CN114694030A

CN114694030A - Landslide detection method, device, equipment and storage medium

Info

Publication number: CN114694030A
Application number: CN202210423650.1A
Authority: CN
Inventors: 荀张媛; 刘恩泽; 张帆; 李成龙
Original assignee: Meihang Remote Sensing Information Co ltd; Aerial Photogrammetry and Remote Sensing Co Ltd
Current assignee: Meihang Remote Sensing Information Co ltd; Aerial Photogrammetry and Remote Sensing Co Ltd
Priority date: 2022-04-21
Filing date: 2022-04-21
Publication date: 2022-07-01

Abstract

The embodiment of the invention provides a landslide detection method, a landslide detection device, landslide detection equipment and a storage medium, which relate to the technical field of geological disaster prediction and comprise the following steps: acquiring pre-earthquake and post-earthquake remote sensing images of a region to be detected from an optical remote sensing image system, registering, and carrying out change detection on the post-earthquake remote sensing image according to the registered pre-earthquake remote sensing image; extracting a predicted deformation area from the post-earthquake non-change area according to the image of the post-earthquake non-change area and a preset deformation rate diagram of the area to be detected, and combining the predicted deformation area with the post-earthquake change area to obtain a target area; calculating a characteristic parameter of each object in the target area according to a gray texture characteristic diagram obtained by calculation of a preset radar image of the area to be detected, the remote sensing image after the earthquake, a preset digital elevation model and a preset deformation rate diagram, and eliminating the area where the non-landslide ground object in the target area is located based on the parameter to obtain the landslide hidden danger area. By adopting the method and the device, the accuracy and the comprehensiveness of landslide hidden danger detection can be improved.

Description

Landslide detection method, device, equipment and storage medium

Technical Field

The invention relates to the technical field of geological disaster prediction, in particular to a landslide detection method, a landslide detection device, landslide detection equipment and a storage medium.

Background

The most common geological disasters of landslide bring huge damage to infrastructures such as roads and houses, seriously threaten the life and property safety of people and also cause obvious surface deformation.

Therefore, it is essential to accurately detect landslides in time in geological disaster prediction.

However, the conventionally used change detection method can quickly identify the region of the surface change, but can only identify the region where the landslide has occurred, and does not identify the secondary landslide that may occur; although the InSAR technology can identify the secondary landslide, the incoherent phenomenon often occurs, data holes occur due to data loss in partial areas, manual intervention is needed, and the accuracy is not high.

Disclosure of Invention

The embodiment of the invention provides a landslide detection method, a landslide detection device and a storage medium, which can improve the accuracy and efficiency of landslide hidden danger detection and ensure the comprehensiveness of geological analysis.

In a first aspect, an embodiment of the present invention provides a landslide detection method, including:

acquiring a pre-earthquake remote sensing image and a post-earthquake remote sensing image of a region to be detected from an optical remote sensing image system;

registering the pre-earthquake remote sensing image to the post-earthquake remote sensing image by taking the post-earthquake remote sensing image as a reference to obtain a registered pre-earthquake remote sensing image;

according to the registered pre-earthquake remote sensing image, change detection is carried out on the post-earthquake remote sensing image to obtain an image of a post-earthquake changed area and an image of a post-earthquake unchanged area;

extracting a prediction deformation region from the post-earthquake non-change region according to the image of the post-earthquake non-change region and a preset deformation rate diagram of the region to be detected; the coordinate system of the preset deformation rate graph is consistent with the coordinate system of the optical remote sensing image system;

calculating a gray texture feature map of a preset radar image according to a gray co-occurrence matrix of the preset radar image of the area to be detected; the coordinate system of the preset radar image is consistent with the coordinate system of the optical remote sensing image system;

merging the post-earthquake change area and the predicted deformation area to obtain a target area;

calculating characteristic parameters of each object in the target area according to the gray texture characteristic diagram, the remote sensing image after the earthquake, a preset digital elevation model of the area to be detected and the preset deformation rate diagram;

and eliminating the region where the non-landslide ground object in the target region is located according to the characteristic parameters of each object in the target region to obtain a landslide hidden danger region.

Optionally, the registering the pre-earthquake remote sensing image to the post-earthquake remote sensing image by using the post-earthquake remote sensing image as a reference to obtain a registered pre-earthquake remote sensing image includes:

performing orthorectification on the pre-earthquake remote sensing image and the post-earthquake remote sensing image to obtain a pre-earthquake orthorectification remote sensing image and a post-earthquake orthorectification remote sensing image;

respectively fusing panchromatic band data and multispectral data in the pre-earthquake ortho-remote sensing image and the post-earthquake ortho-remote sensing image to obtain a fused pre-earthquake remote sensing image and a fused post-earthquake remote sensing image;

and registering the fused remote sensing image before the earthquake to the fused remote sensing image after the earthquake as a reference to obtain the registered remote sensing image before the earthquake.

Optionally, the performing change detection on the post-earthquake remote sensing image according to the registered pre-earthquake remote sensing image to obtain an image of a post-earthquake changed region and an image of a post-earthquake unchanged region includes:

respectively carrying out principal component transformation on the registered pre-earthquake remote sensing image and the post-earthquake remote sensing image to obtain a first principal component image of the registered pre-earthquake remote sensing image and a first principal component image of the post-earthquake remote sensing image; the feature information in the first principal component image of the registered pre-earthquake remote sensing image corresponds to the feature information of the registered pre-earthquake remote sensing image one by one, and the feature information in the first principal component image of the post-earthquake remote sensing image corresponds to the feature information of the post-earthquake remote sensing image one by one;

determining an area in which the position offset between the pixel in the first principal component image of the post-earthquake remote sensing image and the corresponding pixel in the first principal component image of the registered pre-earthquake remote sensing image exceeds a preset offset as an after-earthquake change area in the first principal component image of the post-earthquake remote sensing image;

determining an image of the post-earthquake change region from the post-earthquake remote sensing image according to the post-earthquake change region;

and determining the images of other areas except the post-earthquake changed area in the post-earthquake remote sensing image as the images of the post-earthquake unchanged area.

Optionally, the extracting a predicted deformation region from the post-earthquake non-change region according to the image of the post-earthquake non-change region and the preset deformation information map of the region to be detected includes:

calculating the deformation rate value of each pixel in the image of the post-earthquake non-change area according to the preset deformation rate graph;

and determining an area formed by pixels of which the deformation rate values are within a preset deformation threshold range in the image of the post-earthquake non-change area as the predicted deformation area.

Optionally, the characteristic parameters of each object include: gray texture characteristic values, multispectral values, terrain characteristic values and deformation rate values; calculating the characteristic parameters of each object in the target area according to the gray texture characteristic diagram, the post-earthquake remote sensing image, the preset digital elevation model of the area to be detected and the preset deformation rate diagram, wherein the calculation comprises the following steps:

vectorizing the boundary of the target area to obtain a vector boundary diagram of the target area;

extracting a remote sensing image map of the target area, a digital elevation model of the target area, a deformation rate map of the target area and a gray texture feature map of the target area from the post-earthquake remote sensing image, the preset digital elevation model, the preset deformation rate map and the gray texture feature map respectively by using the vector boundary map;

calculating to obtain a terrain characteristic value of each object in the target area according to the digital elevation model of the target area;

and calculating the gray texture characteristic value, the multispectral value and the deformation rate value of each object in the target area according to the gray texture characteristic image of the target area, the remote sensing image of the target area and the deformation rate image of the target area.

