A method and device for detecting and monitoring changes in buildings and structures over a long period of time.

Abstract

This invention provides a method and apparatus for detecting and monitoring changes in buildings and structures over a long period of time, relating to the technical field of natural resource change monitoring. The method includes: acquiring sample change patch vectors and long-term high-resolution remote sensing image data; determining the long-term high-resolution remote sensing image data slice set corresponding to each change patch in the sample change patch vector; determining the feature data corresponding to the long-term high-resolution remote sensing image data slice set; training a preset random forest model using the feature data and the corresponding building and structure change categories to obtain a target random forest model; and determining the change progress assessment result corresponding to the change patch vector to be processed using the change patch vector to be processed, the long-term high-resolution remote sensing image data corresponding to the change patch vector to be processed, and the target random forest model. This solves the technical problem that existing technologies struggle to assess the progress of changes in buildings and structures.

Description

Technical Field

[0001] This invention relates to the technical field of natural resource change monitoring, and in particular to a method and apparatus for detecting and long-term tracking changes in buildings and structures. Background Technology

[0002] Natural resources, including forests, arable land, and water, are vital to the national economy and people's livelihoods. However, with economic development and accelerated urbanization, the misuse and abuse of natural resources, such as illegal construction, deforestation, and illegal occupation of arable land, have become increasingly serious, causing enormous losses to natural resources. Natural resource monitoring and supervision refers to the work of understanding changes in natural resources themselves and those caused by human activities, based on baseline data from basic and special surveys, to achieve the regulatory goal of "early detection, early prevention, and strict crackdown."

[0003] A crucial aspect of natural resource monitoring and regulation is the discovery and continuous monitoring of changes in natural resources. Buildings and structures are an important type of natural resource monitoring feature. Traditional methods of discovery and continuous monitoring rely on manual screening and public reports, which are inefficient, inaccurate, and complex to process. With the development of remote sensing technology, rapid, large-scale, and short-cycle monitoring of natural resource changes can be achieved. Currently, the most mature method in the industry is based on pre- and post-contemporary imagery, using algorithms such as spectral analysis, machine learning, and deep learning to extract information on changes in buildings and structures. However, this method can only obtain coarse-scale, static information on changes in buildings and structures over two periods, which cannot meet the operational requirements of natural resource monitoring and regulation.

[0004] No effective solutions have yet been proposed to address the above problems. Summary of the Invention

[0005] In view of this, the purpose of the present invention is to provide a method and apparatus for detecting and monitoring changes in buildings and structures over a long period of time, so as to alleviate the technical problem that the prior art is unable to assess the changing state of buildings and structures.

[0006] In a first aspect, embodiments of the present invention provide a method for detecting and monitoring changes in buildings and structures over a long period of time, comprising: acquiring sample change patch vectors and long-term high-resolution remote sensing image data corresponding to the sample change patch vectors; based on the sample change patch vectors and the long-term high-resolution remote sensing image data corresponding to the sample change patch vectors, determining a set of long-term high-resolution remote sensing image data slices corresponding to each change patch in the sample change patch vectors, and determining feature data corresponding to the long-term high-resolution remote sensing image data slices, wherein the feature data includes: geometric feature data and spectral feature data; utilizing the feature data and The building change categories corresponding to the feature data are used to train a preset random forest model to obtain a target random forest model. After obtaining the change patch vector to be processed and the long-term high-resolution remote sensing image data corresponding to the change patch vector to be processed, the change process evaluation result corresponding to the change patch vector to be processed is determined by using the change patch vector to be processed, the long-term high-resolution remote sensing image data corresponding to the change patch vector to be processed, and the target random forest model. The change process evaluation result is used to characterize the start and end phases of the building category change of the change patch in the change patch vector to be processed.

[0007] Further, based on the sample change patch vector and the corresponding long-term high-resolution remote sensing image data, a set of long-term high-resolution remote sensing image data slices corresponding to each change patch in the sample change patch vector is determined, including: preprocessing the long-term high-resolution remote sensing image data corresponding to the sample change patch vector based on the sample change patch vector to obtain the target long-term high-resolution remote sensing image data corresponding to the sample change patch vector, wherein the preprocessing includes: orthorectification processing, resolution unification processing, and coordinate reference system unification processing; and slicing the long-term high-resolution remote sensing image data corresponding to the sample change patch vector based on the position and range information of each change patch in the sample change patch vector to obtain a set of long-term high-resolution remote sensing image data slices corresponding to each change patch in the sample change patch vector.

