CN112053359B - Remote sensing image change detection method and device, electronic equipment and storage medium - Google Patents
Remote sensing image change detection method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN112053359B CN112053359B CN202011056145.5A CN202011056145A CN112053359B CN 112053359 B CN112053359 B CN 112053359B CN 202011056145 A CN202011056145 A CN 202011056145A CN 112053359 B CN112053359 B CN 112053359B
- Authority
- CN
- China
- Prior art keywords
- time sequence
- data
- remote sensing
- pixel
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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
- G06T2207/20092—Interactive image processing based on input by user
- G06T2207/20104—Interactive definition of region of interest [ROI]
-
- 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
- G06T2207/30181—Earth observation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
Abstract
The disclosure provides a remote sensing image change detection method, a remote sensing image change detection device, electronic equipment and a storage medium, wherein the remote sensing image change detection method comprises the following steps: performing pretreatment such as radiation correction, orthographic correction, resampling, registration, cloud and shadow removal, band or index extraction and the like on the periodic time sequence remote sensing image to obtain periodic time sequence data of pixel by pixel and band by band or index; using an empirically set LSTM network structure to learn time sequence data before a monitoring period pixel by pixel to obtain a time sequence evolution LSTM model of a pixel level; predicting each wave band or index data of each pixel in the monitoring period by using each model, and obtaining a non-periodic variation marker graph of each wave band or index in the monitoring period by calculating the difference between the model predicted value and the real monitoring value; and classifying all the aperiodic change marker graphs by using a classifier obtained through training the interesting change samples to obtain a classification detection result of the interesting change.
Description
Technical Field
The present disclosure relates to the field of image processing technologies, and in particular, to a remote sensing image change detection method and apparatus, an electronic device, and a storage medium.
Background
Under the support of the domestic and foreign high-resolution satellite earth observation system, the remote sensing images covering the global and key areas are continuously accumulated and rapidly updated. The periodic time sequence remote sensing image with high time resolution can greatly improve the near real-time monitoring capability of the earth surface change. The periodic time sequence remote sensing image contains rich earth surface regularity change information, and electricity implies earth surface irregularity change information caused by human activities or natural disasters, such as land development, natural disasters and the like. Meanwhile, the periodic time sequence remote sensing image data is mixed with complex disturbance caused by factors such as weather and physical conditions, and the like, so that the change detection is challenged. Therefore, by analyzing the periodic time sequence remote sensing image, the earth surface regular and irregular change information is mined, and interesting changes are found, which is important to improving the change detection capability and pushing the near real-time remote sensing monitoring capability.
Remote sensing image change detection is one of the most classical and most active technical branches in the image processing field, and a detection method is rapid in recent years except a classical dual-phase contrast method and a multi-time phase analysis method. Methods based on time series analysis are evolving from retrospective detection and analysis of historical changes to near real-time monitoring and early warning of current or potential changes. Meanwhile, in addition to the annual changes of the conventional land cover/utilization type, there is increasing concern about irregular, interesting changes occurring on the ground surface, such as insect diseases, drought, geological disasters, and the like.
According to different mechanisms of time sequence analysis and change detection adopted, the periodic time sequence satellite image change detection method can be classified into three main categories: window analysis, timing segmentation and timing prediction. The time sequence prediction method predicts future fluctuation trend based on historical fluctuation characteristics, can detect changes when new observation data appear, has strong continuity and timeliness, can overcome the problem caused by seasonal time phase difference, and is suitable for near real-time monitoring of surface changes. However, the classical time sequence prediction method mostly adopts an explicit and determined regression model and a prediction model, the model form and parameters thereof are too dependent on ground objects and experiences, and the contribution degree of data in different periods is not distinguished, so that the method model is difficult to adaptively expand to other geographic areas or remote sensing index data, and the recent dynamic change trend of the data is difficult to reflect.
Disclosure of Invention
The main purpose of the application is to provide a remote sensing image change detection method, a remote sensing image change detection device, an electronic device and a storage medium, wherein the convenience and the effectiveness of periodic time sequence image change detection can be improved by learning a short-term to long-term dependency relationship in a sequence to generate an implicit self-adaptive sequence prediction model.
