CN108564002A - A kind of remote sensing image time series variation detection method and system - Google Patents

Abstract

An embodiment of the present invention provides a method for detecting changes in time series of remote sensing images, including: removing clouds from the time series images to be measured, and obtaining a cloud-free time series curve of each pixel of the cloud-free time series images in each band; obtaining The location of the seasonal breakpoint in each cloud-free time series curve; based on a preset screening algorithm, it is detected whether the change in the location of the seasonal breakpoint is a real change or a false change. A method and system for detecting changes in time series of remote sensing images provided by the embodiments of the present invention accurately determine the positions of seasonal breakpoints in the time series by using the spatial information of seasonal land surface changes, thereby judging the authenticity of the changes in the breakpoint positions , to achieve effective detection of time series.

Description

A method and system for detecting changes in time series of remote sensing images

technical field

Embodiments of the present invention relate to the technical field of remote sensing image processing, and in particular to a method and system for detecting changes in time series of remote sensing images.

Background technique

Earth observation satellites can provide large-scale surface remote sensing images for a long time, and have great potential in time series change detection. Time series change detection is to use a series of images to qualitatively analyze the time effect of a phenomenon and quantify its transformation. Currently, change detection technology based on medium time and spatial resolution is used in the detection of agricultural land changes, agricultural conditions and forest conditions. , Urban sprawl and other aspects have made great progress.

At present, there are various methods for time series change detection, which can be divided into two types according to their basic processing units: the first type is the processing method based on spatial scale, which takes the entire remote sensing image as the basic unit of analysis, and its Transform into independent components of the time spectrum to distinguish or detect perceptual change dynamics. Common analysis methods include principal component analysis (PCA), multivariate alteration detection (MAD), iteratively reweighted MAD (IR-MAD) and change vector analysis (change vector analysis). , CVA) etc. This type of method can better overcome noise through data dimensionality reduction, but this type of method rarely considers seasonal surface changes; the second type of method is based on direct analysis of pixels, and can detect a temporal mutation of a pixel. Identification of long-term trends and breakpoints, in which the BFAST (Breaks for Additive Season and Trend) algorithm based on econometric laws divides time-series data into three parts: trend, seasonality and noise, which is convenient to apply to other seasonal or non-seasonal Change detection in progress. The CCDC (Continuous Change Detection and Classification) algorithm can also simulate trends, seasonal changes and abrupt changes, and has been well applied in the detection of various land cover change types.

However, the mutation detection of the above two methods is carried out in the trend component after removing the seasonal influence, and the cloud and snow pixels retained in the trend component will be misjudged as the mutation pixel. Moreover, the above two analysis methods only consider the temporal changes of pixels, and do not effectively use the spatial information, especially the surrounding pixels that have a certain spatial correlation with the mutation pixel, and the surrounding pixels can reduce the analysis false alarm rate. The role of in is often not negligible, resulting in poor detection of time series changes.

Contents of the invention

Aiming at the problems existing in the prior art, embodiments of the present invention provide a method and system for detecting changes in time series of remote sensing images, which effectively utilize spatial information of seasonal land surface changes, thereby effectively detecting changes in time series of remote sensing images.

In the first aspect, an embodiment of the present invention provides a method for detecting changes in time series of remote sensing images, including:

S1. Declouding the time series image to be tested, and obtaining the cloud-free time series curve of each pixel in each band of the cloud-free time series image;

S2. Obtain the location of the seasonal breakpoint in each cloud-free time series curve;

S3. Based on a preset screening algorithm, detect whether the change in the position of the seasonal breakpoint is a real change or a false change.

In the second aspect, an embodiment of the present invention provides a remote sensing image time series change detection system, the system includes:

The cloud removal module is used to remove clouds from the time series image to be tested, and obtain the cloud-free time series curve of each pixel of the cloud-free time series image in each band;

A breakpoint obtaining module is used to obtain the seasonal breakpoint position in each cloudless time series curve;

The detection module is configured to detect whether the change in the position of the seasonal breakpoint is a real change or a false change based on a preset screening algorithm.

