Method and device for estimating photovoltaic potential of classified buildings based on remote sensing images

Abstract

The application provides a kind of based on remote sensing image's classification building roof photovoltaic potential estimation method and device, belongs to artificial intelligence field, including: building classification system for photovoltaic potential evaluation is constructed, for obtaining building classification information on remote sensing image, classification system specifically is: building is divided into color steel tile roof industrial and commercial building, concrete roof industrial and commercial building, urban residential house, rural residential house four categories;Four kinds of building corresponding building samples are made;Based on building sample, fine-tune instance segmentation pre-training model extracts the total area of four kinds of building roofs in the region to be measured;According to the total roof area of each type of building and the roof area to be developed, the total photovoltaic potential and the photovoltaic potential to be developed of each type of roof are evaluated, wherein the utilization coefficient is obtained from the roof with installed photovoltaic panels.

Description

Technical Field

[0001] This invention belongs to the field of artificial intelligence, and in particular relates to a method and apparatus for estimating the potential of rooftop photovoltaics of classified buildings based on remote sensing images. Background Technology

[0002] Against the backdrop of the continued development of the distributed photovoltaic industry, the rational planning and utilization of building rooftops has become particularly important for assessing the potential resources for photovoltaic power generation. Currently, there are a large number of building rooftops suitable for installing photovoltaic systems but which are not yet utilized, and the photovoltaic potential of these rooftops is closely related to the type of building.

[0003] Building data plays a central role in assessing the potential of rooftop photovoltaics. This data primarily falls into three categories: first, total building area data for the entire region; second, building area data divided by geographic grids, which includes not only area information but also spatial distribution information; and finally, building footprint data, which is more detailed, containing information such as the specific location, boundaries, and area of ​​individual buildings. The first two types of data typically originate from official statistics or are extrapolated from other multi-source data. However, obtaining large-scale building footprint data is challenging. Therefore, extracting information from remote sensing imagery becomes an effective technical approach, with semantic segmentation and instance segmentation being two main deep learning methods.

[0004] Existing technical methods mainly fall into two categories. The first category focuses on assessing the photovoltaic potential of building rooftops. These studies suffer from several problems: First, most studies fail to consider the diversity of building types. Different building types exhibit significant differences in the suitability and necessity of photovoltaic installation. Treating all building rooftops as equally important photovoltaic carriers may lead to overestimation of the assessment results. Second, while some studies consider building function types during assessment, these studies often rely on auxiliary data such as points of interest (POI), areas of interest (AOI), or urban functional zoning, or the classification granularity is coarse. This auxiliary data may suffer from ambiguity, incomplete coverage, and outdated updates. The classification information is primarily used to statistically analyze the potential of different building types, rather than for specific photovoltaic resource potential calculations for different building types. Finally, although some studies have designed corresponding availability coefficients for different building types in specific regions, these coefficients are mostly based on theoretical analysis, considering local development plans and building distribution, and therefore are not suitable for generalization to other regions.

[0005] The second type of research focuses on technical methods for building classification. These methods typically build upon existing building vector data, utilizing multimodal data to calculate various features for individual buildings, and then using machine learning models such as random forests and XGBoost to identify building functions. Problems with this type of approach include: First, outdated or inaccurate building vector data may reduce the accuracy of building function identification. Second, classification criteria may not be suitable for assessing rooftop photovoltaic potential, as photovoltaic potential assessment requires consideration not only of the building's functional type but also its morphological characteristics. Summary of the Invention

[0006] To address the above technical problems, this invention provides a method and apparatus for estimating the rooftop photovoltaic potential of classified buildings based on remote sensing imagery. First, it establishes building classification standards to meet the needs of rooftop photovoltaic potential assessment. Then, based on remote sensing imagery, it extracts and obtains the availability coefficients of different building types, thereby enabling targeted estimation of rooftop photovoltaic potential for different building types and providing more scientific data support for rooftop photovoltaic development planning. The specific technical solution is as follows:

[0007] A method for estimating the rooftop photovoltaic potential of classified buildings based on remote sensing imagery includes the following steps:

[0008] Step 1: Construct a building classification system for photovoltaic potential assessment to obtain building classification information from remote sensing images. The classification system is as follows: buildings are divided into four categories: industrial and commercial buildings with corrugated steel roofs, industrial and commercial buildings with concrete roofs, urban residential buildings, and rural residential buildings.

