A method for detecting the height of a residential building

Abstract

This invention relates to the field of building height determination technology, specifically to a method for detecting the height of residential buildings. The method includes: acquiring multi-view remote sensing images of residential buildings via satellite, wherein the multi-view remote sensing images contain images of the residential building to be detected and a set of residential buildings with known heights, the set of residential buildings with known heights including multiple residential buildings of known heights; and determining the height of the residential building to be detected based on the multi-view remote sensing images. Therefore, this invention solves the technical problem of low accuracy in determining the height of residential buildings, improves the accuracy of residential building height determination, and is mainly applied to the determination of residential building heights.

Description

Technical Field

[0001] This invention relates to the field of building height determination technology, and specifically to a method for detecting the height of residential buildings. Background Technology

[0002] Building height is a crucial parameter in urban environmental analysis and an essential component of urban microclimate modeling. Rapidly and accurately acquiring building height information is of great significance for various important applications such as environmental management and urban planning. ZY-3 is China's first civilian high-resolution optical transmission mapping satellite. Its onboard cameras have a ground pixel resolution of 2.1 meters for the front-view camera, 3.5 meters for the forward-looking and backward-looking cameras, and 5.8 meters for the multispectral camera. Traditional algorithms that simply extract building heights from single ZY-3 images neglect the potential for inverting building heights based on three-view images from ZY-3, often resulting in low accuracy in determining building heights.

[0003] To address this, this invention proposes a novel algorithm for inverting the height of residential buildings based on three-view imagery from the ZY-3 satellite and incorporating the mathematical concept of Lie group matrices. This correction model avoids the limitations of inverting building heights from a single image. The algorithm is applicable to establishing building height inversion models from three-view imagery, enabling the multi-sensor ZY-3 satellite to realize greater potential in building height inversion. Summary of the Invention

[0004] The summary section of this invention provides a brief overview of the concepts, which will be described in detail in the detailed description section that follows. This summary section is not intended to identify key or essential features of the claimed invention, nor is it intended to limit the scope of the claimed invention.

[0005] To address the technical problem of low accuracy in determining building height, this invention proposes a method for detecting the height of residential buildings.

[0006] This invention provides a method for detecting the height of residential buildings, the method comprising:

[0007] The satellite acquires multi-view remote sensing images of residential buildings. The multi-view remote sensing images capture images of the residential buildings to be detected and a set of residential buildings of known height. The set of residential buildings of known height includes multiple residential buildings of known height.

[0008] The height of the residential building to be inspected is determined based on multi-view remote sensing images of the building.

[0009] Furthermore, determining the height of the residential building to be detected based on the multi-view remote sensing image of the residential building includes:

[0010] Image preprocessing is performed on multi-view remote sensing images to obtain back-view images, front-view images, and front-view images;

[0011] Building shadows were extracted from the rear view, front view, and frontal view images, respectively.

[0012] Determine the shadow fishing net line based on the building's shadow;

[0013] Determine the length of the building's shadow based on the shadow fishing net lines;

[0014] Based on the building shadow length, the number of residential buildings of known height in the set of residential buildings of known height, and the height of the residential buildings of known height, determine the backsight coefficient, frontsight coefficient, and frontal coefficient;

[0015] The building inversion height is determined based on the backsight coefficient, the foresight coefficient, and the frontal coefficient. The building inversion height is the building inversion height in the backsight, the building inversion height in the foresight, or the building inversion height in the frontal view.

[0016] Determine the model matrix based on the building's inverted height;

[0017] The height of the residential building to be detected is determined based on the model matrix.

[0018] Furthermore, the image preprocessing of the multi-view remote sensing images to obtain rear-view images, front-view images, and orthographic images includes:

[0019] Preprocessing of multi-view remote sensing images;

[0020] The back view, front view, and frontal view images are extracted from the preprocessed multi-view remote sensing images.

[0021] Furthermore, determining the shadow fishing line based on the building's shadow includes:

[0022] Based on the sun's azimuth angle during satellite imagery, a fishing net is created by simulating sunlight and intersecting it with the extracted building shadows to obtain the shadow fishing net lines.

[0023] Furthermore, determining the building's shadow length based on the shadow fishing net line includes:

[0024] The box plot method was used to remove outliers within the building shadows in the shaded fishing net lines;

[0025] The average value of the shadow fishing net line is used to determine the length of the building's shadow.

[0026] Furthermore, the formula for determining the backsight coefficient is as follows:

[0027]

[0028] Where, μ b This is the backsight coefficient, where n represents the number of residential buildings of known height in the set of residential buildings of known height, i represents the index of the residential building of known height in the set of residential buildings of known height, and S... b,i Let h be the length of the building shadow of the i-th residential building of known height in the set of residential buildings of known height in the rear view image. b,i It is the height of the i-th residential building with a known height in the set of residential buildings with known heights.