In the third aspect, the embodiment of the present invention provides a remote sensing image time series change detection device, including:

Processor, memory, communication interface and bus; Wherein, described processor, memory, communication interface complete mutual communication through described bus; Described memory stores the program order that can be executed by described processor, and described processing The program instruction can be called by the controller to execute the above-mentioned method for detecting changes in time series of remote sensing images.

In the fourth aspect, an embodiment of the present invention provides a computer program product, the computer program product includes a computer program stored on a non-transitory computer-readable storage medium, the computer program includes program instructions, and when the program instructions are executed When executed by a computer, the computer is made to execute the above method.

In a fifth aspect, an embodiment of the present invention provides a non-transitory computer-readable storage medium, where the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions cause the computer to execute the above method.

A method and system for detecting changes in time series of remote sensing images provided by the embodiments of the present invention accurately determine the positions of seasonal breakpoints in the time series by using the spatial information of seasonal land surface changes, thereby judging the authenticity of the changes in the breakpoint positions , to achieve effective detection of time series.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description These are some embodiments of the present invention. Those skilled in the art can also obtain other drawings based on these drawings without creative work.

Fig. 1 is a flow chart of a method for detecting changes in time series of remote sensing images provided by an embodiment of the present invention;

Fig. 2 is a schematic diagram of a cloudless time series decomposition provided by an embodiment of the present invention;

Fig. 3 is a schematic diagram of a change detection binary map provided by an embodiment of the present invention;

Fig. 4 is a structural diagram of a remote sensing image time series change detection system provided by an embodiment of the present invention;

Fig. 5 is a structural block diagram of a time series change detection device provided by an embodiment of the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are the Some, but not all, embodiments are invented. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

Fig. 1 is a flow chart of a method for detecting changes in time series of remote sensing images provided by an embodiment of the present invention. As shown in Fig. 1, the method includes:

S1. Declouding the time series image to be tested, and obtaining the cloud-free time series curve of each pixel in each band of the cloud-free time series image;

S2. Obtain the location of the seasonal breakpoint in each cloud-free time series curve;

S3. Based on a preset screening algorithm, detect whether the change in the position of the seasonal breakpoint is a real change or a false change.

In the embodiment of the present invention, the time series change detection refers to using a series of images to qualitatively analyze the time effect of a certain phenomenon. Generally, it is to find the breakpoint position of the time effect of the phenomenon, and adjust the time through the change of the breakpoint position. Sequences are analyzed.

It can be understood that by capturing the breakpoint position in the time series, it can be found whether an event has occurred in the object represented by the time series in the past, and then provide the necessary data support for detection.

In the existing technology, for time series change detection, the processing method based on spatial scale or the method of direct analysis based on pixel are generally used. These two methods are carried out in the trend component after removing the seasonal influence, and will The cloud and snow pixels retained in the trend component are misjudged as mutation pixels, but the surrounding pixels that actually have a certain spatial correlation with the mutation pixel play an important role in reducing the false alarm rate of the analysis.

In view of the above-mentioned problems in the prior art, the embodiment of the present invention adopts the time series change detection method as shown in FIG. 1 to realize effective detection of time series.

Specifically, in S1, the time series image to be measured is the object detected by the embodiment of the present invention. Generally, the embodiment of the present invention will collect the remote sensing image of the object to be detected. The remote sensing image is generally obtained by the earth observation satellite, and the commonly used For example: Landsat satellite.

Further, since remote sensing images are collected by satellites, usually some pixel positions are covered by clouds, and the pixels covered by clouds will affect the detection process of the embodiment of the present invention, so it is necessary to perform cloud removal operations on them, so that the measured The cloudy pixels in the time-series images are converted to cloud-free pixels to obtain cloud-free time-series images.