[0009] Step 2: Create corresponding building samples for the four types of buildings;

[0010] Step 3: Based on building samples, fine-tune the instance segmentation pre-trained model to extract the total roof area of ​​the four types of buildings in the test area;

[0011] Step 4: Based on the total roof area of ​​each type of building, calculate the developed roof area and the undeveloped roof area, and assess the total photovoltaic potential and undeveloped photovoltaic potential of each type of roof.

[0012] A device for estimating the potential of rooftop photovoltaic power in classified buildings based on remote sensing imagery, comprising:

[0013] The building classification module constructs a building classification system for photovoltaic potential assessment, which is used to obtain building classification information from remote sensing images. The classification system is specifically divided into four categories: industrial and commercial buildings with corrugated steel roofs, industrial and commercial buildings with concrete roofs, urban residential buildings, and rural residential buildings.

[0014] The building extraction module generates corresponding building samples for four types of buildings.

[0015] The building roof area calculation module, based on the corresponding building samples of four types of buildings, fine-tunes the instance segmentation pre-trained model to extract the total roof area of ​​the four types of buildings in the test area;

[0016] The photovoltaic potential calculation module calculates the developed roof area and the undeveloped roof area based on the total roof area of ​​various types of buildings, and evaluates the total photovoltaic potential and undeveloped photovoltaic potential of each type of roof.

[0017] An electronic device includes: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement the method.

[0018] A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, cause the processor to implement the method described thereon.

[0019] The present invention has the following beneficial effects:

[0020] This invention constructs a building classification system for photovoltaic potential assessment, which can more accurately identify and classify different types of buildings, thereby improving the accuracy of photovoltaic potential assessment. This classification system considers not only the functional characteristics of buildings but also their morphological characteristics, making the assessment results more consistent with reality.

[0021] The method of this invention reduces reliance on auxiliary classification data such as POI, AOI, and urban functional zoning. By fully utilizing the differences in image features on remote sensing images, such as shape, size, color, texture, and shadow, the building classification information required for photovoltaic potential assessment can be directly obtained from remote sensing images, which greatly reduces the complexity of data collection and processing.

[0022] This invention utilizes the advantages of remote sensing imagery, such as wide coverage, rapid updates, and high quality, to enable photovoltaic potential assessment to be conducted more quickly and extensively, especially in large-scale assessments, where this advantage is even more pronounced.

[0023] This invention provides more scientific data support for rooftop photovoltaic development planning. By accurately assessing the rooftop photovoltaic potential of different types of buildings, it can provide strong data support for the planning and implementation of photovoltaic projects, thereby improving the success rate and economic benefits of the projects.

[0024] This invention obtains the area conversion factor from roofs with installed photovoltaic panels, avoiding the limitations of factors derived from theories or experience. This method is more realistic and can more accurately reflect the actual photovoltaic installation potential of roofs of different types of buildings. Attached Figure Description

[0025] Figure 1 This is a general framework diagram of the method of the present invention;

[0026] Figure 2 This is a schematic diagram of an overlapping sliding window and an edge-ignoring model. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other. To achieve the above objectives, this invention adopts the following technical solution.

[0028] The overall framework diagram of the method for estimating the rooftop photovoltaic potential of classified buildings based on remote sensing imagery proposed in this invention is as follows: Figure 1 As shown.

[0029] Step 1: Construct a building classification system for photovoltaic potential assessment to obtain building classification information from remote sensing images;

[0030] Building classification for photovoltaic (PV) potential assessment needs to consider both functional and morphological characteristics. Functionally, rooftop PV can be broadly categorized into commercial / industrial PV and residential PV. Commercial / industrial PV buildings include industrial plants, commercial buildings, government office buildings, schools, hospitals, and public buildings. These buildings typically have high electricity loads and high electricity prices, resulting in higher expected returns for PV projects. Industrial plants, in particular, with their large, flat roofs and minimal shading, are a preferred choice for PV development. Residential PV is primarily used in residential buildings. Due to the complex ownership of rooftops in urban residential buildings, residential PV is mainly concentrated in rural residential areas. However, rural residential buildings face challenges such as small usable area, unclear load-bearing capacity, and lower electricity prices, limiting their PV development potential. Therefore, in terms of priority for distributed PV development, commercial / industrial buildings have the highest potential, followed by rural residential buildings, and finally urban residential buildings.