To achieve the above object, a first aspect of an embodiment of the present application provides a remote sensing image change detection method, including:
preprocessing the periodic time sequence remote sensing image, wherein the periodic time sequence remote sensing image of the t-th time phase is expressed as I t ，I t Is denoted as n, the period is denoted as s, { I t ：t＝1，2，...，m，m+1，...，n}；
Extracting an image spectrum band from the preprocessed periodic time sequence remote sensing image or calculating a remote sensing index to obtain a periodic time sequence spectrum band image or a remote sensing index image, wherein the periodic time sequence spectrum band image or the remote sensing index image is expressed as
Is expressed as i,/or +.>
Extracting periodic time sequence data pixel by pixel and phase by phase from the periodic time sequence spectrum band image or the remote sensing index image, wherein the periodic time sequence data is expressed as
(x, y) representing image pixel coordinates of the periodic time series data, < >>
Dividing the periodic time sequence data into two sections according to time phase t=m to respectively obtain time sequence training data and time sequence monitoring data, wherein the time sequence training data is that
The time sequence monitoring data is
Training a preset long-short-period memory network by utilizing the time sequence training data pixel by pixel, band by band or exponentially to obtain a periodic time sequence band or index data evolution model of a pixel level, wherein the periodic time sequence band or index data evolution model of the pixel level is expressed as
Predicting data except t=m in the time sequence monitoring data by using the periodic time sequence wave band or index data evolution model of the pixel level to obtain time sequence prediction data of each pixel in each wave band or remote sensing index, wherein the time sequence prediction data is expressed as
Calculating a difference between the time sequence prediction data and the time sequence monitoring data, and marking corresponding pixels as non-periodic variation data at corresponding time phases when the difference is within a preset range, wherein the non-periodic variation data is expressed as
Classifying the aperiodic variable marking data with the category number of K+1 by using a supervision classification method to obtain K-class interesting variable marks and one class of other variable marks, wherein the K-class interesting variable marks are expressed as
Marking the periodic time sequence remote sensing images { I }, respectively t ：t＝m+1，...，n, the label is any one of no change, corresponding K-class interesting change and other changes.
Optionally, the preset long-period memory network is expressed as lstm= { I, H, O }, where I represents the number of nodes of the input layer of the preset long-period memory network and I is greater than or equal to S/2, H represents the number of hidden layers of the preset long-period memory network and H is greater than or equal to 1,O and H represents the number of nodes of the output layer of the preset long-period memory network and O is greater than or equal to 1 and less than or equal to n-m.
Optionally, let the difference between the time sequence prediction data and the time sequence monitoring data be
Then
Optionally, let the difference between the time sequence prediction data and the time sequence monitoring data be
The difference is +.>
Wherein (1)>
For the corresponding pixel when the preset long-period memory network is in time phase { t epsilon t } s : s=1, 2,..s } and L is the training error multiple allowed by the prediction error.
Optionally, the preprocessing includes radiation correction, orthographic correction, registration and resampling, cloud and shadow removal.
Optionally, the samples of the supervised classification are the preprocessed satellite image pixel data containing the K-class interest variations.
Optionally, the feature of the supervision classification is the periodic time sequence spectrum band image or the remote sensing index image
In the spectrum band or remote sensing index i.