Furthermore, for time series images, there are usually multiple bands, each band has multiple scene images, and each scene image has multiple pixels, then for each image of each band element, its cloud-free time series curve in the time dimension can be obtained.

In S2, according to the multiple cloud-free time series curves obtained in S1, the seasonal breakpoint position in each cloud-free time series curve is obtained. It can be understood that a time series generally consists of three components, namely : Trend, season and random component. The trend refers to the state or law that the time series presents a continuous upward or downward trend. The season refers to the periodic fluctuation of the time series within a certain period of time. The random component refers to the time series. irregular fluctuations in .

In the embodiment of the present invention, only the information of the seasonal component is obtained by separating the trend, season and random components. This is because the trend component is a linear trend and does not contain breakpoint information, and the seasonal component is fitted by a cosine model, which has obvious Periodicity, and the breakpoint information is included in the seasonal component, and the effective detection of time series can be realized by detecting the authenticity of the position change of the breakpoint information.

In S3, it can be understood that the change of the seasonal breakpoint position determined in S2 is only a preliminary determination result. In order to verify the authenticity of the breakpoint position, it is necessary to select the true Varying breakpoints, thus providing the most accurate analytical breakpoints for time series detection.

A time series change detection method and system provided by the embodiments of the present invention accurately determine the position of the seasonal breakpoint in the time series by using the spatial information of the seasonal land surface change, thereby judging the authenticity of the change of the breakpoint position and realizing Efficient detection of time series.

The embodiment of the present invention takes the collected Landsat time-series image as an example to illustrate the embodiment of the present invention. The Landsat time-series image includes 6 bands. The embodiment of the present invention selects the Landsat TM5 image from 2001 to 2006. There are 138 scenes in total, and each scene image contains a plurality of pixels. In order to realize the corresponding cloud removal operation, the embodiment of the present invention downloads the corresponding cloud mask product (CFmask) of the Landsat image. Each pixel is marked, and the value range of the cloud mask product is 0 to 3, where 0 means that the corresponding pixel has no cloud, and 1 to 3 means that the corresponding pixel has cloud. Interpolation is performed on the cloud pixels, so that all the cloud pixels are converted into cloud-free pixels, so as to further obtain the cloud-free time series curve of each pixel in each band of the cloud-free time series image.

Further, the embodiment of the present invention separates the trend, season, and seasonal component of the random component in the time series, and obtains the seasonal breakpoint position in the seasonal component. Different time series curves contain different numbers of breakpoints. , the embodiment of the present invention needs to obtain the seasonal breakpoint position of each time series curve, for the breakpoint position of the cloudless time series image on the time series curve of each band, through the preset screening algorithm, for each breakpoint The position is screened, and the breakpoint whose breakpoint position is true changes is selected for breakpoint detection.

On the basis of the above-mentioned embodiments, step S1 specifically includes:

Mark each pixel of the time series image to be tested as cloudy or cloudless;

performing interpolation processing on the cloudy pixels, so that all cloudy pixels are converted into cloudless pixels, and the cloud-free time series images are obtained;

Obtain the cloud-free time-series curve of each pixel in each band of the cloud-free time-series image.

It can be understood that, for each pixel in the time series image to be measured, they can be divided into cloudy pixels and cloudless pixels. A cloudy pixel means that the pixel is covered by clouds, and cloudless images Element means that the pixel is not covered by clouds.

Further, through the interpolation of the cloudy pixels, the cloudy pixels can be converted into cloudless pixels, then the time series image to be tested can be converted into a cloudless time series image, and further, it can be obtained from the cloudless time series Obtain a cloud-free time series curve for each pixel in each band in the image.

