A Method for Extracting Information from Prefabricated Steel Sheds Based on Sentinel Imagery
By combining Sentinel imagery and surveying equipment, the problem of low accuracy in identifying color steel shed roofs was solved, enabling efficient and low-cost spatial information monitoring of color steel sheds.
Patent Information
- Application Number
- CN202211231105.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-10-08
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2042-10-08
AI Technical Summary
Existing technologies cannot effectively identify and extract information about color steel roofs. The identification accuracy is low, the labor intensity and cost are high, and the technology lacks specificity, making it difficult to meet the needs of monitoring spatial information changes in color steel roofs.
Preliminary identification was performed using Sentinel imagery combined with normalized water index, normalized vegetation index, and normalized building index. Spectral and textural differences of the corrugated steel sheds were analyzed using normalized surface index and ground texture features. Accurate extraction was then performed using field surveys and mapping equipment.
The process of identifying and extracting prefabricated steel sheds has been simplified, reducing labor intensity and costs, improving identification accuracy, and enabling high-precision monitoring of prefabricated steel sheds.
Smart Images

Figure CN116129252B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a method for extracting information about color steel sheds based on Sentinel imagery, belonging to the field of remote sensing and mapping technology. Background Technology
[0002] With the sustained growth of the national economy, small and medium-sized cities are developing steadily thanks to national policies, and urbanization and urban-rural integration are progressing smoothly. Building types are undergoing a phased transformation, with corrugated steel roofing materials emerging as a transitional product in urban construction. The continuous development of the construction industry has led to the relocation and renovation of numerous urban villages and small factories in old urban areas, resulting in the elimination of a large number of corrugated steel roofing materials. In order to comply with the relocation policies for urbanization, surrounding urban-rural fringe areas have been rebuilding houses, and due to the low price and ease of installation of corrugated steel roofs, this has become a common practice, leading to a continuous increase in the number and illegal construction of corrugated steel roofs in surrounding villages and towns. Corrugated steel roofs themselves pose safety hazards, and their widespread presence in densely populated urban villages and small factories presents certain risks. Furthermore, the rapid turnover of corrugated steel roofs, coupled with the lack of specific geospatial data statistics, makes it difficult to grasp their spatial information changes and hinders targeted management. Currently, there are few related technologies. A similar published patent is CN202111222651.1, but this technology mainly utilizes TM and high-resolution images, preprocessed, to extract buildings through spectral indices and detect changes in building land use over different times. While this can meet certain needs, it cannot construct specific spectral indices for corrugated steel roofs and combine them with multiple image features for extraction. Therefore, it lacks specificity, has high recognition accuracy, and the labor intensity and cost of the recognition operation are relatively high.
[0003] To address this need, there is an urgent need to develop a novel semi-active suspension control method to effectively meet the requirements of practical use. Summary of the Invention
[0004] The purpose of this invention is to overcome the above-mentioned shortcomings and provide a method for extracting information about color steel sheds based on Sentinel images.
[0005] To achieve the above objectives, the present invention is implemented through the following technical solution:
[0006] A method for extracting information about prefabricated steel sheds based on Sentinel imagery includes the following steps:
[0007] S1. Obtain the urban area and vector boundary of the study area. First, set the urban area to be surveyed, and then download the free and publicly available Sentinel-2 multispectral remote sensing images from different periods through the European Space Agency website.
[0008] S2, Data Preprocessing: The downloaded Sentinel data is preprocessed by performing radiometric calibration and atmospheric correction to obtain accurate reflectance information of ground features; the preprocessed image is mosaicked and cropped according to the existing urban vector boundaries to obtain the remote sensing image of the study area.
[0009] S3, pixel unit calculation, identifies and removes urban water bodies, buildings and greenery in the remote sensing area based on the normalized water index MNDWI, normalized vegetation index NDVI and normalized building index NDBI.