[0031] From a structural perspective, building roofs are categorized into corrugated steel roofs and concrete roofs. Corrugated steel roofs are commonly used in low-rise buildings, factories, and warehouses. These roofs have relatively weak load-bearing capacity but a certain slope, making them suitable for direct installation of photovoltaic (PV) modules. Their remote sensing image characteristics include shape, size, color, texture, and shadows. Concrete roofs are commonly used in high-rise buildings, factories, large commercial buildings, and rural residential houses. These roofs have strong load-bearing capacity and are suitable for mounting brackets and installing PV modules at the optimal tilt angle. Their remote sensing image characteristics also include shape, size, color, texture, and shadows.

[0032] Based on the functional and morphological characteristics of the buildings mentioned above, the buildings are classified into four categories: commercial and industrial buildings with corrugated steel roofs, commercial and industrial buildings with concrete roofs, urban residential buildings, and rural residential buildings. This classification can obtain the building classification information required for photovoltaic potential assessment using only differences in image features from remote sensing images, such as shape, size, color, texture, and shadow.

[0033] Step 2: Create corresponding building samples for the four types of buildings to extract the four types of buildings;

[0034] This invention provides a method for batch acquisition of building samples, providing high-quality training data for building classification in photovoltaic potential assessment. This method efficiently generates building samples suitable for the classification system of this invention by combining building footprint data, region of interest (AOI) data, and visual interpretation. The method includes the following steps:

[0035] Data preparation includes: acquiring building footprint data and area of ​​interest (AOI) data;

[0036] Obtaining building footprint data includes: Building footprint refers to the projected outline of a building in a top-down view, which can provide key information such as the building's geographical location, boundary range, and floor area. This data is extracted through remote sensing imagery combined with a deep learning model, but annotation errors may exist.

[0037] Acquiring Area of ​​Interest (AOI) data involves storing AOI data in polygonal vector form, representing areas with specific functions or meanings (such as residential areas, commercial areas, industrial areas, etc.), and including relevant attribute information. The attributes of the AOI data can be used to preliminarily determine the functional type of buildings within the area.

[0038] Sample area selection includes: overlaying and analyzing building footprint data with AOI data, and visually selecting high-quality areas falling within the AOI as sample areas. Downloading the corresponding remote sensing imagery for subsequent building footprint correction and sample annotation.

[0039] Building footprint correction includes addressing two types of errors: allocation noise (labeling errors) and shape noise (inaccurate depiction), which are inherent to building footprint data extracted from historical remote sensing imagery. Therefore, corrections are needed for both types of noise. Specifically, in ArcMap software, the building footprint is moved as a whole relative to the remote sensing imagery to eliminate spatial discrepancies; patches incorrectly extracted as buildings are deleted; unlabeled buildings are added; and inaccurate building boundaries are corrected.

[0040] Sample type labeling, including:

[0041] Preliminary functional attribute acquisition: Based on the attribute information of AOI data, the functional type of buildings in the sample area (such as residential, commercial, industrial, public buildings, etc.) is preliminarily determined.

[0042] Visual interpretation correction: Based on visual interpretation of remote sensing images, the preliminary functional attributes are further corrected to the four building types defined in this invention, namely, industrial and commercial buildings with corrugated steel roofs, industrial and commercial buildings with concrete roofs, urban residential buildings, and rural residential buildings.

[0043] Based on the four types of building samples obtained using the above method, the instance segmentation pre-trained model is further fine-tuned to achieve accurate identification and boundary delineation of different types of buildings. Instance segmentation requires the algorithm not only to identify target objects in an image but also to accurately distinguish the boundaries of each object. This invention selects YOLOv8 as the instance segmentation pre-trained model and chooses from five different scale pre-trained models provided by YOLOv8 based on the dataset size and device performance.

[0044] By fine-tuning the selected YOLOv8 instance segmentation pre-trained model using the four types of building sample sets obtained through the above method, the model can be made capable of recognizing different types of buildings, thus obtaining a fine-tuned instance segmentation model. This model can accurately identify the type of building and precisely delineate its boundaries, providing technical support for subsequent photovoltaic potential assessment.