Specifically, taking Landsat time-series images as an example, the embodiment of the present invention selects Landsat TM5 images from 2001 to 2006, a total of 138 scenes, and downloads the corresponding cloud mask product (CFmask) of the Landsat image at the same time, the cloud mask product Each pixel of each image is marked, and the value range of the cloud mask product is 0 to 3, where 0 indicates that the corresponding pixel is cloudless, and 1 to 3 indicates that the corresponding pixel has clouds, and then according to the interpolation formula Calculate the interpolated cell value, where, Indicates the pixel value after interpolation at time t, and Δt ₁ and Δt ₂ represent the number of days between the first cloud-free date in the next phase and the previous phase of the pixel, respectively.

Further, according to the formula Obtain the cloud-free time-series curve of each pixel of the cloud-free time-series image in each band, where, t∈[1,2,...,T], b∈[1,2,... ,B], loc∈[1,2,...,N], ts _{b, loc} represents the cloud-free time series curve of the locth pixel in the time dimension in the band b, then the processed 2001 to 2006 The Landsat TM5 image in 2010 contains 138 time phases, each image has 6 bands and 1,000,000 pixels.

On the basis of the foregoing embodiments, step S2 specifically includes:

Extract the seasonal component in each cloud-free time series curve;

Based on the structural change test method, the number and location of seasonal breakpoints in the seasonal components are determined.

It is understandable that the time series is usually composed of three components: trend, season, and random components. If these three components can be separated, they can be analyzed one by one, thereby simplifying the detection process.

Specifically, in step S2 provided by the embodiment of the present invention, to extract the seasonal component in each cloud-free time series curve, first use the formula ts _b,loc =T _t +S _t +e _t (t=1,... ,T) Decompose the cloud-free time series curve, where T _t represents the linear trend component, T _t = a _i + b _i t, a represents the intercept of the linear trend, b represents the slope of the linear trend, S _t represents the seasonal component, e _t represents a random component.

FIG. 2 is a schematic diagram of the decomposition of the cloudless time series provided by the embodiment of the present invention. As shown in FIG. 2 , the cloudless time series curve is decomposed into three components: trend, season, and noise. The trend component is a linear trend and does not contain breakpoint information. The seasonal component contains breakpoint information. The seasonal component is fitted by a cosine model and has obvious periodicity. In the embodiment of the present invention, the period of the cloudless time series curve After 3 iterations, it is determined that there is a breakpoint between 2005 and 2006 in the seasonal component, and the position of the breakpoint has a certain degree of reliability within a certain time range.

What needs to be explained is that in the real environment, climate change presents a certain regularity, so the earth surface also presents a certain seasonal regularity. When the local features change, the seasonal components in the time series curve will produce corresponding breakpoints.

Then, the embodiment of the present invention can realize the breakpoint detection of the time series by obtaining the number and position of the seasonal breakpoints in the seasonal component. In the embodiment of the present invention, the seasonal component is used for express:

Will expands to: thereby according to Computes the seasonal component, where S _t represents the breakpoint and between seasonal components, K represents the number of cosine models, f represents the number of image acquisitions in a year, parameter γ _jk represents the amplitude, parameter θ _jk represents the phase, specifically, in the embodiment of the present invention, preferably K is set to 3 and f is set to 23.

It can be understood that in different time series curves, the number of breakpoints included is different. Since the number of Landsat images in the example of the present invention is 23 scenes per year, including 6 years of images, the maximum number of breakpoints is set is 5, the seasonal component is divided into 6 segments at most, and when the position of the breakpoint is determined, the seasonal component between the breakpoints can be obtained by the curve fitting method.

Specifically, the embodiment of the present invention adopts the method of structural change test to obtain the number and position of seasonal breakpoints in the seasonal components.

First, the embodiment of the present invention uses MOSUM to test the seasonal components. If the MOSUM test shows that there are breakpoints in the seasonal components, the number and positions of the seasonal breakpoints are determined according to the structural change test method, wherein, The seasonal coefficients γ _jk , θ _jk can be calculated from the regression equation based on the least squares method.