Optionally, the removing, according to the characteristic parameter of each object in the target region, a region where a non-landslide terrain is located in the target region to obtain a landslide hidden danger region includes:

selecting a plurality of ground object samples in the remote sensing image of the target area by taking an object as a unit; the plurality of surface feature samples comprise all surface feature types in the remote sensing image of the target area, wherein one surface feature sample comprises only one surface feature type;

obtaining the value range of the characteristic parameters of each surface feature type object by using a preset decision tree algorithm according to the characteristic parameters of all the objects in the plurality of surface feature samples;

obtaining the corresponding relation between the characteristic parameter value range of each object in the target area and each surface feature type according to the value range of the characteristic parameter of each surface feature type object, and generating a preset surface feature classification rule of the target area;

classifying the remote sensing image map of the target area by adopting a preset ground object classification rule of the target area to obtain a ground object classification result in the target area;

and according to the land feature classification result, eliminating the region where the non-landslide land feature is located in the target region to obtain the landslide hidden danger region.

Optionally, the terrain feature value comprises: the method comprises the following steps of calculating a slope value, a ground elevation value, a mountain shadow value and a terrain relief value according to the digital elevation model of the target area to obtain a terrain characteristic value of each object in the target area, wherein the steps comprise:

according to the digital elevation model of the target area, calculating to obtain a gradient map, a mountain shadow map, a surface relief map and a ground elevation map of the target area;

and respectively calculating the slope value, the mountain shadow value, the terrain relief value and the ground elevation value of each object in the target area according to the slope map, the mountain shadow map, the terrain relief map and the ground elevation map.

In a second aspect, an embodiment of the present invention further provides a landslide detection apparatus, where the apparatus includes:

the acquisition module is used for acquiring the pre-earthquake remote sensing image and the post-earthquake remote sensing image of the area to be detected from the optical remote sensing image system;

the registration module is used for registering the pre-earthquake remote sensing image to the post-earthquake remote sensing image by taking the post-earthquake remote sensing image as a reference to obtain a registered pre-earthquake remote sensing image;

the detection module is used for carrying out change detection on the remote sensing image after the earthquake according to the registered remote sensing image before the earthquake to obtain an image of a changed area after the earthquake and an image of a non-changed area after the earthquake;

the extraction module is used for extracting a prediction deformation region from the post-earthquake non-change region according to the image of the post-earthquake non-change region and the preset deformation rate diagram of the region to be detected; the coordinate system of the preset deformation rate graph is consistent with the coordinate system of the optical remote sensing image system;

the first calculation module is used for calculating a gray texture feature map of a preset radar image according to a gray co-occurrence matrix of the preset radar image of the area to be detected; the coordinate system of the preset radar image is consistent with the coordinate system of the optical remote sensing image system;

the merging module is used for merging the post-earthquake change area and the prediction deformation area to obtain a target area;

the second calculation module is further used for calculating characteristic parameters of each object in the target area according to the gray texture characteristic map, the remote sensing image after the earthquake, a preset digital elevation model of the area to be detected and the preset deformation rate map;

and the removing module is used for removing the area where the non-landslide ground object in the target area is located according to the characteristic parameters of each object in the target area to obtain a landslide hidden danger area.

In a third aspect, an embodiment of the present invention further provides a landslide detection apparatus, including: a processor, a memory and a bus, wherein the memory stores program instructions executable by the processor, the processor and the memory communicate via the bus when the landslide detection apparatus is operating, and the processor executes the program instructions to perform the steps of the landslide detection method according to any one of the first aspects.

In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, where the storage medium stores a computer program, and the computer program is executed by a processor to perform the landslide detection method according to any one of the first aspects.

The invention provides a landslide detection method, a landslide detection device and a storage medium, which are characterized in that a pre-earthquake remote sensing image and a post-earthquake remote sensing image of a region to be detected, which are acquired from an optical remote sensing image system, are registered, the post-earthquake remote sensing image is subjected to transformation detection according to the registered pre-earthquake remote sensing image to obtain a post-earthquake change region and a post-earthquake non-change region, a prediction deformation region is obtained according to the post-earthquake non-change region and a preset deformation rate diagram, a gray texture characteristic diagram of a preset radar image is obtained according to gray level co-occurrence matrix calculation of a preset radar image of the region to be detected, the post-earthquake change region and the prediction deformation region are combined to obtain a target region, characteristic parameters of each object in the target region are calculated by utilizing a gray level characteristic diagram, the post-earthquake remote sensing image, a preset digital elevation model and a preset deformation rate diagram, and finally the region where the non-landslide ground object is located is removed from the target region according to the characteristic parameters of each object in the target region, and obtaining a landslide hidden danger area. By the method, change detection is carried out on the registered remote sensing image, and a predicted deformation area is obtained by utilizing a deformation rate diagram, so that the finally obtained target area has not only the change area obtained by the change detection, but also the area subjected to deformation, and the information of the target area subjected to landslide analysis detection is more complete and comprehensive; meanwhile, the characteristic parameters of each object in the target area, which are obtained by utilizing the gray texture characteristic map, the remote sensing image after the earthquake, the preset digital elevation model and the preset deformation rate map, are used for eliminating the area of non-landslide land objects in the target area, so that the finally obtained landslide hidden danger area is more accurate, powerful support can be provided for landslide detection, the landslide detection efficiency is improved, geological detection and analysis can be more efficiently and reliably carried out in the face of landslide disasters, and the landslide hidden danger area can be protected in time.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained according to the drawings without inventive efforts.

Fig. 1 is a schematic flow chart of a landslide detection method according to the present invention;

FIG. 2 is a schematic flow chart of a remote sensing image registration method according to the present invention;

FIG. 3 is a schematic flow chart of a variation detection method according to the present invention;

FIG. 4 is a schematic diagram of a process for extracting a predicted deformation region according to the present invention;

FIG. 5 is a schematic flow chart of calculating object feature parameters according to the present invention;

FIG. 6 is a schematic flow chart of obtaining a landslide hazard area according to the present invention;

FIG. 7 is a schematic diagram illustrating a process for calculating a terrain feature value according to the present invention;

FIG. 8 is a schematic view of a landslide detection apparatus according to the present invention;

fig. 9 is a schematic diagram of a landslide detection apparatus provided by the present invention.

Icon: 1000, obtaining a module; 2000, a registration module; 3000, a detection module; 4000, an extraction module; 5000, a first calculation module; 6000, merging the modules; 7000, a second calculation module; 8000, a removing module; 10, a landslide detection device; 11, a processor; 12, a memory; 13, a bus.

Detailed Description

In order to make the objects, 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 and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. The components of embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.

Thus, the following detailed description of the embodiments of the present invention, presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined or explained in subsequent figures.

Before explaining the present invention in detail, an application scenario of the present invention will be described.