[0010] S4. Spectral index calculation for corrugated steel sheds: Based on the spectral curve characteristics analysis of Sentinel-2 images, the values and trends of corrugated steel sheds in the shortwave infrared 1 and shortwave infrared 2 bands show significant differences compared to other land cover types. Based on this characteristic, a spectral index for corrugated steel sheds is constructed and named the Normalized Surface Index (NDSI). Then, based on the textural differences between corrugated steel sheds and other land cover surfaces, land cover texture features are added to assist in information extraction from corrugated steel sheds. The mean, variance, homogeneity, contrast, dissimilarity, entropy, second moment, and correlation of each band of the Sentinel-2 image are calculated using the gray-level co-occurrence matrix to describe the texture characteristics.
[0011] S5, Physical data collection: Field surveys were conducted to select color steel sheds and typical land cover samples in the city, and the numerical differences in land cover characteristics were analyzed and statistically analyzed, serving as samples for accuracy verification; through multiple experiments, appropriate thresholds for various features were selected to achieve the best classification accuracy, and then other land cover types were removed to finally extract the color steel sheds;
[0012] S6, Post-processing of classification results: The classification results are processed using the main analysis to remove small patches and raster vectorization to determine the area of the color steel shed. Then, by processing multiple remote sensing images, the color steel shed extraction results for multiple time periods are obtained, realizing real-time monitoring of the spatial geographic information of the color steel shed.
[0013] Furthermore, in step S3, the Normalized Difference Water Index (MNDWI) calculation function is:
[0014]
[0015] In the formula ρ green ρ swi r1 represents the green light spectral reflectance and the short-wave infrared spectral reflectance;
[0016] The Normalized Difference Vegetation Index (NDVI) is calculated using the following function:
[0017]
[0018] The Normalized Building Index (NDBI) calculation function is as follows:
[0019]
[0020] In the formula ρ nir ρ red Near-infrared spectral reflectance and red spectral reflectance;
[0021] The normalized surface index (NDSI) calculation function is as follows:
[0022]
[0023] In the formula ρ swir1 ρ swir2 For shortwave infrared 1 spectral reflectance and shortwave infrared 2 spectral reflectance.
[0024] Furthermore, in step S4, when comparing the texture differences between the corrugated steel shed and other ground features, the optimal texture features are selected from all statistics by calculating the coefficient of variation and relative differences, thus avoiding a large amount of data redundancy and improving the accuracy of corrugated steel shed extraction.
[0025] Furthermore, when selecting the optimal texture feature from all statistics using the coefficient of variation and relative difference, the specific comparison function is as follows:
[0026] C·V=(σ 2 / μ)×100%
[0027] RD = ((μ1-μ2) / μ1) × 100%
[0028] In the formula, CV is the coefficient of variation, RD is the relative difference, and μ and σ2 are the mean and variance of the texture features of each locality, respectively.
[0029] Furthermore, in step S5, when collecting physical data, on the one hand, theoretical material data of the physical target is obtained by consulting materials; on the other hand, on-site remote sensing mapping is carried out on the physical object to be sampled using surveying equipment.
[0030] Furthermore, during the physical data collection, the axis of the surveying equipment is aligned with the sunlight and the angle of illumination of the target object at the time of collection.