Specifically, the embodiment of the present invention is based on the formula

Obtain the summation result of the moving window on the time series curve, and obtain the number of seasonal breakpoints and the position (time) of occurrence according to the formula sctest=max||TS_SUM(t)||, where TS_SUM represents the summation of the moving window A vector, m represents the size of the moving window.

In the embodiment of the present invention, the seasonal component is firstly calculated with the fluctuation rule by the R language. The function of the MOSUM method is efp(ts _{b, loc} -T _t ~ t, h=h, type='MOSUM'), for each time Sequence curve, make ts _{b, loc} -T _t fit with the corresponding time t, h is the reciprocal of the number of divisions of the seasonal component, preferably, in the embodiment of the present invention, it is set to 1/6, so in the MOSUM method, the corresponding The moving window size is set to 23.

Next, the embodiment of the present invention judges the number and position of seasonal breakpoints by the structural change test method in the R language, and the function of the structural change test method is sctest(efp(ts _{b, loc} - _{T t} ~ t, h=h ,type='MOSUM')), according to the time series volatility rule calculated by the above MOSUM method, record the position corresponding to the maximum value of the fluctuation value as the position of the seasonal breakpoint, in the process of obtaining the seasonal breakpoint , for the seasonal coefficient γ _jk , the calculation process of θ _jk is:

According to the formula α＝[γ _j.1 ,θ _j.1 ,γ _j.2 ,θ _j.2 ,...,γ _jK ,θ _jK ], according to the least squares formula Calculate the seasonality coefficient, where Indicates the time series curve from breakpoint j-1 to breakpoint j, Indicates the seasonal component from breakpoint j-1 to breakpoint j.

Then, through the acquisition process of the above-mentioned seasonal component and the calculation process of the seasonal breakpoint, the number and position of the seasonal breakpoint corresponding to each cloud-free time series can be determined.

It can be understood that, since the results of the testing process may have errors, the embodiment of the present invention preferably uses the result that the number and position of the breakpoints of two adjacent breakpoints no longer changes as the final determination result of the breakpoint position.

On the basis of the foregoing embodiments, step S3 includes:

Carrying out first-level screening for the breakpoint position of the time series curve of each band, the first-level screening is to count the number of bands to which each breakpoint position belongs;

If the number of bands containing the same breakpoint position is greater than the first preset threshold, it is determined that the change in the breakpoint position is a true change;

If the number of bands including the same breakpoint position is less than the first preset threshold, it is determined that the breakpoint position is a false change.

It can be understood that the breakpoint position determined in step S2 is only a preliminary determination process, and the embodiment of the present invention needs to reliably screen the determined breakpoint position, so as to judge the authenticity of the change of each breakpoint position to provide reliable The breakpoint position with the highest reliability.

Specifically, the embodiment of the present invention will perform first-level screening on the breakpoint position of the time series curve of each band. It can be understood that if the time series time curve of each band is determined to be a breakpoint position at a certain position , then the reliability of this point as a breakpoint position should be the highest. On the contrary, if a certain position is judged as a breakpoint position in only a few bands, then the change of this point is likely to be a pseudo-change, that is, the point is a breakpoint The reliability of the point is low.

Specifically, in the embodiment of the present invention, a change in the breakpoint position whose number of bands at the same breakpoint position is greater than the first preset threshold is determined as a true change, and the number of bands containing the same breakpoint position is less than the first preset threshold The breakpoint position of is determined as a pseudo-change.

For example: the Landsat image has 6 bands in total, and each pixel can obtain the time series curves of 6 bands, and the breakpoint positions of the time series curves. In the example of the present invention, if there are 3 or more than 3 bands, it can be judged If a breakpoint occurs at the same position, the breakpoint position is a real change; if only one or two bands determine that a breakpoint occurs somewhere, the breakpoint position is a false change. It can be understood that 3 is the first preset threshold set in the embodiment of the present invention, but the embodiment of the present invention does not limit the specific value of the first threshold, which can be adjusted accordingly according to different actual processing objects.