A second aspect of the embodiments of the present application provides a remote sensing image change detection device, including:
the preprocessing module is used for preprocessing the periodic time sequence remote sensing image, wherein the periodic time sequence remote sensing image of the t-th time phase is expressed as I t ，I t The time sequence length of (1) is represented as n, the period is represented as S, S is not less than 1 and not more than m, { I t ：，＝1，2，...，m，m+1，...，n}；
A first extraction module for extracting periodic time sequence data pixel by pixel and phase by phase from the periodic time sequence spectrum band image or the remote sensing index image, wherein the periodic time sequence data is expressed as
(x, y) representing image pixel coordinates of the periodic time series data, < >>
A second extraction module, configured to divide the periodic time sequence data into two segments according to a time phase t=m, and obtain time sequence training data and time sequence monitoring data respectively, where the time sequence training data is
The time sequence monitoring data is +.>
A segmentation module for dividing the periodic time sequence data into two segments according to time phase t=m to respectively obtain time sequence training data and time sequence monitoring data, wherein the time sequence training data is that
The time sequence monitoring data is +.>
The training module is used for training a preset long-short-period memory network by utilizing the time sequence training data pixel by pixel, band by band or exponentially to obtain a periodic time sequence band or exponential data evolution model of a pixel level, wherein the periodic time sequence band or exponential data evolution model of the pixel level is expressed as
The prediction module is used for predicting data except t=m in the time sequence monitoring data by using the periodic time sequence wave band or index data evolution model of the pixel level to obtain time sequence prediction data of each pixel in each wave band or remote sensing index, wherein the time sequence prediction data is expressed as
A calculation module for calculating the difference between the time sequence prediction data and the time sequence monitoring data, and marking the corresponding pixel as non-periodic variation data at the corresponding time phase when the difference is within a preset range, wherein the non-periodic variation data is expressed as
The classification module is used for classifying the aperiodic variable marking data with the category number of K+1 by using a supervision classification method to obtain K types of interesting variable marks and one type of other variable marks, wherein the K types of interesting variable marks are expressed as
Marking modules for respectively marking the periodic time sequence remote sensing images { I } t : t=m+1,..n } each pixel, labeled as any one of no change, corresponding K-class change of interest, other change.
A third aspect of the embodiments of the present application provides an electronic device, including:
the remote sensing image change detection method is characterized in that the remote sensing image change detection method provided in the first aspect of the embodiment of the application is realized when the processor executes the program.
A fourth aspect of the embodiments of the present application provides a computer readable storage medium having stored thereon a computer program, which when executed by a processor, implements the remote sensing image change detection method provided in the first aspect of the embodiments of the present application.
According to the remote sensing image change detection method, the remote sensing image change detection device, the electronic equipment and the storage medium, the time sequence dynamic change mode is learned through a deep learning artificial neural network long-short-term memory network, and the remote sensing image change detection method, the electronic equipment and the storage medium are an implicit self-adaptive model driven by data learning, so that the problems that the traditional method adopts an explicit ground-driven empirical model, and the regional adaptability is poor, the model form and the parameters need to be optimized manually and the like are solved. Meanwhile, the method adopts a long-period memory network to model different periods or periodic relations in periodic time sequence data, dynamically updates and learns the latest periodic change rule and gradually forgets longer periodic change information, and avoids the problem of interference influence of long-period time sequence data on the model caused by the adoption of a deterministic model in the traditional method; through the advantages of the two aspects, the periodic time sequence remote sensing image change detection result is more robust and reliable.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flow chart of a remote sensing image change detection method according to an embodiment of the present application;
fig. 2 is a schematic structural diagram of a remote sensing image change detection device according to an embodiment of the present application;
FIG. 3 shows a schematic diagram of the hardware architecture of an electronic device;
FIG. 4 is a schematic view of different phase images and aperiodic variable signatures of a certain area of investigation SAVI (soil adjusted vegetation index) provided by the present disclosure;
FIG. 5 is a schematic view of aperiodic change marks in 4 remote sensing index images of a research area provided by the present disclosure;
FIG. 6 is a schematic representation of the detection of non-periodic changes of interest to a region of interest provided by the present disclosure.
Detailed Description
In order to make the application objects, features and advantages of the present application more obvious and understandable, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments herein without making any inventive effort, are intended to be within the scope of the present application.