[0029] Furthermore, the formula for determining the inversion height of the building in the back view is as follows:

[0030]

[0031] Among them, H b,j It is the building inversion height of the j-th residential building in the multi-view remote sensing image, where j is the building number in the multi-view remote sensing image, μ b It is the rearview coefficient, S b,j It is the length of the building shadow in the rear view image of the j-th residential building in the multi-view remote sensing image.

[0032] The present invention has the following beneficial effects:

[0033] This invention proposes a novel algorithm for retrieving the height of residential buildings based on three-view imagery from the ZY-3 satellite and incorporating the mathematical concept of Lie group matrices. This correction model avoids the limitations of retrieving building heights from a single image, improving the accuracy of building height determination. The algorithm is applicable to establishing building height retrieval models from three-view imagery, enabling the multi-sensor ZY-3 satellite to realize greater potential in building height retrieval. Attached Figure Description

[0034] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 A flowchart of a method for detecting the height of a residential building according to the present invention;

[0036] Figure 2 This is a detailed framework diagram according to the present invention. Detailed Implementation

[0037] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the specific implementation methods, structures, features, and effects of the technical solution proposed according to the present invention are described in detail below with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0038] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0039] This invention provides a method for detecting the height of residential buildings, which includes the following steps:

[0040] Multi-view remote sensing images of residential buildings are acquired via satellite;

[0041] The height of the residential building to be inspected is determined based on multi-view remote sensing images of the building.

[0042] The following is a detailed explanation of each of the above steps:

[0043] refer to Figure 1 The flowchart illustrates some embodiments of a method for detecting the height of a residential building according to the present invention. The method for detecting the height of a residential building includes the following steps:

[0044] Step S1: Obtain multi-view remote sensing images of residential buildings via satellite.

[0045] In some embodiments, multi-view remote sensing images of residential buildings can be acquired via satellite.

[0046] Multi-view remote sensing imagery can include: front-view imagery, forward-view imagery, and back-view imagery. Multi-view remote sensing imagery can capture images of residential buildings to be detected and sets of residential buildings of known height. The set of residential buildings of known height includes multiple residential buildings of known height. Residential buildings of known height can be those whose actual height is known. Residential buildings to be detected can be those with a height to be detected. Residential buildings with a height to be detected can be those with an unknown height.

[0047] In practice, the more residential buildings with known heights in the set of residential buildings, the more accurate the height of the residential building to be tested will be.

[0048] Step S2: Determine the height of the residential building to be detected based on the multi-view remote sensing image of the residential building.

[0049] In some embodiments, the height of the residential building to be detected can be determined based on multi-view remote sensing images of the residential building.

[0050] As an example, this step may include the following steps:

[0051] The first step is to perform image preprocessing on the multi-view remote sensing images to obtain back-view images, front-view images, and front-view images.

[0052] For example, this step may include the following sub-steps:

[0053] The first sub-step involves preprocessing the multi-view remote sensing images.

[0054] For example, multi-view remote sensing images can be preprocessed by following the steps of geometric fine correction and image registration, image fusion, cropping, and atmospheric correction.

[0055] The second sub-step involves extracting the back-view image, front-view image, and frontal view image from the preprocessed multi-view remote sensing image.

[0056] The second step is to extract the building shadows from the rear view, front view, and frontal view images respectively.

[0057] For example, the shadows of buildings in the frontal, forward-looking, and back-looking images can be extracted from the preprocessed multi-view remote sensing images using manual interpretation methods.

[0058] The third step is to determine the shadow fishing net line based on the building's shadow.

[0059] For example, based on the sun's azimuth angle during satellite imagery, a fishing net can be created by simulating sunlight and intersecting it with extracted building shadows to obtain the shadow fishing net lines. The extracted building shadows can be from front-view, forward-view, and back-view images.

[0060] The fourth step is to determine the length of the building's shadow based on the shadow fishing net lines.

[0061] For example, this step may include the following sub-steps:

[0062] The first sub-step involves using a box plot to remove outliers within the building shadows in the shaded fishing net lines.

[0063] The second sub-step is to determine the average value of the shadow fishing net lines as the length of the building's shadow.

[0064] In summary, based on the solar azimuth angle captured by satellite, a fishing net can be created to simulate sunlight and intersect with the extracted shadow surface. Outliers within each building's shadow surface are then removed using a box plot method. The average value of the fishing net lines on the building's shadow surface can be taken as the length of the building's shadow surface. This building shadow length can be the shadow length of the building in the rear view image, the front view image, or the forward view image.