[0031] Furthermore, the surveying equipment includes a carrier body, a surveying camera, an infrared camera, a 3D camera, a lidar, a carrier tray, a working head, a linear drive rail, a ring drive rail, a mobile weather station, a wireless data communication circuit, a gravity sensor, and a drive circuit. The carrier tray is a rectangular frame structure, with its upper end face connected to the outer surface of the carrier body and forming an angle of 0°–60° with the axis of the carrier body. The ring drive rail is connected to the lower end face of the carrier tray and is coaxially distributed. The linear drive rail is parallel to the lower end face of the carrier tray and is slidably connected to the carrier tray via the ring drive rail. The axis of the linear drive rail is perpendicular to and intersects the axis of the ring drive rail. There are two working heads, distributed on both sides of the center of the ring drive rail and slidably connected to the linear drive rail. The working head is a rectangular cavity structure with a gravity sensor inside. The lower end face of the working head is connected to a mapping camera, an infrared camera, a 3D camera, and a lidar. The axes of the mapping camera, infrared camera, 3D camera, and lidar are all perpendicular to the lower end face of the support tray. The mobile weather station is connected to the upper end face of the support body, and its axis is perpendicular to the axis of the support body. The wireless data communication circuit and the drive circuit are embedded in the support body. The drive circuit is electrically connected to the support body, the mapping camera, the infrared camera, the 3D camera, the lidar, the linear drive rail, the circular drive rail, the mobile weather station, the wireless data communication circuit, and the gravity sensor. The wireless data communication circuit is also electrically connected to the communication circuit of the support body.
[0032] Furthermore, the carrier body can be any one of a rotary-wing UAV, a fixed-wing UAV, and an airship; the wireless data communication circuit is provided with at least one communication antenna, which is connected to the outer surface of the carrier body; and the driving circuit is a circuit system based on an FPGA chip.
[0033] Compared with traditional information extraction and recognition methods, this invention greatly simplifies the workflow of identifying and extracting information from prefabricated steel sheds, reduces the labor intensity and cost of operations, and greatly improves the accuracy of detection and recognition. At the same time, it can effectively eliminate interference from urban surface buildings, vegetation, etc. during the recognition process, thereby effectively improving the accuracy of prefabricated steel shed recognition. Attached Figure Description
[0034] Figure 1 This is a flowchart of the method of the present invention;
[0035] Figure 2 The structural block diagram of a standard H∞ robust control system;
[0036] Figure 3 This is a schematic diagram of a partial structure of the surveying equipment;
[0037] Figure 4 This is a partial structural diagram showing the connection between the load-bearing pallet and the working head. Detailed Implementation
[0038] like Figure 1 and 2 As shown, a method for extracting information about prefabricated steel sheds based on Sentinel imagery includes the following steps:
[0039] S1. Obtain the urban area and vector boundary of the study area. First, set the urban area to be surveyed, and then download the free and publicly available Sentinel-2 multispectral remote sensing images from different periods through the European Space Agency website.
[0040] S2, Data Preprocessing: The downloaded Sentinel data is preprocessed by performing radiometric calibration and atmospheric correction to obtain accurate reflectance information of ground features; the preprocessed image is mosaicked and cropped according to the existing urban vector boundaries to obtain the remote sensing image of the study area.
[0041] S3, pixel unit calculation, identifies and removes urban water bodies, buildings and greenery in the remote sensing area based on the normalized water index MNDWI, normalized vegetation index NDVI and normalized building index NDBI.
[0042] S4. Spectral index calculation for corrugated steel sheds: Based on the spectral curve characteristics analysis of Sentinel-2 images, the values and trends of corrugated steel sheds in the shortwave infrared 1 and shortwave infrared 2 bands show significant differences compared to other land cover types. Based on this characteristic, a spectral index for corrugated steel sheds is constructed and named the Normalized Surface Index (NDSI). Then, based on the textural differences between corrugated steel sheds and other land cover surfaces, land cover texture features are added to assist in information extraction from corrugated steel sheds. The mean, variance, homogeneity, contrast, dissimilarity, entropy, second moment, and correlation of each band of the Sentinel-2 image are calculated using the gray-level co-occurrence matrix to describe the texture characteristics.
[0043] S5, Physical data collection: Field surveys were conducted to select color steel sheds and typical land cover samples in the city, and the numerical differences in land cover characteristics were analyzed and statistically analyzed, serving as samples for accuracy verification; through multiple experiments, appropriate thresholds for various features were selected to achieve the best classification accuracy, and then other land cover types were removed to finally extract the color steel sheds;
[0044] S6, Post-processing of classification results: The classification results are processed using the main analysis to remove small patches and raster vectorization to determine the area of the color steel shed. Then, by processing multiple remote sensing images, the color steel shed extraction results for multiple time periods are obtained, realizing real-time monitoring of the spatial geographic information of the color steel shed.