[0008] Further, the feature data corresponding to the long-term high-resolution remote sensing image data slice set is determined, including: based on the long-term high-resolution remote sensing image data slice set, determining the geometric feature data between the slice corresponding to the first time in the long-term high-resolution remote sensing image data slice set and the target slice in the long-term high-resolution remote sensing image data slice set, wherein the target slice includes: slices in the long-term high-resolution remote sensing image data slice set other than the slice corresponding to the first time; based on the long-term high-resolution remote sensing image data slice set, determining the spectral feature data between the slice corresponding to the first time in the long-term high-resolution remote sensing image data slice set and the target slice in the long-term high-resolution remote sensing image data slice set, wherein the target slice includes: slices in the long-term high-resolution remote sensing image data slice set other than the slice corresponding to the first time.

[0009] Further, based on the long-term high-resolution remote sensing image data slice set, the geometric feature data between the slice corresponding to the first time in the long-term high-resolution remote sensing image data slice set and the target slice in the long-term high-resolution remote sensing image data slice set are determined, including: performing principal component analysis on the slice corresponding to the first time and the target slice respectively to obtain the first principal component of the slice corresponding to the first time and the first principal component of the target slice; using the Canny model, determining the edge intensity map corresponding to the first principal component of the slice corresponding to the first time and the edge intensity map corresponding to the first principal component of the target slice respectively; calculating the ratio feature data between the edge intensity map corresponding to the first principal component of the slice corresponding to the first time and the edge intensity map corresponding to the first principal component of the target slice; determining the change patch corresponding to the slice corresponding to the first time and the target slice, and determining the aspect ratio of the bounding rectangle corresponding to the change patch and the area ratio between the change patch and the bounding rectangle; and determining the ratio feature data, the aspect ratio, and the area ratio as the geometric feature data.

[0010] Further, based on the long-term high-resolution remote sensing image data slice set, the spectral feature data between the slice corresponding to the first time in the long-term high-resolution remote sensing image data slice set and the target slice in the long-term high-resolution remote sensing image data slice set are determined, including: performing HSV conversion on the true color bands of the slice corresponding to the first time and the true color bands of the target slice, respectively, to obtain a first conversion result and a second conversion result, wherein the true color bands include: red band, green band and blue band; based on the first conversion result, the second conversion result and the rgb2hsv model, determining the first brightness component map of the slice corresponding to the first time and the second brightness component map of the target slice; based on the first brightness component map and the second brightness component map, determining the spectral feature data, wherein the spectral feature data includes: the brightness ratio between the first brightness component map and the second brightness component map, the standard deviation ratio of the first brightness component map, the standard deviation ratio of the second brightness component map, and the standard deviation ratio between the first brightness component map and the second brightness component map.

[0011] Further, using the unprocessed change patch vector, the corresponding long-term high-resolution remote sensing image data, and the target random forest model, the change process assessment result corresponding to the unprocessed change patch vector is determined, including: based on the unprocessed change patch vector and the corresponding long-term high-resolution remote sensing image data, determining the long-term high-resolution remote sensing image data slice set corresponding to each change patch in the unprocessed change patch vector, and determining the long-term high-resolution remote sensing image data slice set corresponding to each change patch in the unprocessed change patch vector; inputting the long-term high-resolution remote sensing image data slice set corresponding to each change patch in the unprocessed change patch vector into the target random forest model to obtain the change process assessment result corresponding to the unprocessed change patch vector.

[0012] Secondly, embodiments of the present invention also provide a device for detecting and monitoring changes in buildings and structures over a long period of time, comprising: an acquisition unit, configured to acquire sample change patch vectors and long-term high-resolution remote sensing image data corresponding to the sample change patch vectors; a determination unit, configured to, based on the sample change patch vectors and the long-term high-resolution remote sensing image data corresponding to the sample change patch vectors, determine a set of long-term high-resolution remote sensing image data slices corresponding to each change patch in the sample change patch vectors, and determine feature data corresponding to the long-term high-resolution remote sensing image data slices, wherein the feature data includes: geometric feature data and spectral feature data; and a training unit, configured to utilize... The feature data and the corresponding building change categories are used to train a preset random forest model to obtain a target random forest model. The evaluation unit is used to determine the change process evaluation result corresponding to the change patch vector to be processed after acquiring the change patch vector to be processed and the long-term high-resolution remote sensing image data corresponding to the change patch vector to be processed, using the change patch vector to be processed, the long-term high-resolution remote sensing image data corresponding to the change patch vector to be processed, and the target random forest model. The change process evaluation result is used to characterize the start and end phases of the building category change of the change patch in the change patch vector to be processed.

[0013] Further, the determining unit is configured to: preprocess the long-term high-resolution remote sensing image data corresponding to the sample change patch vector based on the sample change patch vector to obtain the target long-term high-resolution remote sensing image data corresponding to the sample change patch vector, wherein the preprocessing includes: orthorectification processing, resolution unification processing, and coordinate reference system unification processing; and slice the long-term high-resolution remote sensing image data corresponding to the sample change patch vector based on the position and range information of each change patch in the sample change patch vector to obtain a long-term high-resolution remote sensing image data slice set corresponding to each change patch in the sample change patch vector.