Referring to fig. 1, fig. 1 is a flow chart of a remote sensing image change detection method according to an embodiment of the present application, and the method mainly includes the following steps:
s101, preprocessing a periodic time sequence remote sensing image, wherein the periodic time sequence remote sensing image of the t-th time phase is expressed as I t ，I t Is denoted as n, the period is denoted as s, { I t ：t＝1，2，...，m，m+1，...，n}。
S102, from passing through the pre-treatmentExtracting an image spectrum band from the processed periodic time sequence remote sensing image or calculating a remote sensing index to obtain a periodic time sequence spectrum band image or a remote sensing index image, wherein the periodic time sequence spectrum band image or the remote sensing index image is expressed as
Is expressed as i,/or +.>
S103, extracting periodic time sequence data pixel by pixel and phase by phase from the periodic time sequence spectrum band image or the remote sensing index image, wherein the periodic time sequence data is expressed as
(x, y) representing the image pixel coordinates of the periodic time series data,/and>
s104, dividing the periodic time sequence data into two sections according to time phase t=m to respectively obtain time sequence training data and time sequence monitoring data, wherein the time sequence training data is that
The time sequence monitoring data is
S105, training a preset Long Short-Term Memory (LSTM) network by using the time sequence training data pixel by pixel, band by band or exponentially to obtain a pixel-level periodic time sequence band or index data evolution model, wherein the pixel-level periodic time sequence band or index data evolution model is expressed as
S106, predicting data except t=m in the time sequence monitoring data by using the periodic time sequence wave band or index data evolution model of the pixel level to obtain time sequence prediction data of each pixel in each wave band or remote sensing index, wherein the time sequence prediction data is expressed as
S107, calculating the difference between the time sequence prediction data and the time sequence monitoring data, and marking the corresponding pixel as aperiodic variation marking data at the corresponding time phase when the difference is within a preset range, wherein the aperiodic variation data is expressed as
More, data outside the preset range is marked as unchanged data.
S108, classifying the aperiodic variable marking data with the category number of K+1 by using a supervision classification method to obtain K-type interesting variable marks and one type of other variable marks, wherein the K-type interesting variable marks are expressed as
S109, marking the periodic time sequence remote sensing image { I }, respectively t : t=m+1,..n } each pixel, labeled as any one of no change, corresponding K-class change of interest, other change.
In one embodiment of the present disclosure, the preset long-term memory network is expressed as lstm= { I, H, O }, where I represents the number of nodes of the input layer of the preset long-term memory network and I is greater than or equal to S/2, H represents the number of hidden layers of the preset long-term memory network and H is greater than or equal to 1,O and 1 is greater than or equal to O is less than or equal to n-m.
In one embodiment of the present disclosure, the difference between the time sequence prediction data and the time sequence monitoring data is made as follows
Then->
In one embodiment of the present disclosure, the difference between the time sequence prediction data and the time sequence monitoring data is made as follows
The difference is +.>
Wherein (1)>
For corresponding pixel in time phase { t epsilon t of the preset long-term memory network s : s=1, 2,..s } and L is the training error multiple allowed by the prediction error.
In one embodiment of the present disclosure, the preprocessing includes radiation correction, orthographic correction, registration and resampling, cloud and shadow removal.
In one embodiment of the present disclosure, the sample of the supervised classification is the preprocessed satellite image pixel data containing the K-class interest variations.
In one embodiment of the present disclosure, the supervised classification is characterized by the periodic time-series spectral band images or remote sensing index images
In the spectrum band or remote sensing index i.
Referring to fig. 2, fig. 2 is a schematic structural diagram of a remote sensing image change detection device according to an embodiment of the present application, where the device may be built into an electronic device, and the device mainly includes:
a preprocessing module 201 for preprocessing the periodic time-series remote sensing image, wherein the periodic time-series remote sensing image of the t-th time phase is denoted as I t ，I t Is denoted as n, the period is denoted as S, { I t ：t＝1，2，...，m，m+1，...，n}；
A first extraction module 202 for extracting an image spectrum band or calculating a remote sensing index from the preprocessed periodic time-series remote sensing image to obtain a periodic time-series spectrum band image or a remote sensing index image, wherein the periodic time-series spectrum band image or the remote sensing index image is expressed as
Is denoted as i,
a second extraction module 203 for extracting periodic time series data pixel by pixel and phase by phase from the periodic time series spectrum band image or the remote sensing index image, wherein the periodic time series data is expressed as
(x, y) representing the image pixel coordinates of the periodic time series data,/and>
a segmentation module 204 for dividing the periodic time sequence data into two segments with time phase t=m to obtain time sequence training data and time sequence monitoring data respectively, wherein the time sequence training data is that
The time sequence monitoring data is
The training module 205 is configured to train the preset long-short-period memory network by using the time sequence training data pixel by pixel, band by band or exponentially to obtain a periodic time sequence band or exponential data evolution model at a pixel level, where the periodic time sequence band or exponential data evolution model at the pixel level is expressed as
A prediction module 206, configured to predict data except t=m in the time-series monitoring data by using the periodic time-series band or index data evolution model of the pixel level to obtain time-series prediction data of each pixel in each band or remote sensing index, where the time-series prediction data is expressed as
A calculation module 207 for calculating the difference between the time-series prediction data and the time-series monitoring data, and marking the corresponding pixel as aperiodic variation marking data at the corresponding time phase when the difference is within the preset range, wherein the aperiodic variation marking data is expressed as
A classification module 208, configured to classify the aperiodic variable tag data with a class number of k+1 by using a supervised classification method, to obtain a K-class interesting variable tag and one class of other variable tags, where the K-class interesting variable tag is expressed as
Marking module 209 for marking the periodic time-series remote sensing images { I }, respectively t : t=m+1,..n } each pixel, labeled as any one of no change, corresponding K-class change of interest, other change.