[0045] It is important to note that in step S3, the normalized water index (MNDWI) is calculated using the following function:
[0046]
[0047] In the formula ρ green ρ swir1 For green light spectral reflectance and short-wave infrared spectral reflectance;
[0048] The Normalized Difference Vegetation Index (NDVI) is calculated using the following function:
[0049]
[0050] The Normalized Building Index (NDBI) calculation function is as follows:
[0051]
[0052] In the formula ρ nir ρ red Near-infrared spectral reflectance and red spectral reflectance;
[0053] The normalized surface index (NDSI) calculation function is as follows:
[0054]
[0055] In the formula ρ swir1 ρ swir2 For shortwave infrared 1 spectral reflectance and shortwave infrared 2 spectral reflectance.
[0056] Meanwhile, in step S4, when comparing the texture differences between the corrugated steel shed and other ground features, the optimal texture features are selected from all statistics by calculating the coefficient of variation and relative differences, avoiding a large amount of data redundancy and improving the accuracy of corrugated steel shed extraction.
[0057] Furthermore, when selecting the optimal texture feature from all statistics using the coefficient of variation and relative difference, the specific comparison function is as follows:
[0058] C·V=(σ 2 / μ)×100%
[0059] RD = ((μ1-μ2) / μ1) × 100%
[0060] In the formula, CV is the coefficient of variation, RD is the relative difference, and μ and σ2 are the mean and variance of the texture features of each locality, respectively.
[0061] In this embodiment, during the physical data collection process in step S5, theoretical material data of the physical target is obtained by consulting relevant materials; on the other hand, on-site remote sensing mapping of the sampled object is carried out using surveying equipment.
[0062] In a further optimization, when the physical data is collected, the axis of the surveying equipment is aligned with the angle of sunlight illuminating the target object during the collection time.
[0063] Specifically, the surveying equipment includes a carrier body 1, a surveying camera 2, an infrared camera 3, a 3D camera 4, a lidar 5, a carrier tray 6, a working head 7, a linear drive rail 8, a ring drive rail 9, a mobile weather station 10, a wireless data communication circuit 11, a gravity sensor 12, and a drive circuit 13. The carrier tray 6 is a rectangular frame structure, with its upper end face connected to the outer surface of the carrier body 1 and forming an angle of 0°–60° with the axis of the carrier body 1. The ring drive rail 9 is connected to the lower end face of the carrier tray 6 and is coaxially distributed. The linear drive rail 8 is parallel to the lower end face of the carrier tray 6 and is slidably connected to the carrier tray 6 through the ring drive rail 9. The axis of the linear drive rail 8 is perpendicular to and intersects the axis of the ring drive rail 9. There are two working heads 7, distributed on both sides of the center of the ring drive rail 9 and slidably connected to the linear drive rail 9. The working head 7 is a rectangular cavity structure with a gravity sensor 12 inside. The lower end face of the working head 7 is connected to the surveying camera 2, infrared camera 3, 3D camera 4, and lidar 5, respectively. The axes of the surveying camera 2, infrared camera 3, 3D camera 4, and lidar 5 are all perpendicular to the lower end face of the support tray 6. The mobile weather station 10 is connected to the upper end face of the support body 1, and its axis is perpendicular to the axis of the support body 1. The wireless data communication circuit 11 and the drive circuit 13 are embedded in the support body 1. The drive circuit 13 is electrically connected to the support body 1, the surveying camera 2, infrared camera 3, 3D camera 4, lidar 5, linear drive rail 8, ring drive rail 9, mobile weather station 10, wireless data communication circuit 11, and gravity sensor 12, respectively. The wireless data communication circuit 11 is also electrically connected to the communication circuit of the support body 1.