[0014] Thirdly, embodiments of the present invention also provide an electronic device, including a memory and a processor, wherein the memory is used to store a program that supports the processor in executing the method described in the first aspect above, and the processor is configured to execute the program stored in the memory.

[0015] Fourthly, embodiments of the present invention also provide a computer-readable storage medium on which a computer program is stored.

[0016] In this embodiment of the invention, sample change patch vectors and corresponding long-term high-resolution remote sensing image data are acquired; based on the sample change patch vectors and corresponding long-term high-resolution remote sensing image data, a set of long-term high-resolution remote sensing image data slices corresponding to each change patch in the sample change patch vectors is determined, and feature data corresponding to the set of long-term high-resolution remote sensing image data slices is determined, wherein the feature data includes geometric feature data and spectral feature data; using the feature data and the building change categories corresponding to the feature data, a preset random forest model is trained to obtain a target random forest model; after acquiring the change patch vectors to be processed and the corresponding long-term high-resolution remote sensing image data, a set of long-term high-resolution remote sensing image data slices corresponding to each change patch in the sample change patch vectors is determined, and feature data corresponding to the long-term high-resolution remote sensing image data slices is determined, wherein the feature data includes geometric feature data and spectral feature data; using the feature data and the building change categories corresponding to the feature data, a preset random forest model is trained to obtain a target random forest model; after acquiring the change patch vectors to be processed and the corresponding long-term high-resolution remote sensing image data, a set of long-term high-resolution remote sensing image data slices corresponding to each change patch in the sample change patch vectors is determined, and feature data corresponding to each long-term high-resolution remote sensing image slices is determined, wherein the long-term high-resolution remote sensing image data slices corresponding to each change patch in the sample change patch vectors are determined, and feature data corresponding to each long-term high-resolution remote sensing image slices are ... After processing the long-term high-resolution remote sensing image data corresponding to the change patch vector, the change process assessment result corresponding to the change patch vector is determined using the change patch vector to be processed, the long-term high-resolution remote sensing image data corresponding to the change patch vector to be processed, and the target random forest model. The change process assessment result is used to characterize the start and end phases of the change in the building category of the change patch in the change patch vector to be processed, thereby achieving the purpose of determining the occurrence time and subsequent changes of the building category. This solves the technical problem that existing technologies are difficult to assess the progress of building changes, and thus provides technical support for the time-series tracing and tracking of natural resource monitoring and supervision.

[0017] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained in accordance with the structures particularly pointed out in the description, claims and drawings.

[0018] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0019] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0020] Figure 1 A flowchart illustrating a method for detecting and monitoring changes in buildings and structures over a long period of time, provided in an embodiment of the present invention;

[0021] Figure 2A schematic diagram of a building change detection and long-term tracking monitoring device provided in an embodiment of the present invention;

[0022] Figure 3 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Example 1:

[0025] According to an embodiment of the present invention, an embodiment of a method for detecting and long-term tracking monitoring changes in buildings and structures is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0026] Figure 1 This is a flowchart of a method for detecting and monitoring changes in buildings and structures over a long period of time, according to an embodiment of the present invention. Figure 1 As shown, the method includes the following steps:

[0027] Step S102: Obtain the sample change patch vector and the long-term high-resolution remote sensing image data corresponding to the sample change patch vector;

[0028] It should be noted that the sample change patch vector B records information such as the range, location, preceding and following time phases of the change, and the type of change. Then, based on the layer range of the change patch vector, long-term high-resolution remote sensing image data within the corresponding range is obtained. The time interval for the long-term series is set according to the enforcement objective, including time intervals of days, months, quarters, and years; the time range of the long-term series is such that the starting time phase is the same as the preceding time phase of vector B, and the ending time phase is greater than or equal to the following time phase of the sample change patch vector B. The final obtained long-term high-resolution remote sensing image data is {A1, A2, ..., A...}. n}

[0029] Step S104: Based on the sample change patch vector and the long-term high-resolution remote sensing image data corresponding to the sample change patch vector, determine the long-term high-resolution remote sensing image data slice set corresponding to each change patch in the sample change patch vector, and determine the feature data corresponding to the long-term high-resolution remote sensing image data slice set, wherein the feature data includes: geometric feature data and spectral feature data;

[0030] Step S106: Using the feature data and the building change categories corresponding to the feature data, train the preset random forest model to obtain the target random forest model;

[0031] Step S108: After obtaining the vector of the changed patch to be processed and the long-term high-resolution remote sensing image data corresponding to the vector of the changed patch to be processed, the change process evaluation result corresponding to the vector of the changed patch to be processed is determined by using the vector of the changed patch to be processed, the long-term high-resolution remote sensing image data corresponding to the vector of the changed patch to be processed and the target random forest model. The change process evaluation result is used to characterize the start and end phases of the change in the building category of the changed patch in the vector of the changed patch to be processed.