In one embodiment of the present disclosure, the preset long-term memory network is expressed as lstm= { I, H, O }, where I represents the number of nodes of the input layer of the preset long-term memory network and I is greater than or equal to S/2, H represents the number of hidden layers of the preset long-term memory network and H is greater than or equal to 1,O and 1 is greater than or equal to O is less than or equal to n-m.
In one embodiment of the present disclosure, the difference between the time sequence prediction data and the time sequence monitoring data is made as follows
Then->
In one embodiment of the present disclosure, the difference between the time sequence prediction data and the time sequence monitoring data is made as follows
The difference is +.>
Wherein (1)>
The preset long-term memory network is used for corresponding pixels in time phase { t epsilon t } s : s=1, 2,..s } and L is the training error multiple allowed by the prediction error.
In one embodiment of the present disclosure, the preprocessing includes radiation correction, orthographic correction, registration and resampling, cloud and shadow removal.
In one embodiment of the present disclosure, the sample of the supervised classification is the preprocessed satellite image pixel data containing the K-class interest variations.
In one embodiment of the present disclosure, the supervised classification is characterized by the periodic time-series spectral band images or remote sensing index images
In the spectrum band or remote sensing index i.
Referring to fig. 3, fig. 3 shows a hardware configuration diagram of an electronic device.
The electronic device described in the present embodiment includes:
the remote sensing image change detection method described in the embodiment shown in fig. 1 is implemented by the memory 31, the processor 32, and a computer program stored in the memory 31 and executable on the processor when the processor executes the program.
Further, the electronic device further includes:
at least one input device 33; at least one output device 34.
The memory 31, the processor 32 input device 33 and the output device 34 are connected by a bus 35.
The input device 33 may be a camera, a touch panel, a physical button, a mouse, or the like. The output device 34 may be specifically a display screen.
The Memory 31 may be a high-speed random access Memory (RAM, randomAccess Memory) or a non-volatile Memory (non-volatile Memory), such as a disk Memory. The memory 31 is for storing a set of executable program code and the processor 32 is coupled to the memory 31.
Further, the embodiment of the disclosure further provides a computer readable storage medium, which may be provided in the electronic device in the above embodiments, and the computer readable storage medium may be the electronic device in the embodiment shown in fig. 3. The computer readable storage medium has stored thereon a computer program which when executed by a processor implements the remote sensing image change detection method described in the embodiment shown in fig. 1. Further, the computer-readable medium may be a usb disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, etc. which may store the program code.
More, according to the remote sensing image change detection method, the remote sensing image change detection device, the electronic equipment and the readable storage medium provided by the disclosure, a research area is detected. Fig. 4 to fig. 6 may be specifically referred to, where fig. 4 is a schematic diagram of different time phase images and aperiodic change marks of a certain research area SAVI (soil-adjusted vegetation index) provided by the present disclosure, fig. 5 is a schematic diagram of aperiodic change marks in 4 remote sensing index images of a certain research area provided by the present disclosure, and fig. 6 is a schematic diagram of an interesting aperiodic change detection result of a certain research area provided by the present disclosure.
It should be noted that, each functional module in each embodiment of the present disclosure may be integrated into one processing module, or each module may exist alone physically, or two or more modules may be integrated into one module. The integrated modules may be implemented in hardware or in software functional modules.
The integrated modules, if implemented in the form of software functional modules and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such an understanding, the technical solution of the invention may be embodied essentially or partly in the form of a software product or in part in addition to the prior art.