[0064] In actual surveying, two working heads are used to simultaneously survey the same target object. The working positions of the two working heads are adjusted via linear and circular drive rails, allowing for simultaneous surveying of the same target from two different locations while simultaneously satisfying the needs of surveying targets with different structural positions. This enables correction of target object parameter acquisition, improving the accuracy of parameter acquisition. Furthermore, a gravity sensor can accurately locate the target object's parameter acquisition status during the surveying operation. Simultaneously, a mobile weather station collects air quality, ambient brightness, and meteorological conditions during target object parameter acquisition. The collected data, along with the status and environmental parameter information during the acquisition operation, are then incorporated into steps S5 and S6, improving the accuracy of the data acquisition and identification operation. This also serves as a reference and correction for the accurate ground reflectance information obtained after radiometric calibration and atmospheric correction in step S2, further improving the accuracy of data processing.
[0065] In this embodiment, the carrier body 1 is any one of a rotary-wing UAV, a fixed-wing UAV, and an airship; the wireless data communication circuit 11 is provided with at least one communication antenna 14, which is connected to the outer surface of the carrier body; and the driving circuit 13 is a circuit system based on an FPGA chip.
[0066] Compared with traditional information extraction and recognition methods, this invention greatly simplifies the workflow of identifying and extracting information from prefabricated steel sheds, reduces the labor intensity and cost of operations, and greatly improves the accuracy of detection and recognition. At the same time, it can effectively eliminate interference from urban surface buildings, vegetation, etc. during the recognition process, thereby effectively improving the accuracy of prefabricated steel shed recognition.
[0067] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.
Claims
1. A method for extracting information about color steel sheds based on Sentinel images, characterized in that: The method for extracting information about prefabricated steel sheds based on Sentinel images includes the following steps: S1. Obtain the urban area and vector boundary of the study area. First, set the urban area to be surveyed, and then download free and publicly available Sentinel-2 multispectral remote sensing images from different periods through the website. S2, Data Preprocessing: The downloaded Sentinel data is preprocessed by performing radiometric calibration and atmospheric correction to obtain accurate reflectance information of ground features; the preprocessed image is mosaicked and cropped according to the existing urban vector boundaries to obtain the remote sensing image of the study area. S3, pixel unit calculation, identifies and removes urban water bodies, buildings and greenery in the remote sensing area based on the normalized water index MNDWI, normalized vegetation index NDVI and normalized building index NDBI. S4. Spectral index calculation for corrugated steel sheds: Based on the spectral curve characteristics analysis of Sentinel-2 land cover, the values and trends of corrugated steel sheds in the shortwave infrared 1 and shortwave infrared 2 bands show significant differences from other land cover types. Based on this characteristic, a spectral index for corrugated steel sheds is constructed and named the Normalized Surface Index (NDSI). Then, based on the texture differences between corrugated steel sheds and other land cover surfaces, land cover texture features are added to assist in the information extraction of corrugated steel sheds. The mean, variance, homogeneity, contrast, dissimilarity, entropy, second moment, and correlation of each band of the Sentinel-2 image are calculated using the gray-level co-occurrence matrix to describe the texture features. S5, Physical data collection: Field surveys were conducted to select color steel sheds and typical land cover samples in the city, and the numerical differences in land cover characteristics were analyzed and statistically analyzed, serving as samples for accuracy verification; through multiple experiments, appropriate thresholds for various features were selected to achieve the best classification accuracy, and then other land cover types were removed to finally extract the color steel sheds; S6, Post-processing of classification results: The classification results are processed using the main analysis to remove small patches and raster vectorization to determine the area of the color steel shed. Then, by processing multiple remote sensing images, the color steel shed extraction results for multiple time periods are obtained, realizing real-time monitoring of the spatial geographic information of the color steel shed.