[0032] In this embodiment of the invention, sample change patch vectors and corresponding long-term high-resolution remote sensing image data are acquired; based on the sample change patch vectors and corresponding long-term high-resolution remote sensing image data, a set of long-term high-resolution remote sensing image data slices corresponding to each change patch in the sample change patch vectors is determined, and feature data corresponding to the set of long-term high-resolution remote sensing image data slices is determined, wherein the feature data includes geometric feature data and spectral feature data; using the feature data and the building change categories corresponding to the feature data, a preset random forest model is trained to obtain a target random forest model; after acquiring the change patch vectors to be processed and the corresponding long-term high-resolution remote sensing image data, a set of long-term high-resolution remote sensing image data slices corresponding to each change patch in the sample change patch vectors is determined, and feature data corresponding to the long-term high-resolution remote sensing image data slices is determined, wherein the feature data includes geometric feature data and spectral feature data; using the feature data and the building change categories corresponding to the feature data, a preset random forest model is trained to obtain a target random forest model; after acquiring the change patch vectors to be processed and the corresponding long-term high-resolution remote sensing image data, a set of long-term high-resolution remote sensing image data slices corresponding to each change patch in the sample change patch vectors is determined, and feature data corresponding to each long-term high-resolution remote sensing image slices is determined, wherein the long-term high-resolution remote sensing image data slices corresponding to each change patch in the sample change patch vectors are determined, and feature data corresponding to each long-term high-resolution remote sensing image slices are ... After processing the long-term high-resolution remote sensing image data corresponding to the change patch vector, the change process assessment result corresponding to the change patch vector is determined using the change patch vector to be processed, the long-term high-resolution remote sensing image data corresponding to the change patch vector to be processed, and the target random forest model. The change process assessment result is used to characterize the start and end phases of the change in the building category of the change patch in the change patch vector to be processed, thereby achieving the purpose of determining the occurrence time and subsequent changes of the building category. This solves the technical problem that existing technologies are difficult to assess the progress of building changes, and thus provides technical support for the time-series tracing and tracking of natural resource monitoring and supervision.

[0033] In this embodiment of the invention, based on the sample change patch vector and the long-term high-resolution remote sensing image data corresponding to the sample change patch vector, the long-term high-resolution remote sensing image data slice set corresponding to each change patch in the sample change patch vector is determined, including the following steps:

[0034] Based on the sample change patch vector, the long-term high-resolution remote sensing image data corresponding to the sample change patch vector is preprocessed to obtain the target long-term high-resolution remote sensing image data corresponding to the sample change patch vector. The preprocessing includes: orthorectification processing, resolution unification processing, and coordinate reference system one processing.

[0035] Based on the location and range information of each change patch in the sample change patch vector, the long-term high-resolution remote sensing image data corresponding to the sample change patch vector is sliced ​​to obtain a set of long-term high-resolution remote sensing image data slices corresponding to each change patch in the change patch vector to be evaluated.

[0036] In this embodiment of the invention, long-term high-resolution remote sensing image data {A1,A2,...,A} corresponding to sample change patch vectors were obtained. n After obtaining the sample change patch vector B, the long-term high-resolution remote sensing image data and the sample change patch vector need to be standardized. First, orthorectification is performed on the base period image A1.

[0037] Then, image A2 to A n Orthorectification is performed using A1 as the reference base map to ensure that the positions of the image pixels correspond one-to-one with a small error.

[0038] To further reduce pixel position error, taking base period A1 as a reference, from A2 to A... n Register with A1 respectively.

[0039] If the image types are different, the image resolution needs to be standardized. Using the lowest resolution image as a reference, other temporal images should be resampled to the reference resolution. Check if the coordinate reference systems of the images and vectors are consistent. If they are different, use base period A1 as a reference and resample A2 to A... n The image and vector B are respectively converted into the A1 coordinate reference system.

[0040] The sample change patch vector B contains multiple change patches, with B1 being one of them. The following processing is performed using a single change patch B1 as the processing unit: First, the position and extent of the current change patch B1 are determined, and then the location within the long-term image {A1, A...} is found. 2, ...,A n The image blocks at corresponding positions and ranges in the array are sliced ​​to form a slice set {C1,C2,...,C}. nThe slice set is defined by the bounding rectangle of the variation patch B1.

[0041] To eliminate interference from values ​​outside the patch's range, the slice only has values ​​within the range of the changed patch B1; the portion outside B1 is assigned a null value. Additionally, the change category b corresponding to the changed patch B1 is obtained.

[0042] In this embodiment of the invention, determining the feature data corresponding to the long-term high-resolution remote sensing image data slice set includes the following steps:

[0043] Based on the long-term high-resolution remote sensing image data slice set, the geometric feature data between the slice corresponding to the first time in the long-term high-resolution remote sensing image data slice set and the target slice in the long-term high-resolution remote sensing image data slice set are determined respectively. The target slice includes: slices in the long-term high-resolution remote sensing image data slice set other than the slice corresponding to the first time.