[0065] The fifth step is to determine the backsight coefficient, frontsight coefficient, and frontal sight coefficient based on the building shadow length, the number of residential buildings with known heights in the set of residential buildings with known heights, and the height of the residential buildings with known heights.

[0066] For example, the formula for determining the backsight coefficient can be:

[0067]

[0068] Where, μ b This is the backsight coefficient (backsight ratio coefficient), where n represents the number of residential buildings of known height in the set of residential buildings of known height (the number of sample points with known building heights selected), i represents the index of the residential building of known height in the set of residential buildings of known height, and S... b,i Let h be the length of the building shadow of the i-th residential building of known height in the set of residential buildings of known height (the length of the shadow surface of the building in the back view of the selected sample point), h. b,i It is the height of the i-th residential building with known height in the set of residential buildings (the actual height of the building corresponding to the shaded surface of the building when viewed from the front of the sample point).

[0069] For example, the formula for determining the backsight coefficient can be:

[0070]

[0071] Where, μ b It is the backsight coefficient, where X represents the number of residential buildings of different heights in the set of residential buildings of known height, x represents the serial number of the residential buildings of different heights in the set of residential buildings of known height, and S... b,x Let h be the length of the building shadow of the x-th residential building of known height in the set of residential buildings of known height in the rear view image. b,x It is the height of the x-th residential building with a known height in the set of residential buildings with known heights.

[0072] When determining the forward and frontal coefficients, you can refer to the calculation method for determining the backsight coefficient.

[0073] The sixth step is to determine the building's inversion height based on the backsight coefficient, frontsight coefficient, and frontal coefficient.

[0074] Among them, the building inversion height is the building inversion height from the back view, the building inversion height from the front view, or the building inversion height from the front view.

[0075] For example, the formula for determining the inversion height of a building in the back view can be:

[0076]

[0077] Among them, H b,j It is the building inversion height of the j-th residential building in the multi-view remote sensing image, where j is the building number in the multi-view remote sensing image, μ b It is the rearview coefficient, S b,j It is the length of the building shadow in the rear view image of the j-th residential building in the multi-view remote sensing image.

[0078] The estimated height of other buildings can be calculated using the backsight coefficient (building inversion height).

[0079] When determining the inversion height of a building in front view or the building in front of you, you can refer to the calculation method for determining the inversion height of a building in back view.

[0080] Step 7: Determine the model matrix based on the building's inverted height.

[0081] The model matrix is ​​the building inversion height matrix. The elements of the building inversion height matrix are the building inversion heights.

[0082] In the acquired multi-view remote sensing images, the frontal view, forward view, and back view images were all taken by different sensors on the same satellite at the same time and from different angles. The relationship between the three is expressed and distinguished by the rows and columns in the definition of the Lie group matrix, thereby establishing the building height inversion matrix.

[0083] Step 8: Determine the height of the residential building to be detected based on the model matrix.

[0084] The three-view and multi-view remote sensing images provided by the ZY-3 satellite are images taken by the satellite at the same time from different angles. Traditional matrix thinking is used to construct matrices for buildings in these three-view images, which requires considering factors such as imaging angles. However, a Lie group is defined as a smooth manifold, which can be understood as a smooth, curved, and nonlinear high-dimensional space. Therefore, we can treat the buildings in the frontal, forward-view, and back-view images as vectors of different dimensions, considering them as a unified Lie group matrix, and based on this, establish the following formula:

[0085] Hn×1 =[R] n×l ×[γ] l×1 +ε

[0086] Where n represents the number of sample residential buildings in the sample residential building set, H is the actual building height, R is the n×l building inversion height matrix, γ is the l×1 model coefficient, ε is the correction coefficient, and H n×1 With [R] n×l These are matrices constructed from the corresponding data of the selected samples. The selected samples can be residential buildings of known height in multi-view remote sensing images. Among the building inversion heights obtained from the frontal, forward-looking, and back-looking images using the scaling factor method mentioned above, the sample residential buildings are those whose heights are closer to the actual heights of the residential buildings in all three images.

[0087] The matrix correction model described above can be simply viewed as a Fourier transform, and theoretically, by using its inverse process, the coefficients γ and ε of the model can be obtained by substituting data that meet the conditions.

[0088] For example, given the known height and the inverted building height, the data of residential buildings with known heights selected from multi-view remote sensing images are those whose inverted building heights, obtained through the scaling factor method mentioned above, are closest to the actual height of the residential buildings in the frontal, forward-looking, and back-looking images. These data are then substituted into the Lie group matrix operation H. n×1 =[R] n×l ×[γ] l×1 The following system of linear equations can be obtained from +ε:

[0089] a*N1+b*B1+c*F1+§=H1

[0090] a*N2+b*B2+c*F2+§=H2

[0091] a*N3+b*B3+c*F3+§=H3

[0092] a*N4+b*B4+c*F4+§=H4

[0093] Where a, b, and c are [γ] l×1 The model coefficients in H. i For H n×1 The matrix elements in N. i F i B i For [R] n×l The matrix elements are represented by i = 1, 2, 3…n. n represents the number of sample residential buildings in the sample residential building set. § is the correction coefficient ε. The sample residential building set consists of each individual sample residential building.