2. The method for extracting information about color steel sheds based on Sentinel images according to claim 1, characterized in that: In step S3, the Normalized Difference Water Index (MNDWI) calculation function is: In the formula For green light spectral reflectance and short-wave infrared spectral reflectance; The Normalized Difference Vegetation Index (NDVI) is calculated using the following function: The Normalized Building Index (NDBI) calculation function is as follows: In the formula , representing near-infrared spectral reflectance and red spectral reflectance; The normalized surface index (NDSI) calculation function is as follows: In the formula For shortwave infrared 1 spectral reflectance and shortwave infrared 2 spectral reflectance.
3. The method for extracting information about color steel sheds based on Sentinel images according to claim 1, characterized in that: In step S4, when comparing the texture differences between the corrugated steel shed and other ground features, the optimal texture features are selected from all statistics by calculating the coefficient of variation and relative difference, thus avoiding a large amount of data redundancy and improving the accuracy of corrugated steel shed extraction.
4. A method for extracting information about color steel sheds based on Sentinel images according to claim 1 or 3, characterized in that: When selecting the optimal texture feature from all statistics using the coefficient of variation and relative difference, the specific comparison function is as follows: In the formula Here, denoted as the coefficient of variation, and RD as the relative difference. These represent the mean and variance of the texture features of each locality.
5. The method for extracting information about color steel sheds based on Sentinel images according to claim 1, characterized in that: In step S5, when collecting physical data, on the one hand, theoretical material data of the physical target is obtained by consulting literature; on the other hand, on-site remote sensing mapping is carried out on the physical object to be sampled using surveying equipment.
6. A method for extracting information about color steel sheds based on Sentinel images according to claim 1 or 5, characterized in that: The aforementioned data collection process involves ensuring that the axis of the surveying equipment is aligned with the sunlight and the angle of illumination of the target object during the data collection time.
7. The method for extracting information about color steel sheds based on Sentinel images according to claim 5, characterized in that: The surveying equipment includes a carrier body, a surveying camera, an infrared camera, a 3D camera, a lidar, a carrier tray, a working head, a linear drive rail, a circular drive rail, a mobile weather station, a wireless data communication circuit, a gravity sensor, and a drive circuit. The carrier tray is a rectangular frame structure, with its upper surface connected to the outer surface of the carrier body and forming an angle of 0°–60° with the carrier body's axis. The circular drive rail is connected to the lower surface of the carrier tray and coaxially distributed. The linear drive rail is parallel to the lower surface of the carrier tray and slidably connected to the carrier tray via the circular drive rail. The axis of the linear drive rail is perpendicular to and intersects the axis of the circular drive rail. Two working heads are located on either side of the center of the circular drive rail and slidably connected to the linear drive rail. The working head is a rectangular cavity structure with a gravity sensor inside. The lower end face of the working head is connected to a mapping camera, an infrared camera, a 3D camera, and a lidar. The axes of the mapping camera, infrared camera, 3D camera, and lidar are all perpendicular to the lower end face of the support tray. The mobile weather station is connected to the upper end face of the support body, and its axis is perpendicular to the axis of the support body. The wireless data communication circuit and the drive circuit are embedded in the support body. The drive circuit is electrically connected to the support body, the mapping camera, the infrared camera, the 3D camera, the lidar, the linear drive rail, the circular drive rail, the mobile weather station, the wireless data communication circuit, and the gravity sensor. The wireless data communication circuit is also electrically connected to the communication circuit of the support body.
8. The method for extracting information about color steel sheds based on Sentinel images according to claim 7, characterized in that: The carrier body can be any one of a rotary-wing UAV, a fixed-wing UAV, and an airship; the wireless data communication circuit is provided with at least one communication antenna, which is connected to the outer surface of the carrier body; and the driving circuit is a circuit system based on an FPGA chip.
Citation Information
Patent Citations
Building extraction method based on TM-high-resolution remote sensing image fusion
CN113989640A
Method to Correct Satellite Data to Surface Reflectance Using Scene Statistics
US20210142447A1
Method and apparatus for extracting mountain landscape buildings based on high-resolution remote sensing images
US20210166020A1