[0044] Based on the long-term high-resolution remote sensing image data slice set, the spectral feature data between the slice corresponding to the first time in the long-term high-resolution remote sensing image data slice set and the target slice in the long-term high-resolution remote sensing image data slice set are determined respectively. The target slice includes slices in the long-term high-resolution remote sensing image data slice set other than the slice corresponding to the first time.

[0045] In this embodiment of the invention, the slice set {C1,C2,...,C} is obtained. i, ...,C n Afterwards, to further analyze the change information, feature calculations are needed. First, obtain the combined time phase {C1, C2, ..., C3} of the base period slice C1 (i.e., the slice corresponding to the first time) and other time phase slices (i.e., the target slice). i Then calculate the combined phase {C1,C}. i Geometric and spectral features F i .

[0046] Geometric features include edge variation features and shape parameter features. The calculation steps for edge variation features are as follows: First, calculate slices C1 and C2 respectively. i Perform principal component analysis to obtain the corresponding first principal component C. 1_1p c and C i_1pc Then, for C 1_1pc With C i_1pc To perform Canny edge detection, the Canny model can be called using the feature module of the public library Skimage to obtain the edge intensity map C. 1_edge With C i_edge Then, the ratio features r1 and r2 are calculated based on the edge intensity map:

[0047]

[0048]

[0049] Where, ∑C 1_edge With ∑C i_edge This represents the summation of edge intensity maps, ∑ ＜0 (C i_edge -C 1_edge ) and ∑ ＞0 (C i_edge -C 1_edge ) respectively represent the values ​​for C i_edge -C 1_edge Summing of values ​​less than 0 and summing of values ​​greater than 0.

[0050] The steps for calculating shape parameter features are: locating C1 and C... i For the corresponding feature B1, calculate the outer rectangle of B1; then, based on the outer rectangle and feature B1, calculate the aspect ratio r3 and area ratio r4.

[0051]

[0052]

[0053] Where l and w correspond to the length and width of the circumscribed rectangle, respectively, and a e Let 'a' and 'a' represent the areas of the circumscribed rectangle and patch B1, respectively. Regarding these four geometric features, 'r1' reflects the change in edge complexity between preceding and following time phases, 'r2' reflects the degree of change in the later time phase relative to the earlier time phase, and 'r3' and 'r4' reflect the shape of the patch, which is beneficial for characterizing structures with prominent shape features such as strip-shaped and block-shaped structures. Using ratios eliminates the influence of dimensions, focusing only on relative features, and to some extent improves the generalization ability of the engineering project.

[0054] The calculation steps for spectral characteristics are as follows: First, compare the previous time slice C1 with the subsequent time slice C... i HSV conversion of the red, green, and blue true-color bands can be performed by calling the rgb2hsv model from the color module of the public library Skimage to obtain the luminance component map C. 1_v With C i_v Then, based on C 1_v With C i_v Calculate the luminance ratio r5, the standard deviation ratio of the preceding phase r6, the standard deviation ratio of the following phase r7, and the standard deviation ratio r8:

[0055]

[0056]

[0057]

[0058]

[0059] Where, ∑C i_v With ∑C 1_v This represents the summation of the luminance component maps, std(C 1_v ) and std(C i_v The expression represents the standard deviation of the luminance component map, where mean(C) represents the standard deviation of the luminance component map. 1_v ) and mean(C i_v The expression () represents the mean value of the brightness component map. Among the four spectral features, r5 reflects the change in brightness, r6 and r7 reflect the brightness distribution of structures within the patch at different time points, respectively, and r8 reflects the change in brightness distribution at different time points.

[0060] The final characteristic data combination F of the change process was obtained. i ={r1,r2,r3,r4,r5,r6,r7,r8}, since all features are dimensionless ratio features, the influence of dimensional absolute values ​​is weakened. Therefore, it does not depend on specific image values ​​and has a good effect on improving the ability of engineering generalization. The combination of geometric features and spectral features provides a more comprehensive characterization of building information and has a good effect on distinguishing different building variations.

[0061] The following is a detailed explanation of step S106.

[0062] To predict the probability of changes between different time phases, a machine learning prediction model needs to be trained. After obtaining the feature data and the corresponding building change categories, a random forest model is constructed using the obtained feature data F. s For input, the category of change L s Train the model for the labels.