On the basis of the above-mentioned embodiment, after performing the first-level screening on the breakpoint position of the time series curve of each band, step S3 also includes:

The breakpoint position determined to be a true change in the first-level screening is used as the candidate breakpoint position for the second-level screening, and the second-level screening is: based on the preset change intensity calculation model, calculate the candidate breakpoint position corresponding to If the change intensity is greater than the second preset threshold, it is determined that the change in the position of the candidate breakpoint is a true change;

If the change intensity is smaller than a second preset threshold, it is determined that the change of the position of the candidate breakpoint is a false change.

It can be understood that, on the basis of the first-level screening performed in the above embodiments, the embodiment of the present invention also provides a second-level screening, so as to more accurately determine the authenticity of the breakpoint position change.

Specifically, the second-level screening process is as follows:

The parameters A=[α _1L ,α _2L ,α _3L ,α _4L ,α _5L ,α _6L ] of the six bands corresponding to Landsat on the left side of the breakpoint and the parameter B=[α _1R , α _2R , α _3R , α _4R , α _5R , α _6R ] as input parameters, through the change vector detection model and Obtain the change intensity |D|, where α _b represents the seasonal component parameter at band b, and judge whether there is a change according to the preset threshold value. If the change intensity is greater than the second preset threshold value, it is determined that the breakpoint position is a true change. If the change If the intensity is less than the second preset threshold, the position of the breakpoint is a false change, and the second preset threshold can be set according to actual conditions.

It can be understood that the candidate breakpoint positions are the breakpoint positions determined as true changes in the first-level screening.

On the basis of the above-mentioned embodiments, after performing the second-level screening on the breakpoint positions determined to be true changes in the first-level screening as candidate breakpoint positions, step S3 also includes:

The breakpoint position judged to be a true change in the second-level screening is used as the target breakpoint position to perform the third-level screening. The third-level screening is: obtain the change detection binary map corresponding to the target breakpoint position, if the If the number of pixels changed in the change detection binary image in the neighborhood window is greater than the third preset threshold, it is determined that the change in the position of the target breakpoint is a true change;

If the number of pixels changed in the change detection binary image within the domain window is less than a third preset threshold, it is determined that the change in the position of the target breakpoint is a false change.

It can be understood that, on the basis of the first-level screening and the second-level screening in the above embodiments, the embodiment of the present invention also provides a third-level screening, so as to more accurately determine the authenticity of the breakpoint position change.

In the embodiment of the present invention, the breakpoint position determined as a true change in the second-level screening is defined as the target breakpoint position in the embodiment of the present invention, and the change detection binary map of the target breakpoint position is counted in the spatial dimension , as shown in FIG. 3 . FIG. 3 is a schematic diagram of a change detection binary image provided by an embodiment of the present invention, wherein black represents changed pixels, and white represents unchanged pixels.

After obtaining the change detection binary map, it is necessary to detect the number of changed pixels contained in each window. It can be understood that the size of the window is preset in the embodiment of the present invention. If the change detection binary map neighborhood window The number of changed pixels is greater than or equal to the third preset threshold, then the breakpoint position is a true change, otherwise, if the number of changed pixels in the neighborhood window of the step change detection binary image is less than the third preset threshold, Then the position of the breakpoint is a pseudo-change.

For example: set the size of the neighborhood window to 3*3, if the number of changing pixels in the window is greater than or equal to 4, then the pixel in the center of the window is a real change, if the number of changing pixels in the window is less than 4, then the center of the window Pseudo-variation at the pixel. Then 4 is the third preset threshold described in the embodiment of the present invention. Similarly, the embodiment of the present invention does not limit the specific value of the third preset threshold, which can be adjusted accordingly according to different actual processing objects.

Through the three-level screening scheme provided by the embodiment of the present invention, the position of the breakpoint can be accurately determined, thereby improving the accuracy of time series detection.