It should be noted that, for the sake of simplicity of description, the foregoing method embodiments are all expressed as a series of combinations of actions, but it should be understood by those skilled in the art that the present invention is not limited by the order of actions described, as some steps may be performed in other order or simultaneously in accordance with the present invention. Further, those skilled in the art will appreciate that the embodiments described in the specification are all preferred embodiments, and that the acts and modules referred to are not necessarily all required for the present invention.
In the foregoing embodiments, the descriptions of the embodiments are emphasized, and for parts of one embodiment that are not described in detail, reference may be made to the related descriptions of other embodiments.
The foregoing describes a remote sensing image change detection method, apparatus, electronic device and readable storage medium provided by the present invention, and those skilled in the art should not understand the present invention to limit the scope of the present invention in terms of the specific implementation and application range according to the concepts of the embodiments of the present invention.
Claims (10)
1. A remote sensing image change detection method, comprising:
preprocessing the periodic time sequence remote sensing image, wherein the periodic time sequence remote sensing image of the t-th time phase is expressed as I t ，I t The time sequence length of (1) is represented as n, the period is represented as S, S is not less than 1 and not more than m, { I t ：t＝1，2，...，m，m+1，...，n}；
Extracting an image spectrum band from the preprocessed periodic time sequence remote sensing image or calculating a remote sensing index to obtain a periodic time sequence spectrum band image or a remote sensing index image, wherein the periodic time sequence spectrum band image or the remote sensing index image is expressed as
Is expressed as i,/or +.>
Extracting periodic time sequence data pixel by pixel and phase by phase from the periodic time sequence spectrum band image or the remote sensing index image, wherein the periodic time sequence dataRepresented as
(x, y) representing image pixel coordinates of the periodic time series data, < >>
Dividing the periodic time sequence data into two sections according to time phase t=m to respectively obtain time sequence training data and time sequence monitoring data, wherein the time sequence training data is that
The time sequence monitoring data is
Training a preset long-short-period memory network by utilizing the time sequence training data pixel by pixel, band by band or exponentially to obtain a periodic time sequence band or index data evolution model of a pixel level, wherein the periodic time sequence band or index data evolution model of the pixel level is expressed as
Predicting data except t=m in the time sequence monitoring data by using the periodic time sequence wave band or index data evolution model of the pixel level to obtain time sequence prediction data of each pixel in each wave band or remote sensing index, wherein the time sequence prediction data is expressed as
Calculating a difference between the time sequence prediction data and the time sequence monitoring data, and marking corresponding pixels as non-periodic variation data at corresponding time phases when the difference is within a preset range, wherein the non-periodic variation data is expressed as
Classifying the aperiodic variable data by using a supervision classification method, wherein the classification number of the aperiodic variable data is K+1, and obtaining K-class interesting change marks and one class of other change marks, wherein the K-class interesting change marks are expressed as
Marking the periodic time sequence remote sensing images { I }, respectively t : t=m+1,..n } each pixel, labeled as any one of no change, corresponding K-class change of interest, other change.
2. The remote sensing image change detection method according to claim 1, wherein the preset long-period memory network is expressed as lstm= { I, H, O }, where I represents the number of nodes of the input layer of the preset long-period memory network and I is greater than or equal to S/2, H represents the number of hidden layers of the preset long-period memory network and H is greater than or equal to 1,O and 1 is less than or equal to O is less than or equal to n-m.
3. The method according to claim 1, wherein the difference between the time-series prediction data and the time-series monitoring data is
Then->
4. The method according to claim 1, wherein the difference between the time-series prediction data and the time-series monitoring data is
The difference is +.>
Wherein (1)>
For the corresponding pixel when the preset long-period memory network is in time phase { t epsilon t } s : training error of s=1, 2,.. s All sets of phases for which the phase of the time series training data satisfies the S-th position condition within each period S are satisfied.
5. The method of claim 1, wherein the preprocessing includes radiation correction, orthographic correction, registration and resampling, cloud and shadow removal.
6. The method of claim 1 or 5, wherein the samples of the supervised classification are the preprocessed satellite image pixel data containing the K-class interest variations.