[0063] The RandomForestClassifier model from the ensemble module of the open-source library Sklearn can be used for model training; then the temporal combinations to be detected {C1, C... i Using the feature Fi as input, a trained random forest model is used for probability prediction to obtain the probability Pi corresponding to the change category b; finally, all temporal combinations of the current patch B1 are traversed to obtain the long-term probability sequence P = {0, P2, ..., P}. n}

[0064] In this embodiment of the invention, step S108 includes the following steps:

[0065] Based on the vector of the to-be-processed changed patches and the long-temporal high-resolution remote sensing image data corresponding to the vector of the to-be-processed changed patches, determine the set of slices of the long-temporal high-resolution remote sensing image data corresponding to each changed patch in the vector of the to-be-evaluated changed patches, and determine the set of slices of the long-temporal high-resolution remote sensing image data corresponding to each changed patch in the vector of the to-be-evaluated changed patches;

[0066] Input the set of slices of the long-temporal high-resolution remote sensing image data corresponding to each changed patch in the vector of the to-be-evaluated changed patches into the target random forest model, and obtain the change process evaluation result corresponding to the vector of the to-be-processed changed patches.

[0067] In the embodiment of the present invention, obtain the vector B of the to-be-processed changed patches y The long-temporal probability sequence P = {0, P2,..., P n}, then next, obtain the change process information. First, calculate the change gradient T i :

[0068] T i = P i - P i-1

[0069] Obtain the gradient sequence T = {P2, P3 - P2,..., P n - P n-1}; then, find the phase S corresponding to the first gradient that satisfies T > t1 in T, where t1 is a threshold, and S is the starting phase when the change occurs; for all phases after S, satisfying |T| > t2 indicates that the change still exists, and satisfying |T| < t2 indicates that the change has been lifted, where t2 is a threshold. According to the change status of the building structures at the phase S when the change occurs and the subsequent phases, the evolution process of the change of the building structures over time can be listed, which can provide a judgment basis for law enforcement traceability and law enforcement tracking.

[0070] Embodiment 2:

[0071] The embodiment of the present invention further provides a building structure change discovery and long-temporal tracking monitoring device, which is used to execute the building structure change discovery and long-temporal tracking monitoring method provided in the above content of the embodiment of the present invention. The following is a specific introduction to the building structure change discovery and long-temporal tracking monitoring device provided in the embodiment of the present invention.

[0072] As Figure 2 shown, Figure 2 is a schematic diagram of the above building structure change discovery and long-temporal tracking monitoring device. The building structure change discovery and long-temporal tracking monitoring device includes:

[0073] Acquisition unit 10 is used to acquire sample change patch vectors and long-term high-resolution remote sensing image data corresponding to the sample change patch vectors;

[0074] The determining unit 20 is used to determine, based on the sample change patch vector and the long-term high-resolution remote sensing image data corresponding to the sample change patch vector, the long-term high-resolution remote sensing image data slice set corresponding to each change patch in the sample change patch vector, and to determine the feature data corresponding to the long-term high-resolution remote sensing image data slice set, wherein the feature data includes: geometric feature data and spectral feature data;

[0075] Training unit 30 is used to train a preset random forest model using the feature data and the building change categories corresponding to the feature data to obtain a target random forest model;

[0076] The evaluation unit 40 is used to determine the change process evaluation result corresponding to the change patch vector to be processed after acquiring the change patch vector to be processed and the long-term high-resolution remote sensing image data corresponding to the change patch vector to be processed, using the change patch vector to be processed, the long-term high-resolution remote sensing image data corresponding to the change patch vector to be processed, and the target random forest model. The change process evaluation result is used to characterize the start and end phases of the change in the building category of the change patch in the change patch vector to be processed.

[0077] In this embodiment of the invention, sample change patch vectors and corresponding long-term high-resolution remote sensing image data are acquired; based on the sample change patch vectors and corresponding long-term high-resolution remote sensing image data, a set of long-term high-resolution remote sensing image data slices corresponding to each change patch in the sample change patch vectors is determined, and feature data corresponding to the set of long-term high-resolution remote sensing image data slices is determined, wherein the feature data includes geometric feature data and spectral feature data; using the feature data and the building change categories corresponding to the feature data, a preset random forest model is trained to obtain a target random forest model; after acquiring the change patch vectors to be processed and the corresponding long-term high-resolution remote sensing image data, a set of long-term high-resolution remote sensing image data slices corresponding to each change patch in the sample change patch vectors is determined, and feature data corresponding to the long-term high-resolution remote sensing image data slices is determined, wherein the feature data includes geometric feature data and spectral feature data; using the feature data and the building change categories corresponding to the feature data, a preset random forest model is trained to obtain a target random forest model; after acquiring the change patch vectors to be processed and the corresponding long-term high-resolution remote sensing image data, a set of long-term high-resolution remote sensing image data slices corresponding to each change patch in the sample change patch vectors is determined, and feature data corresponding to each long-term high-resolution remote sensing image slices is determined, wherein the long-term high-resolution remote sensing image data slices corresponding to each change patch in the sample change patch vectors are determined, and feature data corresponding to each long-term high-resolution remote sensing image slices are ... After processing the long-term high-resolution remote sensing image data corresponding to the change patch vector, the change process assessment result corresponding to the change patch vector is determined using the change patch vector to be processed, the long-term high-resolution remote sensing image data corresponding to the change patch vector to be processed, and the target random forest model. The change process assessment result is used to characterize the start and end phases of the change in the building category of the change patch in the change patch vector to be processed, thereby achieving the purpose of determining the occurrence time and subsequent changes of the building category. This solves the technical problem that existing technologies are difficult to assess the progress of building changes, and thus provides technical support for the time-series tracing and tracking of natural resource monitoring and supervision.