[0094] The final inversion height of the residential building to be tested is the height of the residential building to be tested.

[0095] It is important to note that the amount of data selected for substitution must satisfy the condition that the linear algebraic equation has a unique solution, and the error should be as small as possible. In order to obtain a data amount with smaller errors for substitution calculation, the box plot method mentioned above for eliminating fishing net lines with large errors was used to filter the estimated height data of different buildings, and the building height data with smaller errors was substituted into the calculation.

[0096] A detailed framework diagram of this invention can be seen as follows: Figure 2 As shown.

[0097] This invention proposes a novel algorithm for inverting the height of residential buildings based on three-view imagery from the ZY-3 satellite and incorporating the mathematical concept of Lie group matrices. This correction model avoids the limitations of inverting building heights from a single image. The algorithm is applicable to establishing building height inversion models from three-view imagery, enabling the multi-sensor ZY-3 satellite to realize greater potential in building height inversion.

[0098] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for detecting the height of residential buildings, characterized in that, Includes the following steps: The satellite acquires multi-view remote sensing images of residential buildings. The multi-view remote sensing images capture images of the residential buildings to be detected and a set of residential buildings of known height. The set of residential buildings of known height includes multiple residential buildings of known height. The height of the residential building to be inspected is determined based on multi-view remote sensing images of the building. The step of determining the height of the residential building to be detected based on multi-view remote sensing images of residential buildings includes: Image preprocessing is performed on multi-view remote sensing images to obtain back-view images, front-view images, and front-view images; Building shadows were extracted from the rear view, front view, and frontal view images, respectively. Determine the shadow fishing net line based on the building's shadow; Determine the length of the building's shadow based on the shadow fishing net lines; Based on the building shadow length, the number of residential buildings of known height in the set of residential buildings of known height, and the height of the residential buildings of known height, determine the backsight coefficient, frontsight coefficient, and frontal coefficient; The building inversion height is determined based on the backsight coefficient, the foresight coefficient, and the frontal coefficient. The building inversion height is the building inversion height in the backsight, the building inversion height in the foresight, or the building inversion height in the frontal view. Determine the model matrix based on the building's inverted height; Based on the model matrix, the height of the residential building to be detected is determined. The model matrix is ​​the building inversion height matrix, and the elements of the building inversion height matrix are the building inversion heights. In the acquired multi-view remote sensing images, the frontal view, forward view, and back view images are all taken by different sensors on the same satellite at the same time and from different angles. The relationship between the three is expressed and distinguished by the rows and columns in the definition of the Lie group matrix, thereby establishing the building height inversion matrix.

2. The method for detecting the height of a residential building according to claim 1, characterized in that, The image preprocessing of multi-view remote sensing images to obtain rear-view, front-view, and orthographic images includes: Preprocessing of multi-view remote sensing images; The back view, front view, and frontal view images are extracted from the preprocessed multi-view remote sensing images.

3. The method for detecting the height of a residential building according to claim 1, characterized in that, The method of determining the shadow fishing net line based on the building shadow includes: Based on the sun's azimuth angle during satellite imagery, a fishing net is created by simulating sunlight and intersecting it with the extracted building shadows to obtain the shadow fishing net lines.

4. The method for detecting the height of a residential building according to claim 1, characterized in that, The determination of the building's shadow length based on the shadow fishing net line includes: The box plot method was used to remove outliers within the building shadows in the shaded fishing net lines; The average value of the shadow fishing net line is used to determine the length of the building's shadow.

5. The method for detecting the height of a residential building according to claim 1, characterized in that, The formula for determining the backsight coefficient is: in, This is the backsight coefficient, where n represents the number of residential buildings of known height in the set of residential buildings of known height, and i represents the index of the residential buildings of known height in the set of residential buildings of known height. Let be the length of the building shadow in the rear view image of the i-th residential building of known height in the set of residential buildings of known height. It is the height of the i-th residential building with a known height in the set of residential buildings with known heights.

6. The method for detecting the height of a residential building according to claim 1, characterized in that, The formula for determining the inversion height of a building in the back view is: in, It is the building inversion height of the j-th residential building in the multi-view remote sensing image, where j is the building number in the multi-view remote sensing image. It is the rearview coefficient. It is the length of the building shadow in the rear view image of the j-th residential building in the multi-view remote sensing image.

Images