Landslide mostly occurs in mountainous areas and is often accompanied by other secondary disasters. However, due to rugged terrain and inconvenient traffic in mountainous areas, field geological analysis is difficult to be performed by manpower in a short time, and conventional landslide detection methods have certain limitations, for example, a landslide method based on change detection can quickly acquire a change area, but the situation of over-identification generally exists, and secondary landslide which may occur is not identified; InSAR (Interferometric Synthetic Aperture Radar) can identify the hidden danger of secondary landslide after earthquake, but the problem of coherence loss is one of the biggest obstacles; although the accuracy of land feature classification can be improved based on an object-oriented method, landslide identification is carried out in a large range with low efficiency, and is insensitive to landslide hidden dangers without disasters, and the characteristics of optical images such as spectrum and texture cannot meet the requirements of quantitative and rapid landslide information extraction at present. These limitations in landslide detection result in failure to provide a reliable landslide detection method to ensure the accuracy of the result.

Based on the above, the invention provides a landslide detection method, a landslide detection device and a storage medium, wherein a pre-earthquake remote sensing image and a post-earthquake remote sensing image of a to-be-detected area acquired from an optical remote sensing image system are registered, the post-earthquake remote sensing image is transformed and detected according to the registered pre-earthquake remote sensing image to obtain a post-earthquake variable area and a post-earthquake non-variable area, a predicted deformation area is obtained according to the post-earthquake non-variable area and a preset deformation rate diagram, a gray texture characteristic diagram of a preset radar image is obtained according to the gray co-occurrence matrix calculation of the preset radar image of the to-be-detected area, the post-earthquake variable area and the predicted deformation area are combined to obtain a target area, then a gray texture characteristic diagram, the post-earthquake remote sensing image, a preset digital elevation model and a preset deformation rate diagram are used for calculating characteristic parameters of each object in the target area, and finally the area where the non-landslide ground object is located is removed in the target area according to the characteristic parameters of each object in the target area, and obtaining a landslide hidden danger area. The landslide detection method provided in the following embodiment of the present invention may be executed by a landslide detection device, which may be a desktop computer, a notebook computer, or the like, or an intelligent terminal that is convenient to carry, and the present invention is not limited thereto.

The following is an explanation by way of various embodiments in conjunction with the accompanying drawings. Fig. 1 is a schematic flow chart of a landslide detection method provided by the present invention. As shown in fig. 1, the landslide detection method includes:

s110, acquiring a pre-earthquake remote sensing image and a post-earthquake remote sensing image of the area to be detected from the optical remote sensing image system.

The remote sensing technology has the advantages of high imaging speed, wide coverage range, low cost and the like, and is an effective means for monitoring disasters and environments. The optical remote sensing image system is generally composed of a remote sensor, a remote sensing platform and an information transmission device. The remote sensor is the basis for forming an optical remote sensing image system and can be a synthetic aperture radar or a multispectral scanner and the like. In this embodiment, unmanned aerial vehicle remote sensing technology can be used, and aerial photography is carried out in the area to be detected, so that a panoramic optical remote sensing image of the area to be detected is formed.

Through the information transmission device, the ground can acquire the pre-earthquake remote sensing image and the post-earthquake remote sensing image of the area to be detected, which are shot by the remote sensor, from the optical remote sensing image system. Wherein, the remote sensing image before the earthquake and the remote sensing image after the earthquake are both optical remote sensing images.

And S120, registering the pre-earthquake remote sensing image to the post-earthquake remote sensing image by taking the post-earthquake remote sensing image as a reference to obtain the registered pre-earthquake remote sensing image.

Because the acquisition time of the post-earthquake remote sensing image is inconsistent with that of the pre-earthquake remote sensing image, in order to eliminate distortion and prevent the increase of the error rate of subsequent detection due to image dislocation, two remote sensing images need to be preprocessed. In this embodiment, the remote sensing image after earthquake is taken as a reference, and the remote sensing image before earthquake is registered to the remote sensing image after earthquake, so that the coordinates of the remote sensing image before earthquake are consistent with those of the remote sensing image after earthquake, and the registered remote sensing image before earthquake is obtained. Optionally, in the registration process, feature points in the two remote sensing images to be registered may be determined first, and spatial relationship conversion is performed according to the feature points, so that the pre-earthquake remote sensing image may be registered to the coordinate system of the post-earthquake remote sensing image according to the conversion matrix, and the registered pre-earthquake remote sensing image is obtained. In a possible implementation manner, a registration algorithm compiled by an open source program platform may be used to perform automatic registration of the two images, which is not limited by the present invention.

S130, according to the registered pre-earthquake remote sensing image, change detection is carried out on the post-earthquake remote sensing image, and an image of a post-earthquake changed area and an image of a post-earthquake unchanged area are obtained.

In order to obtain the difference between the states observed by different events in the area to be detected, the remote sensing image after the earthquake needs to be transformed and detected according to the registered remote sensing image before the earthquake, so that the image of the changed area after the earthquake and the image of the unchanged area after the earthquake are obtained. In general, the appropriate threshold values may be selected to identify changed and unchanged images based on the selected change detection method, such as algebraic algorithms, transformations, classification, etc. The image of the change area after the earthquake is the image of the image area with obvious surface change compared with the registered remote sensing image before the earthquake, the area is the area suffered from the primary disaster, and the area suffered from the disaster can be locked in time through the change detection result.

And S140, extracting a prediction deformation area from the post-earthquake non-change area according to the image of the post-earthquake non-change area and the preset deformation rate diagram of the area to be detected.

And the coordinate system of the preset deformation rate graph is consistent with the coordinate system of the optical remote sensing image system.

In order to improve the accuracy of landslide detection, in addition to paying attention to the image of the post-earthquake changed area where significant earth surface changes occur in the change detection, attention should be paid to the post-earthquake unchanged area where secondary disasters may occur. In this embodiment, a predicted deformation region is extracted from the post-earthquake non-change region according to an image of the post-earthquake non-change region and a preset deformation rate diagram of the region to be detected, where the predicted deformation region is a region predicted to be in the post-earthquake non-change region and likely to have a secondary disaster.

Specifically, an SAR (Synthetic Aperture Radar) image of a region to be detected is processed by technologies such as Stacking (interference pattern superposition technology), PS (Persistent Scatterer) and SBAS (Small base line Subsets) to obtain an unprocessed preset deformation rate map of the region to be detected; then, in order to facilitate subsequent processing, projection processing needs to be performed on the unprocessed preset deformation rate diagram of the region to be detected, so that the coordinate system of the processed preset deformation rate diagram is consistent with the coordinate system of the optical remote sensing image system. In a possible implementation manner, the resampling processing may be performed on the preset deformation rate map, so that the resolution of the preset deformation rate map is consistent with the image resolution in the optical remote sensing image system.

S150, calculating a gray texture feature map of the preset radar image according to the gray co-occurrence matrix of the preset radar image of the area to be detected.

And the coordinate system of the preset radar image is consistent with the coordinate system of the optical remote sensing image system.

In this embodiment, an unprocessed post-earthquake SAR image, that is, an unprocessed post-earthquake radar image, may be obtained from the SAR image of the region to be detected. The unprocessed post-earthquake radar image is corrected and filtered through an image processing platform, such as an ENVI platform, so that the processed post-earthquake radar image is obtained. And then, taking the post-earthquake remote sensing image as a reference, and registering the post-earthquake radar image to the post-earthquake remote sensing image to obtain the registered post-earthquake radar image. And then, projecting and resampling the registered post-earthquake radar image to enable the post-earthquake radar image to be consistent with a coordinate system of the optical remote sensing image system and consistent with resolution, and obtaining a preset radar image.

According to the preset radar image, a Gray-level co-occurrence matrix (GLCM) of the preset radar image can be calculated.

Specifically, the statistical rule that the gray scale (i, j) of a pixel appears in a certain direction θ and a certain distance s in an image can be used to reflect the gray scale texture feature of the image, and the feature is generally represented by a matrix, which is a gray scale co-occurrence matrix.

In this embodiment, let θ be: 0 °, 45 °, 90 °, 135 °.

Assuming that the gray level of the preset radar image is N, the gray level co-occurrence matrix P of the preset radar image is an N × N square matrix, and taking any point (x, y) on the preset radar image and a point (Δ x, Δ y) at a distance s in the θ direction of the point, the gray level of the point (x, y) is (i, j). Moving the point (x, y) over the entire image, the nxn (i, j) values of the point (x, y) are obtained. And (3) counting the occurrence frequency of each (i, j) value for the whole image, and then arranging the (i, j) values into a square matrix to form a gray level co-occurrence matrix of the preset radar image. At the same time, (i, j) are normalized to the probability p (i, j) of occurrence of such gray value (i, j) according to the total number of occurrences of (i, j).

And then obtaining eight gray texture feature maps of the preset radar image according to formulas (1) to (8) based on the gray co-occurrence matrix of the preset radar image, wherein the eight gray texture feature maps comprise: homogeneity map, angular second moment map, contrast map, entropy map, correlation map, mean sum map, variance sum map, and dissimilarity map. The specific process is as follows:

obtaining a uniformity map of a preset radar image according to a formula (1):

where HOM is Homogeneity (Homogeneity), a measure of how homogeneous an image is.

Obtaining an angular second moment diagram of the preset radar image according to the formula (2):

wherein, the ASM is an angle second moment (angle second moment), which represents the uniformity of the image gray scale texture variation and reflects the image gray scale distribution and the texture thickness.

Obtaining a contrast map of the preset radar image according to a formula (3):

where CON is contrast (contrast), which reflects the sharpness of the image texture.

Obtaining an entropy diagram of the preset radar image according to a formula (4):

wherein ENT is Entropy (Entropy) and reflects the amount of information contained in the image.

Obtaining a correlation chart of the preset radar image according to the formula (5):

where COR is Correlation (Correlation), which is a measure of the linear relationship of image gray scales, and reflects the directionality of image texture.

Obtaining a mean value graph of a preset radar image according to a formula (6):

the MEA is a Mean value (Mean), which is generally referred to as a Mean value for short, and reflects the degree of regularity of the image texture.

Obtaining the variance and the graph of the preset radar image according to the formula (7):

where VAR is a sum of Variance (Variance), generally referred to as Variance for short, and represents a discrete value of the image luminance value.

Obtaining a difference map of the preset radar image according to the formula (8):

DIS is variance (variance) and indicates the size of the difference in the gray level of the image.

Based on the steps, eight gray-scale texture feature maps of the preset radar image can be obtained.

And S160, combining the post-earthquake change area and the prediction deformation area to obtain a target area.

By combining the obtained predicted deformation region and the after-earthquake change region, a target region including the after-earthquake change region and the predicted deformation region can be obtained, and the target region is a region in which landslide hazard detection is required in this embodiment.

S170, calculating characteristic parameters of each object in the target area according to the gray texture characteristic map, the remote sensing image after the earthquake, the preset digital elevation model of the area to be detected and the preset deformation rate map.

Before the calculation, the target region needs to be divided to obtain a plurality of objects of the target region.

Specifically, the target region may be segmented by a multi-scale segmentation method, so as to obtain a plurality of objects of the target region.

The quality of the segmentation result directly affects the image classification accuracy. Among them, the setting of the segmentation scale has a great influence on the segmentation result. In one possible implementation, to avoid segmenting the object too small or too large, an optimal segmentation scale may be determined before segmentation. And performing segmentation according to the optimal segmentation scale, so that the obtained multiple objects of the target area can effectively distinguish the ground objects.

In order to compensate for the limitation of landslide detection performed by an optical remote sensing image alone, the embodiment performs comprehensive judgment according to the characteristics of the optical remote sensing image, the gray texture characteristics of the SAR image, the topographic characteristics of the area to be detected, and the deformation rate characteristics of the SAR image.

Firstly, an unprocessed preset DEM (Digital Elevation Model) of an area to be detected can be downloaded from high-precision terrain grid data, and the unprocessed preset DEM data is projected and resampled to be consistent with a coordinate system and resolution of an optical remote sensing image system, so that the processed preset Digital Elevation Model of the area to be detected is obtained.

And then, calculating characteristic parameters of each object in the target area according to the obtained eight gray texture characteristic graphs, the remote sensing image after the earthquake, a preset digital elevation model of the area to be detected and a preset deformation rate graph. The characteristic parameters of the object can indicate the characteristics of the optical remote sensing image of the object, the gray texture characteristics of the SAR image, the terrain characteristics and the deformation rate characteristics of the SAR image.

And S180, eliminating the area where the non-landslide ground object in the target area is located according to the characteristic parameters of each object in the target area to obtain a landslide hidden danger area.

According to the characteristic parameters of each object in the target area, areas where non-landslide land features are located are removed, namely areas where some land features such as roads, artificial buildings, water bodies and the like which do not slide are removed, the areas are possibly detected to have surface changes due to road blockage, house collapse and water level changes, but are areas where the non-landslide land features are located, the areas are removed in the target area, and therefore a landslide hidden danger area can be obtained, and the area with landslide hidden dangers is obtained. Specifically, in the area of the landslide hazard, some areas are areas where primary landslide has occurred and where a secondary landslide risk exists, and some areas are areas where landslide has not occurred but where a landslide risk exists due to a change in the earth's surface. The landslide hidden danger areas are obtained according to the characteristic parameters of each object in the target area, so that the landslide hidden danger areas can be judged more accurately, the landslide hidden danger areas can be further protected, and further disasters can be prevented from occurring.

In the embodiment, the change detection is carried out on the registered remote sensing image, and the deformation rate graph is utilized to obtain the predicted deformation area, so that the finally obtained target area not only has the change area obtained by the change detection, but also has the area subjected to deformation, and the target area information for analyzing landslide detection is more complete and comprehensive; meanwhile, the characteristic parameters of each object in the target area, which are calculated by using the gray texture characteristic map, the remote sensing image after the earthquake, the preset digital elevation model and the preset deformation rate map, are used for removing the area of the non-landslide land object in the target area, so that the finally obtained landslide hidden danger area is more accurate, powerful support can be provided for landslide detection, the landslide detection efficiency is improved, geological detection and analysis can be more efficiently and reliably carried out in the face of landslide disaster, and the landslide hidden danger area can be protected in time.

On the basis of the landslide detection method provided by the above fig. 1, the present invention also provides a possible implementation manner of the remote sensing image registration method. Fig. 2 is a schematic flow chart of a remote sensing image registration method provided by the present invention. As shown in fig. 2, in S120, the registering the pre-earthquake remote sensing image to the post-earthquake remote sensing image with the post-earthquake remote sensing image as a reference to obtain a registered pre-earthquake remote sensing image includes:

and S122, performing ortho-rectification on the pre-earthquake remote sensing image and the post-earthquake remote sensing image to obtain the pre-earthquake ortho-sensing image and the post-earthquake ortho-sensing image.

In order to ensure the registration accuracy, before the registration, the remote sensing image before the earthquake and the remote sensing image after the earthquake need to be subjected to orthorectification. For the optical remote sensing images, each optical remote sensing image comprises panchromatic waveband data and four-waveband multispectral data. Specifically, a complete RPC model can be obtained by using an RPC (Rational Polynomial Coefficients) file and an RPC model form carried by optical remote sensing image data; and performing orthorectification on panchromatic waveband data and four-waveband multispectral data of the pre-earthquake remote sensing image and the post-earthquake remote sensing image respectively by utilizing a preset digital elevation model and a complete RPC model of the area to be detected, and obtaining the pre-earthquake orthometric remote sensing image and the post-earthquake orthometric remote sensing image respectively, wherein the pre-earthquake orthometric remote sensing image comprises the panchromatic waveband data and the four-waveband multispectral data. In a possible implementation mode, radiation correction can be further carried out on the remote sensing image before the earthquake and the remote sensing image after the earthquake, so that the problem of radiation distortion caused by atmospheric refraction and solar altitude is solved.

And S124, respectively fusing panchromatic band data and multispectral data in the pre-earthquake ortho-remote sensing image and the post-earthquake ortho-remote sensing image to obtain a fused pre-earthquake remote sensing image and a fused post-earthquake remote sensing image.

After the orthorectification, the images also need to be fused in order to enrich the spectral information of the images. Specifically, a preset image fusion algorithm, such as NNDiffuse Pan imaging algorithm, may be used to fuse panchromatic band data and multispectral data in the pre-earthquake ortho-remote sensing image and the post-earthquake ortho-remote sensing image, respectively, so as to obtain a fused pre-earthquake remote sensing image and a fused post-earthquake remote sensing image.

And S126, registering the fused remote sensing image before the earthquake to the fused remote sensing image after the earthquake as a reference by taking the fused remote sensing image after the earthquake as a reference to obtain a registered remote sensing image before the earthquake.

After the fusion processing, the fused remote sensing image after earthquake is taken as a reference, and the fused remote sensing image before earthquake is registered to the fused remote sensing image after earthquake as a reference, so that the registered remote sensing image before earthquake is obtained. The specific registration process is as described in S120, and is not described again.

In this embodiment, the optical remote sensing image is preprocessed before registration, so that the registration accuracy can be improved, and interference of invalid image information is avoided.

On the basis of the landslide detection method provided by the above fig. 1, the present invention also passes a possible implementation manner of the change detection method. Fig. 3 is a schematic flow chart of a change detection method according to the present invention. As shown in fig. 3, in step S130, performing change detection on the post-earthquake remote sensing image according to the registered pre-earthquake remote sensing image to obtain an image of a post-earthquake changed region and an image of a post-earthquake unchanged region, including:

s132, principal component transformation is respectively carried out on the registered pre-earthquake remote sensing image and the registered post-earthquake remote sensing image, and a first principal component image of the registered pre-earthquake remote sensing image and a first principal component image of the registered post-earthquake remote sensing image are obtained.

And the ground object information in the first principal component image of the remote sensing image after the earthquake corresponds to the ground object information of the remote sensing image after the earthquake one by one.

Specifically, the feature information includes feature position information and feature type information.

The principal component transform can represent a remote sensing image containing a large number of spectral bands with a small number of bands. The information of the image after principal component transformation is hardly lost, but the data size can be obviously reduced. In this embodiment, principal component transformation is performed on the registered pre-earthquake remote sensing image and post-earthquake remote sensing image respectively, so that a plurality of principal component images of the registered pre-earthquake remote sensing image and a plurality of principal component images of the post-earthquake remote sensing image are obtained, and a first principal component image of the registered pre-earthquake remote sensing image and a first principal component image of the post-earthquake remote sensing image are selected. The first principal component image is an image contained in a first principal component of the remote sensing image after principal component transformation. By observing the images included in the first principal component of the two images after principal component conversion, all the feature information included in the two images can be obtained with a small data amount.

And S134, determining a region in which the position offset between the pixels in the first principal component image of the post-earthquake remote sensing image and the corresponding pixels in the first principal component image of the registered pre-earthquake remote sensing image exceeds a preset offset as a post-earthquake change region in the first principal component image of the post-earthquake remote sensing image.

Each pixel in the first principal component image of the post-earthquake remote sensing image corresponds to each pixel in the first principal component image of the pre-earthquake remote sensing image after registration; comparing the position information of each pixel in the first principal component image of the post-earthquake remote sensing image with the position information of each corresponding pixel in the first principal component image of the pre-earthquake remote sensing image after registration to obtain the position offset between each pair of pixels in the two principal component images; presetting offset, and determining a region corresponding to the pixel with the position offset exceeding the preset offset as a post-earthquake change region in the first principal component image of the post-earthquake remote sensing image. That is, the position offset between the pixel positions in the post-earthquake change region in the first principal component image of the post-earthquake remote sensing image and the corresponding pixel positions in the first principal component image of the post-earthquake remote sensing image after registration exceeds the preset offset.

And S136, determining an image of the post-earthquake change area from the post-earthquake remote sensing image according to the post-earthquake change area.

Since the feature information in the first principal component image of the post-earthquake remote sensing image corresponds to the feature information of the registered pre-earthquake remote sensing image one by one, the image of the post-earthquake change region can be determined from the post-earthquake remote sensing image according to the post-earthquake change region in the first principal component image of the post-earthquake remote sensing image.

And S138, determining the images of other areas except the post-earthquake change area in the post-earthquake remote sensing image as the images of the post-earthquake non-change area.

After the image of the post-earthquake changed region is determined, determining the images of other regions except the post-earthquake changed region in the post-earthquake remote sensing image as the images of the post-earthquake unchanged region. That is, compared with the remote sensing image after the earthquake, the image of the unchanged area after the earthquake has no obvious surface deformation.

In the embodiment, the change detection of the remote sensing image is carried out by utilizing the principal component transformation, so that the surface feature information of the image is presented while the image data volume is reduced, the workload is reduced, and the changed and unchanged areas are simply, quickly and efficiently analyzed.

On the basis of the landslide detection method provided by the above fig. 1, the present invention also provides a possible implementation manner of extracting a predicted deformation region. FIG. 4 is a schematic flow chart of extracting a predicted deformation region according to the present invention. As shown in fig. 4, in the step S140, extracting the predicted deformation region from the post-earthquake non-change region according to the image of the post-earthquake non-change region and the preset deformation information map of the region to be detected includes:

and S142, calculating the deformation rate value of each pixel in the image of the post-earthquake non-change area according to the preset deformation rate graph.

And according to the position of each pixel in the processed deformation rate diagram of the region to be detected, the position of each pixel in the image of the post-earthquake non-change region can be corresponded, so that the deformation rate value of each pixel in the image of the post-earthquake non-change region is calculated.

S144, determining an area formed by pixels of which the deformation rate value is within a preset deformation threshold range in the image of the post-earthquake non-change area as a predicted deformation area.

Setting a preset deformation threshold range according to the deformation rate value of each pixel in the image of the post-earthquake non-change area, extracting an area formed by pixels with the deformation rate values within the preset deformation threshold range as a predicted deformation area, wherein the predicted deformation area is an area predicted to possibly have secondary disasters in the post-earthquake non-change area.

In the embodiment, the deformation prediction region is obtained by using the preset deformation rate graph and the optical remote sensing image, so that the landslide detection result is more complete, the identification of the hidden danger region is increased, and the problem that secondary landslide cannot be detected is avoided.

On the basis of the landslide detection method provided by the above fig. 1, the present invention also provides a possible implementation manner of calculating the object feature parameters. Fig. 5 is a schematic flow chart of calculating object feature parameters according to the present invention. As shown in fig. 5, the characteristic parameters of each object include: gray texture characteristic values, multispectral values, terrain characteristic values and deformation rate values; in the above S150, calculating the characteristic parameters of each object in the target area according to the gray texture feature map, the post-earthquake remote sensing image, the preset digital elevation model of the area to be detected, and the preset deformation rate map includes:

and S210, vectorizing the boundary of the target area to obtain a vector boundary diagram of the target area.

And S220, extracting a remote sensing image map of the target area, a digital elevation model of the target area, a deformation rate map of the target area and a gray texture feature map of the target area from the remote sensing image after the earthquake, a preset digital elevation model, a preset deformation rate map and a gray texture feature map respectively by using a vector boundary map.

Because the obtained boundary of the target area is fuzzy and fine, the vectorization of the boundary of the target area is needed in order to obtain the clear boundary of the target area, so that the vector boundary image of the target area is obtained by determining the boundary, and therefore, the vector boundary image can be adopted to draw out the remote sensing image map of the target area from the remote sensing image after the earthquake, and the remote sensing image map of the target area is extracted and obtained; the method comprises the steps that a vector boundary graph is adopted, a remote sensing image map of a target area is outlined from a preset digital elevation model, and therefore a digital elevation model of the target area is extracted; adopting a vector boundary graph to outline a remote sensing image map of a target area from a preset deformation rate graph so as to extract and obtain a deformation rate graph of the target area; adopting a vector boundary graph, and drawing a gray texture feature graph of the target area from the gray texture feature graph so as to extract the gray texture feature graph of the target area; alternatively, if eight gray-scale texture feature maps of the preset radar image are obtained in S140, eight gray-scale texture feature maps of the target region may be extracted from the eight gray-scale texture feature maps by using the vector boundary map.

After the remote sensing image map of the target area is obtained, before calculation, the multispectral data in the remote sensing image map of the target area can be segmented to obtain a plurality of objects of the target area.

Specifically, a multi-scale segmentation method may be used to segment multispectral data in the remote sensing image map of the target region, so as to obtain a plurality of objects in the target region.

And S230, calculating to obtain a terrain characteristic value of each object in the target area according to the digital elevation model of the target area.

The method comprises the steps of firstly obtaining terrain characteristic values of all pixels forming an object in each object in a target area according to a digital elevation model of the target area, then respectively calculating the terrain characteristic value average value of all pixels forming the object in each object, wherein the average value of the terrain characteristic values of all pixels in one object is the terrain characteristic value of the object.

And S240, calculating a gray texture characteristic value, a multispectral value and a deformation rate value of each object in the target area according to the gray texture characteristic image of the target area, the remote sensing image of the target area and the deformation rate image of the target area.

According to the gray texture feature map of the target area, the remote sensing image map of the target area and the deformation rate map of the target area, eight gray texture feature values, four-waveband multispectral values and deformation rate values of all pixels forming the object in each object in the target area are calculated. The eight gray texture feature values include: homogeneity values, angular second moment values, contrast values, entropy values, relevance values, mean sum value maps, variance sum values, and dissimilarity values.

Secondly, respectively calculating the average value of each gray scale texture characteristic value, the average value of four-waveband multispectral values and the average value of deformation rate in the eight gray scale texture characteristic values of all pixels forming the object in each object.

Then, the average value of the gray scale texture feature values of all the pixels in one object is the gray scale texture feature value of the object, for example, the average value of the uniformity values of all the pixels in one object is the uniformity value of the object; the average value of the four-band multispectral values of all pixels in an object is the four-band multispectral value of the object; the average value of the four-band multispectral values of all the pixels in one object is the four-band multispectral value of the object.

And finally obtaining the terrain characteristic value, the deformation rate value, the multispectral value and the gray texture characteristic value of each object in the target area, and obtaining the characteristic parameters of each object in the target area.

In this embodiment, by calculating the characteristic parameters of each object, the landslide in the target area can be divided based on the characteristics of the optical remote sensing image of the target area, the gray texture characteristics of the SAR image, the topographic characteristics, and the deformation rate characteristics of the SAR image, so that the accuracy of landslide detection is improved.

On the basis of the method for calculating the characteristic parameters of the object provided by the above fig. 5, the present invention also provides a possible implementation manner for obtaining the landslide hazard area. Fig. 6 is a schematic flow chart for obtaining a landslide hazard area according to the present invention. As shown in fig. 6, in the step S180, the removing, according to the feature parameter of each object in the target region, a region where the non-landslide terrain is located in the target region to obtain the landslide hazard region includes:

and S310, selecting a plurality of ground object samples in the remote sensing image of the target area by taking the object as a unit.

The method comprises the steps that a plurality of surface feature samples comprise all surface feature types in a remote sensing image of a target area, wherein one surface feature sample comprises only one surface feature type;

the remote sensing image of the target area contains various surface features, visual analysis is needed to obtain surface feature classification rules, and a plurality of surface feature samples in the remote sensing image of the target area are selected by taking an object as a unit. For example, a first sample is selected as a water body sample, a second sample is selected as a mountain body sample, and so on, and the selected multiple sample types need to include all the surface feature types in the remote sensing image of the target area. Furthermore, it is necessary to ensure that only one type of feature is contained in a single sample.

And S320, obtaining the value range of the characteristic parameters of each surface feature type object by using a preset decision tree algorithm according to the characteristic parameters of all the objects in the plurality of surface feature samples.

The basic composition unit of each surface feature sample is an object, each object has a corresponding characteristic parameter, And the value range of the characteristic parameter corresponding to the object of each surface feature type can be obtained by using a preset decision Tree algorithm, such as a CART (Classification And Regression Tree) algorithm, according to the characteristic parameters of all the objects in the plurality of surface feature samples.

And S330, obtaining the corresponding relation between the value range of the characteristic parameter of each object in the target area and each surface feature type according to the value range of the characteristic parameter of each surface feature type object, and generating a preset surface feature classification rule of the target area.

Because the plurality of surface feature samples contain all surface feature types in the remote sensing image map of the target area, the corresponding relation between the characteristic parameter value range of each object in the target area and each surface feature type can be obtained according to the characteristic parameter value range of each surface feature type object, namely, the value range of the characteristic parameter of each surface feature type object can be compared according to the characteristic parameter value of each object in the target area, so that the surface feature type of each object can be obtained, and further, a preset surface feature classification rule of the target area is generated.

And S340, classifying the remote sensing image map of the target area by adopting a preset ground object classification rule of the target area to obtain a ground object classification result in the target area.

Due to the fact that the representation forms of the ground objects in the target area are various, in order to accurately and finally divide the landslide hidden danger area, the remote sensing image map of the target area needs to be classified by using a preset ground object classification rule, and a ground object classification result in the target area is obtained. For example, the remote sensing image of the target area is classified into types of land and objects such as water, vegetation, bare land, roads, mountains, and artificial buildings.

And S350, according to the ground feature classification result, eliminating the region where the non-landslide ground features in the target region are located to obtain a landslide hidden danger region.

According to the land feature classification result of the remote sensing image map of the target area, regions of land features such as roads, artificial buildings, water bodies and the like which cannot slide are removed, and finally the potential danger regions of the slide are obtained.

In this embodiment, by obtaining the characteristic parameters corresponding to the surface feature objects, the preset surface feature classification rules adapted to each object in the target area can be generated, the surface feature conditions in the target area can be better reflected, misjudgment caused by all visual classification is avoided, the target surface features are classified, non-landslide surface features are convenient to reject, landslide hidden danger areas are obtained quickly and effectively, landslide analysis areas are reduced, and unnecessary analysis is omitted.

On the basis of the calculation of the object characteristic parameters provided by the above-mentioned fig. 5, the present invention also provides a possible implementation manner of calculating the terrain characteristic values. Fig. 7 is a schematic flowchart of a process for calculating a terrain feature value according to the present invention. As shown in fig. 7, the topographic feature values include: the slope value, the ground elevation value, the mountain shadow value, the relief value, in the above-mentioned S320, after selecting a plurality of surface feature samples in the remote sensing image map of the target area with the object as the unit, include:

and S410, calculating to obtain a slope map, a mountain shadow map, a surface relief map and a ground elevation map of the target area according to the digital elevation model of the target area.

Specifically, a gradient map of the target area is calculated according to equation (9);

wherein slope is slope, f_xIs the elevation change rate in the X direction in the digital elevation model of the target area, f_yThe elevation change rate in the Y direction in the digital elevation model of the target area is obtained;

calculating a mountain shadow map of the target area according to formula (10);

hillshade＝255×(cos(zenith_rad)×cos(slope_rad))+

(sin(zenith_rad)×sin(slope_rad)×cos(azimuth_rad-aspect_rad)) (10)

hillshade∈[0,255],zenith_rad,

azimuth_rad,aspect_rad∈[0,2π]

wherein hillshade is mountain shadow value, zenith_radRadian number of sun zenith angle for obtained remote sensing image，slope_radIs the number of slope arcs in the digital elevation model of the target area, azimuth_radMeasure the number of radians in the direction of the sun's rays for the acquired remote-sensing image_radIs the number of slope arc degrees in the digital elevation model of the target area.

Calculating a surface relief map and a surface elevation map of the target area according to the formula (11);

R＝H_max-H_min,R＞0,H_max,H_min∈R (11)

wherein R is the ground waviness, H_maxIs the maximum elevation, H, within a fixed analysis window in a digital elevation model of the target area_minIs the lowest elevation within the corresponding fixed analysis window in the digital elevation model for the target area.

And S420, respectively calculating a slope value, a mountain shadow value, a terrain relief value and a ground elevation value of each object in the target area according to the slope map, the mountain shadow map, the terrain relief map and the ground elevation map.

Firstly, obtaining the slope value, the ground elevation value, the mountain shadow value and the terrain relief value of all pixels forming the object in each object in the target area according to the slope map, the mountain shadow map, the terrain relief map and the ground elevation map of the target area respectively.

Secondly, respectively calculating the average value of the slope values of all pixels forming the object, the average value of the mountain shadow values, the average value of the surface relief values and the average value of the ground elevation values in each object, wherein the average value of the slope values of all pixels in one object is the slope value of the object, the average value of the mountain shadow values of all pixels in one object is the mountain shadow value of the object, the average value of the surface relief values of all pixels in one object is the surface relief value of the object, and the average value of the ground elevation values of all pixels in one object is the ground elevation value of the object.

And finally obtaining the slope value, the ground elevation value, the mountain shadow value and the terrain relief value of each object in the target area.

By using the characteristic parameters, the characteristic parameters of all the objects in a plurality of surface feature samples can be obtained, and then the value range of the characteristic parameters of each surface feature type object is obtained by using a preset decision tree algorithm.

In this embodiment, the slope value, the ground elevation value, the mountain shadow value, and the terrain relief value of each object in the target area are calculated by using the digital elevation model of the target area, so that the terrain feature of the target area is obtained, and the determination of the landslide area is facilitated by using the terrain feature.

The following describes a landslide detection device and a landslide detection apparatus provided by the present application for implementation, and specific implementation processes and technical effects thereof are referred to above, and will not be described again below.

Fig. 8 is a schematic diagram of a landslide detection apparatus provided in the present invention, and as shown in fig. 8, the landslide detection apparatus includes:

the acquisition module 1000 is used for acquiring a pre-earthquake remote sensing image and a post-earthquake remote sensing image of a region to be detected from the optical remote sensing image system;

the registration module 2000 is configured to register the pre-earthquake remote sensing image to the post-earthquake remote sensing image by using the post-earthquake remote sensing image as a reference, so as to obtain a registered pre-earthquake remote sensing image;

the detection module 3000 is configured to perform change detection on the post-earthquake remote sensing image according to the registered pre-earthquake remote sensing image to obtain an image of a post-earthquake changed region and an image of a post-earthquake unchanged region;

the extraction module 4000 is configured to extract a predicted deformation region from the post-earthquake non-change region according to the image of the post-earthquake non-change region and the preset deformation rate map of the region to be detected; the coordinate system of the preset deformation rate graph is consistent with the coordinate system of the optical remote sensing image system;

the first calculation module 5000 is configured to calculate a gray level texture feature map of a preset radar image according to a gray level co-occurrence matrix of the preset radar image of the to-be-detected region; the coordinate system of the preset radar image is consistent with the coordinate system of the optical remote sensing image system;

the merging module 6000 is used for merging the post-earthquake change area and the prediction deformation area to obtain a target area;

the second calculation module 7000 is further configured to calculate a characteristic parameter of each object in the target region according to the gray texture feature map, the post-earthquake remote sensing image, the preset digital elevation model of the region to be detected, and the preset deformation rate map;

and the eliminating module 8000 is configured to eliminate a region where a non-landslide ground object in the target region is located according to the characteristic parameter of each object in the target region, so as to obtain a landslide hidden danger region.

Optionally, the registration module 2000 is further configured to perform orthorectification on the pre-earthquake remote sensing image and the post-earthquake remote sensing image to obtain a pre-earthquake orthometric remote sensing image and a post-earthquake orthometric remote sensing image; respectively fusing panchromatic band data and multispectral data in the pre-earthquake ortho-remote sensing image and the post-earthquake ortho-remote sensing image to obtain a fused pre-earthquake remote sensing image and a fused post-earthquake remote sensing image; and taking the fused post-earthquake remote sensing image as a reference, and registering the fused pre-earthquake remote sensing image to the fused post-earthquake remote sensing image as a reference to obtain a registered pre-earthquake remote sensing image.

Optionally, the detection module 3000 is further specifically configured to perform principal component transformation on the registered pre-earthquake remote sensing image and post-earthquake remote sensing image respectively to obtain a first principal component image of the registered pre-earthquake remote sensing image and a first principal component image of the post-earthquake remote sensing image; the feature information in the first principal component image of the registered pre-earthquake remote sensing image corresponds to the feature information of the registered pre-earthquake remote sensing image one by one, and the feature information in the first principal component image of the post-earthquake remote sensing image corresponds to the feature information of the post-earthquake remote sensing image one by one; determining an area in which the position offset between the pixel in the first principal component image of the post-earthquake remote sensing image and the corresponding pixel in the first principal component image of the registered pre-earthquake remote sensing image exceeds a preset offset as a post-earthquake change area in the first principal component image of the post-earthquake remote sensing image; determining an image of the post-earthquake change region from the post-earthquake remote sensing image according to the post-earthquake change region; and determining the images of other areas except the post-earthquake changed area in the post-earthquake remote sensing image as the images of the post-earthquake unchanged area.

Optionally, the extracting module 4000 is further specifically configured to calculate a deformation rate value of each pixel in the image of the post-earthquake unchanged area according to a preset deformation rate map; and determining an area formed by pixels of which the deformation rate values are within a preset deformation threshold range in the image of the post-earthquake non-change area as a predicted deformation area.

Optionally, the characteristic parameters of each object include: gray texture characteristic values, multispectral values, terrain characteristic values and deformation rate values.

Optionally, the extracting module 4000 is further configured to perform vectorization on the boundary of the target region to obtain a vector boundary map of the target region; and respectively extracting a remote sensing image map of the target area, a digital elevation model of the target area, a deformation rate map of the target area and a gray texture feature map of the target area from the post-earthquake remote sensing image, a preset digital elevation model, a preset deformation rate map and the gray texture feature map by adopting a vector boundary map.

Optionally, the second calculating module 7000 is further specifically configured to calculate, according to the digital elevation model of the target area, a terrain feature value of each object in the target area; and calculating a gray texture characteristic value, a multispectral value and a deformation rate value of each object in the target area according to the gray texture characteristic image of the target area, the remote sensing image of the target area and the deformation rate image of the target area.

Optionally, the removing module 8000 is further configured to specifically select a plurality of surface feature samples in the remote sensing image of the target area by using the object as a unit; the multiple surface feature samples comprise all surface feature types in the remote sensing image of the target area, wherein one surface feature sample comprises only one surface feature type; obtaining the value range of the characteristic parameters of each surface feature type object by using a preset decision tree algorithm according to the characteristic parameters of all objects in a plurality of surface feature samples; according to the value range of the characteristic parameter of each surface feature type object, obtaining the corresponding relation between the value range of the characteristic parameter of each object in the target area and each surface feature type, and generating a preset surface feature classification rule of the target area; classifying the remote sensing image map of the target area by adopting a preset ground object classification rule of the target area to obtain a ground object classification result in the target area; and according to the land feature classification result, eliminating the region where the non-landslide land feature is located in the target region to obtain a landslide hidden danger region.

Optionally, the terrain characteristic value comprises: slope value, ground elevation value, mountain shadow value and topographic relief value.

Optionally, the second calculation module 7000 is further specifically configured to calculate a slope map, a mountain shadow map, a relief map, and a ground elevation map of the target area according to the digital elevation model of the target area; and respectively calculating the slope value, the mountain shadow value, the terrain relief value and the ground elevation value of each object in the target area according to the slope map, the mountain shadow map, the terrain relief map and the ground elevation map.

These above modules may be one or more integrated circuits configured to implement the above methods, such as: one or more Application Specific Integrated Circuits (ASICs), or one or more Digital Signal Processors (DSPs), or one or more Field Programmable Gate Arrays (FPGAs), among others. For another example, when one of the above modules is implemented in the form of a Processing element scheduler code, the Processing element may be a general-purpose processor, such as a Central Processing Unit (CPU) or other processor capable of calling program code. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).

Fig. 9 is a schematic diagram of a landslide detection apparatus provided in the present invention, which may be a computing apparatus or a server with a computing processing function.

The landslide detection apparatus 10 includes: a processor 11, a storage medium 12 and a bus 13, the storage medium 12 storing program instructions executable by the processor 11, the processor 11 communicating with the storage medium 12 via the bus 13 when the landslide detection apparatus 10 is executed, the processor 11 executing the program instructions to perform the above-described method embodiments. The specific implementation and technical effects are similar, and are not described herein again.

Optionally, the present invention also provides a program product, for example a computer-readable storage medium, comprising a program which, when being executed by a processor, is adapted to carry out the above-mentioned method embodiments.

In the embodiments provided in the present invention, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one type of logical functional division, and other divisions may be realized in practice, for example, multiple units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional unit.

The integrated unit implemented in the form of a software functional unit may be stored in a computer readable storage medium. The software functional unit is stored in a storage medium and includes several instructions to enable a computer device (which may be a personal computer, a server, or a network device) or a processor (processor) to execute some steps of the methods according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a portable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other media capable of storing program codes.

The above description is only for the specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the appended claims.

Claims

1. A landslide detection method, the method comprising:

2. The method according to claim 1, wherein the registering the pre-earthquake remote sensing image to the post-earthquake remote sensing image by taking the post-earthquake remote sensing image as a reference to obtain a registered pre-earthquake remote sensing image comprises:

3. The method according to claim 1, wherein the performing change detection on the post-earthquake remote sensing image according to the registered pre-earthquake remote sensing image to obtain an image of a post-earthquake changed region and an image of a post-earthquake unchanged region comprises:

determining an area in which the position offset between the pixel in the first principal component image of the post-earthquake remote sensing image and the corresponding pixel in the first principal component image of the registered pre-earthquake remote sensing image exceeds a preset offset as a post-earthquake change area in the first principal component image of the post-earthquake remote sensing image;

4. The method according to claim 1, wherein the extracting a predicted deformation region from the post-earthquake unchanged region according to the image of the post-earthquake unchanged region and the preset deformation information map of the region to be detected comprises:

5. The method of claim 1, wherein the characteristic parameters of each object comprise: gray texture characteristic values, multispectral values, terrain characteristic values and deformation rate values; calculating the characteristic parameters of each object in the target area according to the gray texture characteristic diagram, the post-earthquake remote sensing image, the preset digital elevation model of the area to be detected and the preset deformation rate diagram, wherein the calculation comprises the following steps:

extracting a remote sensing image of the target area, a digital elevation model of the target area, a deformation rate diagram of the target area and a gray texture feature diagram of the target area from the post-earthquake remote sensing image, the preset digital elevation model, the preset deformation rate diagram and the gray texture feature diagram respectively by using the vector boundary diagram;

according to the digital elevation model of the target area, calculating to obtain a terrain characteristic value of each object in the target area;

6. The method according to claim 5, wherein the step of eliminating the region where the non-landslide terrain is located in the target region according to the characteristic parameters of each object in the target region to obtain a landslide hazard region comprises the steps of:

7. The method of claim 5, wherein the terrain feature values comprise: the method comprises the following steps of calculating a slope value, a ground elevation value, a mountain shadow value and a terrain relief value according to the digital elevation model of the target area to obtain a terrain characteristic value of each object in the target area, wherein the steps comprise:

calculating to obtain a slope map, a mountain shadow map, a surface relief map and a ground elevation map of the target area according to the digital elevation model of the target area;

8. A landslide detection device, wherein the device comprises:

the acquisition module is used for acquiring a pre-earthquake remote sensing image and a post-earthquake remote sensing image of the area to be detected from the optical remote sensing image system;

9. A landslide detection apparatus, comprising: a processor, a memory and a bus, the memory storing program instructions executable by the processor, the processor and the memory communicating via the bus when the landslide detection apparatus is in operation, the processor executing the program instructions to perform the steps of the landslide detection method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that the storage medium has stored thereon a computer program which, when being executed by a processor, carries out the landslide detection method according to any one of claims 1-7.