[0078] Example 3:

[0079] This invention also provides an electronic device, including a memory and a processor. The memory is used to store a program that supports the processor in executing the method described in Embodiment 1 above, and the processor is configured to execute the program stored in the memory.

[0080] See Figure 3 The present invention also provides an electronic device 100, including: a processor 50, a memory 51, a bus 52 and a communication interface 53, wherein the processor 50, the communication interface 53 and the memory 51 are connected through the bus 52; the processor 50 is used to execute executable modules, such as computer programs, stored in the memory 51.

[0081] The memory 51 may include high-speed random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 53 (which can be wired or wireless), such as the Internet, wide area network, local area network, metropolitan area network, etc.

[0082] Bus 52 can be an ISA bus, PCI bus, or EISA bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 3 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0083] The memory 51 is used to store programs. After receiving an execution instruction, the processor 50 executes the programs. The method executed by the device for defining the flow process disclosed in any of the foregoing embodiments of the present invention can be applied to the processor 50 or implemented by the processor 50.

[0084] Processor 50 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of processor 50 or by instructions in software form. Processor 50 can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this invention can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory 51. The processor 50 reads the information in memory 51 and, in conjunction with its hardware, completes the steps of the above method.

[0085] Example 4:

[0086] This invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the method described in Embodiment 1 above.

[0087] Furthermore, in the description of the embodiments of the present invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in the present invention based on the specific circumstances.

[0088] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0089] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the shown or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0090] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0091] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0092] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for detecting and monitoring changes in buildings and structures over a long period of time, characterized in that, include: Acquire the sample change patch vector and the long-term high-resolution remote sensing image data corresponding to the sample change patch vector; Based on the sample change patch vector and the long-term high-resolution remote sensing image data corresponding to the sample change patch vector, the long-term high-resolution remote sensing image data slice set corresponding to each change patch in the sample change patch vector is determined, and the feature data corresponding to the long-term high-resolution remote sensing image data slice set is determined, wherein the feature data includes: geometric feature data and spectral feature data; Using the feature data and the building change categories corresponding to the feature data, a preset random forest model is trained to obtain a target random forest model; After obtaining the vector of the change patch to be processed and the long-term high-resolution remote sensing image data corresponding to the vector of the change patch to be processed, the change process assessment result corresponding to the vector of the change patch to be processed is determined by using the vector of the change patch to be processed, the long-term high-resolution remote sensing image data corresponding to the vector of the change patch to be processed and the target random forest model. The change process assessment result is used to characterize the start and end phases of the building changes of the change patch in the vector of the change patch to be processed.

2. The method according to claim 1, characterized in that, Based on the sample change patch vector and the corresponding long-term high-resolution remote sensing image data, a set of long-term high-resolution remote sensing image data slices corresponding to each change patch in the sample change patch vector is determined, including: Based on the sample change patch vector, the long-term high-resolution remote sensing image data corresponding to the sample change patch vector is preprocessed to obtain the target long-term high-resolution remote sensing image data corresponding to the sample change patch vector. The preprocessing includes: orthorectification processing, resolution unification processing, and coordinate reference system one processing. Based on the location and range information of each change patch in the sample change patch vector, the long-term high-resolution remote sensing image data corresponding to the sample change patch vector is sliced ​​to obtain a long-term high-resolution remote sensing image data slice set corresponding to each change patch in the sample change patch vector.

3. The method according to claim 1, characterized in that, The feature data corresponding to the long-term high-resolution remote sensing image data slice set were determined, including: Based on the long-term high-resolution remote sensing image data slice set, the geometric feature data between the slice corresponding to the first time in the long-term high-resolution remote sensing image data slice set and the target slice in the long-term high-resolution remote sensing image data slice set are determined respectively. The target slice includes: slices in the long-term high-resolution remote sensing image data slice set other than the slice corresponding to the first time. Based on the long-term high-resolution remote sensing image data slice set, the spectral feature data between the slice corresponding to the first time in the long-term high-resolution remote sensing image data slice set and the target slice in the long-term high-resolution remote sensing image data slice set are determined respectively. The target slice includes slices in the long-term high-resolution remote sensing image data slice set other than the slice corresponding to the first time.