7. The method according to claim 1, wherein the feature of the supervised classification is the periodic time-series spectral band image or the remote sensing index image
In the spectrum band or remote sensing index i.
8. A remote sensing image change detection apparatus, comprising:
the preprocessing module is used for preprocessing the periodic time sequence remote sensing image, wherein the periodic time sequence remote sensing image of the t-th time phase is expressed as I t ，I t Is denoted as n, the period is denoted as S,1≤S≤m，{I t ：t＝1，2，...，m，m+1，...，n}；
A first extraction module, configured to extract an image spectrum band from the preprocessed periodic time-series remote sensing image or calculate a remote sensing index, to obtain a periodic time-series spectrum band image or a remote sensing index image, where the periodic time-series spectrum band image or the remote sensing index image is expressed as
Is denoted as i,
a second extraction module for extracting periodic time sequence data pixel by pixel and phase by phase from the periodic time sequence spectrum band image or the remote sensing index image, wherein the periodic time sequence data is expressed as
(x, y) representing image pixel coordinates of the periodic time series data, < >>
A segmentation module for dividing the periodic time sequence data into two segments according to time phase t=m to respectively obtain time sequence training data and time sequence monitoring data, wherein the time sequence training data is that
The time sequence monitoring data is
Training module for utilizing pixel by pixel, band by band or exponentiallyTraining a preset long-short-period memory network by using the time sequence training data to obtain a periodic time sequence wave band or index data evolution model of a pixel level, wherein the periodic time sequence wave band or index data evolution model of the pixel level is expressed as
The prediction module is used for predicting data except t=m in the time sequence monitoring data by using the periodic time sequence wave band or index data evolution model of the pixel level to obtain time sequence prediction data of each pixel in each wave band or remote sensing index, wherein the time sequence prediction data is expressed as
A calculation module for calculating the difference between the time sequence prediction data and the time sequence monitoring data, and marking the corresponding pixel as non-periodic variation data at the corresponding time phase when the difference is within a preset range, wherein the non-periodic variation data is expressed as
The classification module is used for classifying the aperiodic variable data with the category number of K+1 by using a supervision classification method to obtain K-class interesting change marks and one class of other change marks, wherein the K-class interesting change marks are expressed as
Marking modules for respectively marking the periodic time sequence remote sensing images { I } t : t=m+1,..n } each pixel, labeled as any one of no change, corresponding K-class change of interest, other change.
9. An electronic device, comprising: a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor, when executing the computer program, performs the steps of the remote sensing image change detection method of any one of claims 1 to 7.
10. A computer readable storage medium having stored thereon a computer program, wherein the computer program, when executed by a processor, implements the steps of the remote sensing image change detection method of any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011056145.5A CN112053359B (en) | 2020-09-30 | 2020-09-30 | Remote sensing image change detection method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011056145.5A CN112053359B (en) | 2020-09-30 | 2020-09-30 | Remote sensing image change detection method and device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112053359A CN112053359A (en) | 2020-12-08 |
| CN112053359B true CN112053359B (en) | 2023-05-26 |
Family
ID=73605419
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011056145.5A Active CN112053359B (en) | 2020-09-30 | 2020-09-30 | Remote sensing image change detection method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112053359B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112652028A (en) * | 2021-01-20 | 2021-04-13 | 四川测绘地理信息局测绘技术服务中心 | Method for extracting pine information of single plant infected pine wood nematode disease based on RGB image |
| CN113222997A (en) * | 2021-03-31 | 2021-08-06 | 上海商汤智能科技有限公司 | Neural network generation method, neural network image processing device, electronic device, and medium |
| CN114120137B (en) * | 2021-10-19 | 2023-07-25 | 桂林理工大学 | Time-sequence vegetation remote sensing image-based wetland element time-space evolution monitoring method |
| CN115795255B (en) * | 2022-09-21 | 2024-03-26 | 深圳大学 | Method, device, medium and terminal for detecting time sequence change of wetland |
