infrared target detection system and detection method | |
| CN112364865B (en) | Method for detecting small moving target in complex scene | |
| CN110570442A (en) | Contour detection method under complex background, terminal device and storage medium | |
| CN110706208A (en) | Infrared dim target detection method based on tensor mean square minimum error | |
| CN108596032B (en) | Detection method, device, equipment and medium for fighting behavior in video | |
| CN110222704A (en) | A kind of Weakly supervised object detection method and device | |
| CN106778822B (en) | Image straight line detection method based on funnel transformation | |
| CN107729903A (en) | SAR image object detection method based on area probability statistics and significance analysis | |
| CN116229419B (en) | Pedestrian detection method and device | |
| CN107301655A (en) | A kind of video movement target method for detecting based on background modeling | |
| Zhou et al. | AMSP-UOD: When vortex convolution and stochastic perturbation meet underwater object detection | |
| CN115761606A (en) | Box electric energy meter identification method and device based on image processing | |
| CN103903258A (en) | Method for detecting changes of remote sensing image based on order statistic spectral clustering | |
| CN111079797B (en) | Image classification method and device and storage medium | |
| Khryashchev et al. | Evaluation of face image quality metrics in person identification problem | |
| Nian et al. | Image Enhancement Algorithm based on Local Contrast for Convolutional Neural Network-Based Infrared Target Recognition:* Note: Subtitles are not captured in Xplore and should not be used | |
| Kim et al. | Analyzing the performance of canny edge detection on interpolated Images | |
| Li et al. | SAR object detection with a saliency method based on PCA and global contrast |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190705 |