Fig. 4 is a structural diagram of a remote sensing image time series change detection system provided by an embodiment of the present invention. As shown in Fig. 4, the system includes: a cloud removal module 1, a breakpoint acquisition module 2 and a detection module 3, wherein:

The cloud removal module 1 is used to remove clouds from the time series image to be tested, and obtain the cloud-free time series curve of each pixel of the cloud-free time series image in each band;

Breakpoint acquisition module 2 is used to obtain the seasonal breakpoint position in each cloudless time series curve;

The detection module 3 is used to detect whether the change in the position of the seasonal breakpoint is a real change or a false change based on a preset screening algorithm.

Specifically, how to use the cloud removal module 1, the breakpoint acquisition module 2 and the detection module 3 to communicate with the vehicle data can be used to implement the technical solution of the embodiment of the time series change detection method shown in Figure 1, its implementation principle and technical effect are similar, here I won't repeat them here.

A remote sensing image time series change detection system provided by an embodiment of the present invention accurately determines the position of the seasonal breakpoint in the time series by using the spatial information of the seasonal land surface change, thereby judging the authenticity of the change of the breakpoint position and realizing Efficient detection of time series.

An embodiment of the present invention provides a remote sensing image time series change detection device, including: at least one processor; and at least one memory connected to the processor in communication, wherein:

FIG. 5 is a structural block diagram of a time series change detection device provided by an embodiment of the present invention. Referring to FIG. 5, the time series change detection device includes: a processor (processor) 810, a communication interface (CommunicationsInterface) 820, and a memory 830 and the bus 840 , wherein the processor 810 , the communication interface 820 , and the memory 830 communicate with each other through the bus 840 . The communication interface 820 can be used for information transmission between the server and the time series change detection device. The processor 810 can call the logic instructions in the memory 830 to perform the following method: S1. Declouding the time-series image to be tested to obtain the cloud-free time-series curve of each pixel of the cloud-free time-series image in each band; S2. Obtain the position of the seasonal breakpoint in each cloud-free time series curve; S3. Based on a preset screening algorithm, detect whether the change in the position of the seasonal breakpoint is a real change or a false change.

An embodiment of the present invention discloses a computer program product, the computer program product includes a computer program stored on a non-transitory computer-readable storage medium, the computer program includes program instructions, and when the program instructions are executed by a computer, The computer can execute the method provided by the above-mentioned method embodiments, for example, including: S1, performing cloud removal on the time series image to be tested, and obtaining the cloud-free time series curve of each pixel of the cloud-free time series image in each band; S2 . Acquiring the position of the seasonal breakpoint in each cloud-free time series curve; S3. Based on a preset screening algorithm, detecting whether the change in the position of the seasonal breakpoint is a real change or a false change.

An embodiment of the present invention provides a non-transitory computer-readable storage medium, the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions cause the computer to execute the methods provided in the above method embodiments, for example Including: S1. Declouding the time series image to be tested to obtain the cloud-free time-series curve of each pixel in each band of the cloud-free time-series image; S2. Obtaining the seasonal breakpoint in each cloud-free time-series curve Position; S3. Based on a preset screening algorithm, detect whether the change in the position of the seasonal breakpoint is a real change or a false change.

Those of ordinary skill in the art can understand that all or part of the steps for realizing the above-mentioned method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the It includes the steps of the above method embodiments; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other various media that can store program codes.

Through the above description of the implementations, those skilled in the art can clearly understand that each implementation can be implemented by means of software plus a necessary general hardware platform, and of course also by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the prior art can be embodied in the form of software products, and the computer software products can be stored in computer-readable storage media, such as ROM/RAM, magnetic discs, optical discs, etc., including several instructions to make a computer device (which may be a personal computer, server, or network device, etc.) execute the methods described in various embodiments or some parts of the embodiments.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it can still be Modifications are made to the technical solutions described in the foregoing embodiments, or equivalent replacements are made to some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims (10)

1. a kind of remote sensing image time series variation detection method, which is characterized in that including：

S1, cloud is carried out to time series image to be measured, obtain cloudless time series image each wave band each pixel Cloudless time-serial position；

S2, seasonal breakpoint location in each cloudless time-serial position is obtained；

S3, it is based on preset filtering algorithm, the variation for detecting the seasonal breakpoint location is true variation or pseudo- variation.