| CN116912698B (en) * | 2023-09-04 | 2023-12-08 | 中国地质大学（武汉） | Land coverage change monitoring method and device, electronic equipment and storage medium |
| CN117253155B (en) * | 2023-11-17 | 2024-03-15 | 山东大学 | Human activity detection method and system based on deep learning |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10325370B1 (en) * | 2016-05-31 | 2019-06-18 | University Of New Brunswick | Method and system of coregistration of remote sensing images |
| CN111723732A (en) * | 2020-06-18 | 2020-09-29 | 西安电子科技大学 | Optical remote sensing image change detection method, storage medium and computing device |
-
2020
- 2020-09-30 CN CN202011056145.5A patent/CN112053359B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10325370B1 (en) * | 2016-05-31 | 2019-06-18 | University Of New Brunswick | Method and system of coregistration of remote sensing images |
| CN111723732A (en) * | 2020-06-18 | 2020-09-29 | 西安电子科技大学 | Optical remote sensing image change detection method, storage medium and computing device |
Non-Patent Citations (2)
| Title |
|---|
| Learning Spectral-Spatial-Temporal Features via a Recurrent Convolutional Neural Network for Change Detection in Multispectral Imagery;Lichao Mou,et al.;《IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING》;20190228;第924-935页 * |
| 基于深度学习的城市高分遥感图像变化检测方法的研究;陈璐，管霜霜;《计算机应用研究》;20200630;第320-323页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112053359A (en) | 2020-12-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112053359B (en) | Remote sensing image change detection method and device, electronic equipment and storage medium | |
| Tehrani et al. | Machine learning and landslide studies: recent advances and applications | |
| Zhao et al. | Monthly mapping of forest harvesting using dense time series Sentinel-1 SAR imagery and deep learning | |
| Chen et al. | Estimating tropical cyclone intensity by satellite imagery utilizing convolutional neural networks | |
| Feizizadeh et al. | A comparison of the integrated fuzzy object-based deep learning approach and three machine learning techniques for land use/cover change monitoring and environmental impacts assessment | |
| Xu et al. | Extraction of urban built-up areas from nighttime lights using artificial neural network | |
| Hamedianfar et al. | Large-scale urban mapping using integrated geographic object-based image analysis and artificial bee colony optimization from worldview-3 data | |
| Phaneendra Kumar et al. | Whale optimization-based band selection technique for hyperspectral image classification | |
| Tong et al. | Detection of urban sprawl using a genetic algorithm-evolved artificial neural network classification in remote sensing: a case study in Jiading and Putuo districts of Shanghai, China | |
| Sahoo et al. | Inter-calibration of DMSP-OLS and SNPP-VIIRS-DNB annual nighttime light composites using machine learning | |
| Li et al. | Sloping farmland identification using hierarchical classification in the Xi-He region of China | |
| Su et al. | Classification of MODIS images combining surface temperature and texture features using the Support Vector Machine method for estimation of the extent of sea ice in the frozen Bohai Bay, China | |
| Malik et al. | Computer vision models for comparing spatial patterns: understanding spatial scale | |
| Rhif et al. | Veg-W2TCN: a parallel hybrid forecasting framework for non-stationary time series using wavelet and temporal convolution network model | |
| Vashisht et al. | Crop yield prediction using improved extreme learning machine | |
| Okubo et al. | Land use/cover classification of a complex agricultural landscape using single-dated very high spatial resolution satellite-sensed imagery | |
| Chakraborty et al. | A reflection on image classifications for forest ecology management: towards landscape mapping and monitoring | |
| Dahiya et al. | A comparative analysis of different land-use and land-cover classifiers using hyperspectral data | |
| Gangappa et al. | Classification of land cover images using modified water wave Optimization-based hybrid classifier | |
| Xi et al. | Intra-annual land cover mapping and dynamics analysis with dense satellite image time series: A spatiotemporal cube based spatiotemporal contextual method | |
| Kumawat et al. | Time-Variant Satellite Vegetation Classification Enabled by Hybrid Metaheuristic-Based Adaptive Time-Weighted Dynamic Time Warping | |
| Wang et al. | High‐Resolution Remote‐Sensing Image‐Change Detection Based on Morphological Attribute Profiles and Decision Fusion | |
| Abbes et al. | Prediction changes for nonstationary multi-temporal satellite images using HMM | |
| CN112967286B (en) | Method and device for detecting newly added construction land | |
| Mewes et al. | An agent-based extension for object-based image analysis for the delineation of irrigated agriculture from remote sensing data |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |