Building disaster-bearing body remote sensing recognition method based on angular point features

Abstract

The invention discloses a building disaster-bearing body remote sensing recognition method based on angular point features, and relates to the technical field of building disaster-bearing body recognition, and the method comprises the following steps: S1, collecting high-resolution optical remote sensing images in a concentrated manner, and carrying out the preprocessing of the high-resolution optical remote sensing images; s2, extracting image angular points of the preprocessed high-resolution optical remote sensing image set; s3, determining newly added angular points according to the image angular points; and S4, generating an outer boundary according to the newly added angular points, and completing remote sensing identification of the building disaster-bearing body. According to the method, a ground object sample does not need to be manufactured to carry out model pre-training, the building disaster-bearing body can be recognized only by depending on changes of image corner point features, the omission phenomenon in manual visual interpretation is avoided, and the method can be suitable for two scenes of recognition of a newly-added building disaster-bearing body before a disaster and recognition of a damaged building disaster-bearing body after the disaster.

Description

Technical Field

[0001] This invention relates to the field of building disaster-bearing body identification technology, specifically to a remote sensing identification method for building disaster-bearing bodies based on corner features. Background Technology

[0002] In the field of geological disaster monitoring, the disaster-bearing body is the ultimate carrier of disaster risk. The energy and destructiveness of geological disasters (such as landslides and debris flows) are ultimately quantified and manifested through the damage they cause to the disaster-bearing body (such as building collapses, road interruptions, and casualties). It is not only a key factor in determining disaster losses, but also the core objective and object of disaster prevention and mitigation work. The ultimate goal of all geological disaster monitoring, early warning, assessment, and control projects is to protect the disaster-bearing body.

[0003] Buildings that bear the hazard of disasters refer to structures and buildings that may be damaged or lose their functionality in the event of natural disasters or other catastrophic events. They are one of the core targets of disaster prevention and mitigation efforts. Monitoring changes in these buildings is of great significance for pre-disaster risk assessment and post-disaster emergency response.

[0004] Currently, the identification of building disaster-bearing bodies is mainly achieved through manual visual interpretation or artificial intelligence interpretation. Manual visual interpretation requires excessive manpower and is prone to omissions in practical use. Artificial intelligence interpretation, on the other hand, requires providing the model with a sufficient number of different types of samples for training and learning in advance, resulting in high initial costs and an inability to handle complex terrain scenarios. Summary of the Invention

[0005] To address the above problems, this invention proposes a remote sensing identification method for building disaster-bearing bodies based on corner features.

[0006] The technical solution of this invention is: a remote sensing identification method for building disaster-bearing bodies based on corner features, comprising the following steps:

[0007] S1. Acquire high-resolution optical remote sensing images and perform preprocessing;

[0008] S2. Extract the corner points of the preprocessed high-resolution optical remote sensing image set;

[0009] S3. Determine the new corner points based on the image corner points;

[0010] S4. Based on the newly added corner points, generate the outer boundary and complete the remote sensing identification of the building's disaster-bearing body.

[0011] Furthermore, in S1, the preprocessing specifically includes: radiometric calibration, atmospheric correction, orthorectification, and image fusion.

[0012] Furthermore, in S2, when extracting corner points from images, the results of corner point extraction in non-building change areas are consistent, while the results of corner point extraction in building change areas show the greatest difference.

[0013] Furthermore, S3 includes the following sub-steps:

[0014] S31. Use a corner extraction algorithm to extract the corner set of the first image and the corner set of the second image in the high-resolution optical remote sensing image set.

[0015] S32. Overlay the corner data of the first image's corner set and the corner data of the second image's corner set in geospatial space, and generate a circle with a radius of [missing information - likely a radius value]. The area;

[0016] S33, Select radius as The corner points of the first building are obtained from the set of corner points of all corner points of the second image within the area;

[0017] S34. Subtract the first building corner point set from the corner point set of the second image to obtain the second building corner point set;

[0018] S35, radius When the radius is equal to the minimum radius threshold, repeat steps S32 to S34 to obtain the second set of building corner points minus the radius. When the maximum radius threshold is equal, repeat steps S32 to S34 to obtain the second set of building corner points, and obtain the new corner points.

[0019] Furthermore, in S33, the set of the first building corner points The expression is:

[0020] ;

[0021] in, For the corner points in the first set of corner points, The set of corner points of the first image. The set of corner points in the set of corner points of the first image Corner point The radius is The neighborhood of.

[0022] Furthermore, in S34, the set of the second building corner points The expression is:

[0023] ;

[0024] in, The set of corner points in the set of corner points of the first image Corner point The radius is The neighborhood, Let this be the set of the first building corner points. For a specific radius value, The radius is [0, 1].

[0025] Furthermore, in S4, convex hull polygon generation is used to transform newly added corner points into outer boundaries.

[0026] The beneficial effects of this invention are as follows: This invention provides a remote sensing identification method for building disaster-bearing bodies based on corner features, applicable to the identification of building disaster-bearing bodies in high-resolution optical remote sensing images. This method eliminates the need for pre-training of models by creating ground feature samples; it identifies building disaster-bearing bodies solely based on changes in image corner features, avoiding omissions in manual visual interpretation. It is applicable to both the identification of newly added building disaster-bearing bodies before a disaster and the identification of damaged building disaster-bearing bodies after a disaster. Attached Figure Description

[0027] Figure 1 A flowchart of a remote sensing identification method for building disaster-bearing bodies based on corner features;

[0028] Figure 2 Here is a diagram illustrating the principle of adding corner point extraction;

[0029] Figure 3 This is a schematic diagram of the process of extracting new corner points. Detailed Implementation

[0030] The embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0031] like Figure 1 As shown, this invention provides a remote sensing identification method for building disaster-bearing bodies based on corner features, comprising the following steps:

[0032] S1. Acquire high-resolution optical remote sensing images and perform preprocessing;

[0033] S2. Extract the corner points of the preprocessed high-resolution optical remote sensing image set;

[0034] S3. Determine the new corner points based on the image corner points;

[0035] S4. Based on the newly added corner points, generate the outer boundary and complete the remote sensing identification of the building's disaster-bearing body.

[0036] In this embodiment of the invention, in S1, the preprocessing specifically includes: radiometric calibration, atmospheric correction, orthorectification, and image fusion.

[0037] In this embodiment of the invention, during S2, when extracting corner points from images, the corner point extraction results in non-building change areas are consistent, while the corner point extraction results in building change areas show the greatest difference.

[0038] Currently, corner extraction algorithms include Harris corner detection, Fast corner detection, and ORB corner detection, among others. These algorithms differ in computational complexity and robustness, requiring consideration of specific application scenarios. For example, the Harris algorithm is simple and fast, suitable for general applications; the FAST algorithm is computationally fast, suitable for real-time applications; and the ORB algorithm has some robustness to rotation. This method only needs to ensure that the same corner extraction method is used for both image periods, guaranteeing consistent corner extraction results in non-building change areas while maximizing the difference in corner extraction results in building change areas.

[0039] In this embodiment of the invention, S3 includes the following sub-steps:

[0040] S31. Use a corner extraction algorithm to extract the corner set of the first image and the corner set of the second image in the high-resolution optical remote sensing image set.

[0041] S32. Overlay the corner data of the first image's corner set and the corner data of the second image's corner set in geospatial space, and generate a circle with a radius of [missing information - likely a radius value]. The area;

[0042] S33, Select radius as The corner points of the first building are obtained from the set of corner points of all corner points of the second image within the area;

[0043] S34. Subtract the first building corner point set from the corner point set of the second image to obtain the second building corner point set;

[0044] S35, radius When the radius is equal to the minimum radius threshold, repeat steps S32 to S34 to obtain the second set of building corner points minus the radius. When the maximum radius threshold is equal, repeat steps S32 to S34 to obtain the second set of building corner points, and obtain the new corner points.

[0045] In this embodiment of the invention, in S33, the first set of building corner points The expression is:

[0046] ;

[0047] in, For the corner points in the first set of corner points, The set of corner points of the first image. The set of corner points in the set of corner points of the first image Corner point The radius is The neighborhood of.

[0048] In this embodiment of the invention, in S34, the second set of building corner points The expression is:

[0049] ;

[0050] in, The set of corner points in the set of corner points of the first image Corner point The radius is The neighborhood, Let this be the set of the first building corner points. For a specific radius value, The radius is [0, 1].

[0051] Changes in corner features are highly correlated with changes in buildings; new corners will be generated in areas with newly added buildings or severely damaged buildings. Therefore, by utilizing the spatial relationship between corners in two image sets and employing buffer analysis, corners of newly added or severely damaged buildings can be extracted.

[0052] Assume there are two remote sensing images, a and b, acquired sequentially. Image b contains newly added or severely damaged buildings. Figure 2 As shown, a corner extraction algorithm is used to obtain the set of corner points in two images. and Overlay corner data in geospatial space to achieve... any corner point inside Centered on, the generation radius is The region, symbolically represented as the neighborhood. Then select all within the area. The corner points in the set are obtained Then use minus This will give you the set of building corner points. When the radius for The set of building corner points at time is denoted as .

[0053] For buildings present in both imagery periods, those with small positional deviations can be filtered out when r is small. As r increases, fewer background noise points are extracted from the corners, and the proportion of newly added corners from newly constructed or severely damaged buildings gradually increases. However, when r exceeds the interval distance between newly added corners, the proportion of newly added corners gradually decreases, while the proportion of linear background noise points such as roads and embankments gradually increases. To reduce the impact of background noise points while retaining the most effective newly added corner information, statistical methods are used to obtain minimum and maximum radius thresholds. and ,use minus You can then obtain information about the newly added corner points.

[0054] and The specific values ​​are obtained through statistical methods.

[0055] The newly extracted corner points in this step include not only corner points at the outer boundary of buildings but also corner points inside buildings. Outer boundary corner points are mainly generated by newly constructed buildings, while internal corner points are mainly generated by severely damaged buildings. Generally, to extract corner points from newly constructed buildings, two images from the non-disaster period with a relatively long image interval can be selected for identification. For post-disaster emergency response, to avoid interference from newly constructed buildings, two images from the period before and after the disaster can be selected for identification.

[0056] In this embodiment of the invention, in S4, the newly added corner points are transformed into outer boundaries using convex hull polygon generation.

[0057] The newly added corner points extracted in step 3 are transformed into the outer boundary of the newly added or severely damaged building disaster-bearing body by using convex hull polygon generation, so as to make the maximum use of the boundary information included in the corner points.

[0058] In embodiments of the present invention, such as Figure 3 As shown, sub-image a is a remote sensing image from 2022, and sub-image b is a remote sensing image from 2023. The red area represents the corner points in the image obtained using the Harris corner extraction algorithm. Statistical methods were used to determine the maximum radius threshold of 100m and the minimum radius threshold of 30m suitable for separating newly added building corner points in the two-year images. The newly added building corner points in this area in 2023 compared to 2022 are: As shown in the red area of ​​subgraph c.

[0059] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of this invention.

Claims

1. A remote sensing identification method for building disaster-bearing bodies based on corner features, characterized in that, Includes the following steps: S1. Acquire high-resolution optical remote sensing images and perform preprocessing; S2. Extract the corner points of the preprocessed high-resolution optical remote sensing image set; S3. Determine the new corner points based on the image corner points; S4. Based on the newly added corner points, generate the outer boundary and complete the remote sensing identification of the building's disaster-bearing body.

2. The remote sensing identification method for building disaster-bearing bodies based on corner features according to claim 1, characterized in that, In S1, the preprocessing specifically includes: radiometric calibration, atmospheric correction, orthorectification, and image fusion.

3. The remote sensing identification method for building disaster-bearing bodies based on corner features according to claim 1, characterized in that, In S2, when extracting corner points from images, the results of corner point extraction in non-building change areas are consistent, while the results of corner point extraction in building change areas show the greatest difference.

4. The remote sensing identification method for building disaster-bearing bodies based on corner features according to claim 1, characterized in that, S3 includes the following sub-steps: S31. Use a corner extraction algorithm to extract the corner set of the first image and the corner set of the second image in the high-resolution optical remote sensing image set. S32. Overlay the corner data of the first image's corner set and the corner data of the second image's corner set in geospatial space, and generate a circle with a radius of [missing information - likely a radius value]. The area; S33, Select radius as The corner points of the first building are obtained from the set of corner points of all corner points of the second image within the area; S34. Subtract the first building corner point set from the corner point set of the second image to obtain the second building corner point set; S35, radius When the radius is equal to the minimum radius threshold, repeat steps S32 to S34 to obtain the second set of building corner points minus the radius. When the maximum radius threshold is equal, repeat steps S32 to S34 to obtain the second set of building corner points, and obtain the new corner points.

5. The remote sensing identification method for building disaster-bearing bodies based on corner features according to claim 4, characterized in that, In S33, the first set of building corner points The expression is: ； in, For the corner points in the first set of corner points, The set of corner points of the first image. The set of corner points in the set of corner points of the first image Corner point The radius is The neighborhood of.

6. The remote sensing identification method for building disaster-bearing bodies based on corner features according to claim 4, characterized in that, In S34, the second set of building corner points The expression is: ； in, The set of corner points in the set of corner points of the first image Corner point The radius is The neighborhood, Let this be the set of the first building corner points. For a specific radius value, The radius is [0, 1].

7. The remote sensing identification method for building disaster-bearing bodies based on corner features according to claim 1, characterized in that, In step S4, convex hull polygon generation is used to transform newly added corner points into outer boundaries.

Images