2. according to the method described in claim 1, it is characterized in that, step S1 is specifically included：

Each pixel of time series image to be measured is labeled as to have cloud pixel or cloudless pixel；

There is cloud pixel to carry out interpolation processing to described, so that all have cloud pixel to be converted to cloudless pixel, when obtaining described cloudless Between sequential images；

Obtain the cloudless time-serial position of each pixel of the cloudless time series image in each wave band.

3. according to the method described in claim 1, it is characterized in that, step S2 is specifically included：

Seasonal ingredient in each cloudless time-serial position of extraction；

Based on structure change method of testing, the number of seasonal breakpoint and position in the seasonal ingredient are determined.

4. according to the method described in claim 1, it is characterized in that, step S3 includes：

First order screening is carried out to the breakpoint location of the time-serial position of each wave band, the first order screening is to each Wave band number belonging to breakpoint location is counted, if including the wave band number of same breakpoint location be more than the first predetermined threshold value, Then judge that the breakpoint location variation is really to change；

If including the wave band number of same breakpoint location is less than first predetermined threshold value, judge that the breakpoint location becomes for puppet Change.

5. according to the method described in claim 4, it is characterized in that, the time-serial position to each wave band breakpoint After position carries out first order screening, step S3 further includes：

It is determined as that the breakpoint location really changed carries out second level screening as Candidate point position during the first order is screened, described the Two level is screened：Based on preset change intensity computation model, the corresponding change intensity in the Candidate point position is calculated, if institute It states change intensity and is more than the second predetermined threshold value, then judge that the Candidate point change in location is really to change；

If the change intensity is less than the second predetermined threshold value, judge that the Candidate point change in location changes for puppet.

6. according to the method described in claim 5, it is characterized in that, being determined as the breakpoint bit really changed in screening the first order It sets after carrying out second level screening as Candidate point position, step S3 further includes：

It is determined as that the breakpoint location really changed carries out third level screening as target breakpoint location during the second level is screened, described the Three level screen is：The corresponding variation detection binary map of the target breakpoint location is obtained, if the variation detects binary map in neighbour The pixel number changed in the window of domain is more than third predetermined threshold value, then judges that the target breakpoint location variation is really to change；

If the pixel number that changes in the window of field of variation detection binary map is less than third predetermined threshold value, described in judgement The variation of target breakpoint location is pseudo- variation.

7. a kind of remote sensing image time series variation detecting system, which is characterized in that the system comprises：

Cloud module is gone, for carrying out cloud to time series image to be measured, obtains cloudless time series image in each wave band The cloudless time-serial position of each pixel；

Breakpoint acquisition module, for obtaining seasonal breakpoint location in each cloudless time-serial position；

Detection module, for being based on preset filtering algorithm, the variation for detecting the seasonal breakpoint location is true variation or pseudo- Variation.

8. a kind of computer equipment, which is characterized in that including memory and processor, the processor and the memory pass through Bus completes mutual communication；The memory is stored with the program instruction that can be executed by the processor, the processor Described program instruction is called to be able to carry out the method as described in claim 1 to 6 is any.

9. a kind of computer program product, which is characterized in that the computer program product includes being stored in non-transient computer Computer program on readable storage medium storing program for executing, the computer program include program instruction, when described program is instructed by computer When execution, the computer is made to execute such as claim 1 to 6 any one of them method.

10. a kind of non-transient computer readable storage medium, which is characterized in that the non-transient computer readable storage medium is deposited Computer instruction is stored up, the computer instruction makes the computer execute such as claim 1 to 6 any one of them method.