4. The method according to claim 3, characterized in that, Based on the long-term high-resolution remote sensing image data slice set, the geometric feature data between the slice corresponding to the first time in the long-term high-resolution remote sensing image data slice set and the target slice in the long-term high-resolution remote sensing image data slice set are determined, including: Principal component analysis was performed on the slice corresponding to the first time and the target slice respectively to obtain the first principal component of the slice corresponding to the first time and the first principal component of the target slice. Using the Canny model, the edge intensity maps corresponding to the first principal components of the slice at the first time and the edge intensity maps corresponding to the first principal components of the target slice are determined respectively. Calculate the ratio feature data between the edge intensity map corresponding to the first principal component of the slice at the first time and the edge intensity map corresponding to the first principal component of the target slice; The slice corresponding to the first time and the change patch corresponding to the target slice are determined, and the aspect ratio of the bounding rectangle corresponding to the change patch and the area ratio between the change patch and the bounding rectangle are determined. The ratio feature data, the aspect ratio, and the area ratio are determined as the geometric feature data.

5. The method according to claim 3, characterized in that, Based on the long-term high-resolution remote sensing image data slice set, the spectral feature data between the slice corresponding to the first time in the long-term high-resolution remote sensing image data slice set and the target slice in the long-term high-resolution remote sensing image data slice set are determined, including: HSV conversion is performed on the true color bands of the slice corresponding to the first time and the true color bands of the target slice respectively to obtain a first conversion result and a second conversion result, wherein the true color bands include: red band, green band and blue band; Based on the first conversion result, the second conversion result, and the rgb2hsv model, the first luminance component map of the slice corresponding to the first time and the second luminance component map of the target slice are determined. Based on the first luminance component map and the second luminance component map, spectral feature data is determined, wherein the spectral feature data includes: the luminance ratio between the first luminance component map and the second luminance component map, the standard deviation ratio of the first luminance component map, the standard deviation ratio of the second luminance component map, and the standard deviation ratio between the first luminance component map and the second luminance component map.

6. The method according to claim 1, characterized in that, Using the unprocessed change patch vector, the corresponding long-term high-resolution remote sensing image data, and the target random forest model, the change process assessment result corresponding to the unprocessed change patch vector is determined, including: Based on the vector of the changed patch to be processed and the long-term high-resolution remote sensing image data corresponding to the vector of the changed patch to be processed, the long-term high-resolution remote sensing image data slice set corresponding to each changed patch in the vector of the sample changed patch is determined, and the long-term high-resolution remote sensing image data slice set corresponding to each changed patch in the vector of the changed patch to be processed is determined. The long-term high-resolution remote sensing image data slice set corresponding to each changed patch in the vector of changed patches to be processed is input into the target random forest model to obtain the evaluation result of the change process corresponding to the vector of changed patches to be processed.

7. A device for detecting and monitoring changes in buildings and structures over a long period of time, characterized in that, include: The acquisition unit is used to acquire the sample change patch vector and the long-term high-resolution remote sensing image data corresponding to the sample change patch vector; The determining unit is configured to, based on the sample change patch vector and the long-term high-resolution remote sensing image data corresponding to the sample change patch vector, determine the long-term high-resolution remote sensing image data slice set corresponding to each change patch in the sample change patch vector, and determine the feature data corresponding to the long-term high-resolution remote sensing image data slice set, wherein the feature data includes: geometric feature data and spectral feature data; The training unit is used to train a preset random forest model using the feature data and the building change categories corresponding to the feature data, so as to obtain a target random forest model. The evaluation unit, after acquiring the vector of the changed patch to be processed and the long-term high-resolution remote sensing image data corresponding to the vector of the changed patch to be processed, uses the vector of the changed patch to be processed, the long-term high-resolution remote sensing image data corresponding to the vector of the changed patch to be processed, and the target random forest model to determine the evaluation result of the change process corresponding to the vector of the changed patch to be processed. The evaluation result of the change process is used to characterize the start and end phases of the change in the building category of the changed patch in the vector of the changed patch to be processed.

8. The apparatus according to claim 7, characterized in that, The determining unit is used for: Based on the sample change patch vector, the long-term high-resolution remote sensing image data corresponding to the sample change patch vector is preprocessed to obtain the target long-term high-resolution remote sensing image data corresponding to the sample change patch vector. The preprocessing includes: orthorectification processing, resolution unification processing, and coordinate reference system one processing. Based on the location and range information of each change patch in the sample change patch vector, the long-term high-resolution remote sensing image data corresponding to the sample change patch vector is sliced ​​to obtain a long-term high-resolution remote sensing image data slice set corresponding to each change patch in the sample change patch vector.

9. An electronic device, characterized in that, The device includes a memory and a processor, the memory being used to store a program that enables the processor to execute the method according to any one of claims 1 to 6, and the processor being configured to execute the program stored in the memory.

10. A computer-readable storage medium storing a computer program thereon, characterized in that, When a computer program is run by a processor, it performs the steps of the method described in any one of claims 1 to 6.

Images