[0045] Step 3: The fine-tuned instance segmentation model will be used to extract information on four types of buildings over a large area;

[0046] When applying the trained model to extract information on four types of buildings over a large area, the remote sensing image needs to be cropped into fixed squares for extraction, and then the extracted results are stitched together. This invention uses a sliding overlapping window for extraction and borrows from existing edge-ignoring strategies for stitching. Specifically, it retains only targets falling within the effective area at the center of the window. For targets that fall exactly on the edge of the effective area, they are retained only if they fall on the right or bottom edge. Figure 2As shown, the solid-line rectangle represents the original remote sensing image, the thin dashed rectangle represents the filled remote sensing image, the thick dashed square (A) represents the sliding window, and the diamond-filled square (a) represents the effective region within the sliding window. The side length of A is set to the size of the input image of the YOLOv8 instance segmentation model, i.e., 640, the side length of a is set to 500, the step size of the sliding window is also 500, and the percentage of region a occupying the range of A is set to approximately 0.6. In the final stitching result, each pixel is predicted only once and is not an edge pixel. To ensure that the effective region can completely cover the original image, the original image is first filled with 70 pixels on each side before extraction.

[0047] The complete process of large-scale extraction and post-processing includes: (1) Traversing the remote sensing images of each township-level administrative region, using the above-mentioned sliding overlapping window method, extracting window by window using the trained model, and recording the information of each window, the bounding box within the window, and the mask. (2) Performing the following processing for each window: reading the pixel coordinates of the sliding window and converting them to latitude and longitude coordinates; calculating the pixel coordinates of the effective area and converting them to latitude and longitude coordinates; using the idea of ​​edge discarding, filtering the effective bounding boxes and converting the pixel coordinates to latitude and longitude coordinates; reading the mask within the effective bounding box. (3) Stitching the results of each window together, storing the sliding window, effective area, and bounding box in shpfile format, and storing the mask in tif format. (4) Using the community AOI and industrial and commercial AOI to perform two-step error correction on the mask. (5) Converting the corrected mask from tif format to shpfile format, setting the projection coordinate system and calculating the area, using the boundary of the township-level administrative region as the mask, and calculating the total area of ​​the roofs of various types of buildings under each township-level administrative region.

[0048] The aforementioned "two-step error correction" refers to the recognition results of the instance segmentation model optimized by this invention based on AOI data. Compared to commercial and industrial buildings with corrugated steel roofs and rural residential buildings, some buildings in the two categories of concrete-roofed commercial and industrial buildings and urban residential buildings have similar image features. Therefore, AOI data is used to correct the confusion between these two categories, with the following rules: First, if some buildings fall within the community AOI but are identified as concrete-roofed commercial and industrial buildings, they are corrected to urban residential buildings; second, if some buildings fall within the commercial and industrial AOI but are identified as urban residential buildings, they are corrected to concrete-roofed commercial and industrial buildings.

[0049] Step 4: Assess the untapped photovoltaic potential and total photovoltaic potential of each type of roof;

[0050] After obtaining the total roof area of ​​different types of buildings within the area, it is necessary to convert the roof area into usable roof area and area suitable for installing photovoltaic (PV) panels. Converting from roof area to usable roof area is to exclude the shaded area caused by elevator shafts, parapet walls, etc., as well as the area required for maintenance after PV installation. Converting from usable roof area to area suitable for PV panel installation takes into account the PV panel installation method. PV panel installation is divided into flat installation and optimal tilt angle installation. The optimal tilt angle refers to using brackets to support the PV panels so that they can receive solar radiation to the maximum extent; the determining factor for the optimal tilt angle is latitude. The optimal spacing is to maximize the use of roof space while ensuring that the front row of PV panels does not block the rear row of PV panels; the determining factors for the optimal spacing are installation tilt angle, roof slope and gradient, and PV panel size. If it is a flat installation, the area suitable for installing PV panels is considered equal to the usable roof area. In rough estimation, these two are not distinguished. This invention obtains the usability coefficient from roof area to area suitable for installing PV panels from practice. Specifically, based on the obtained building vector, the tiles in the remote sensing image are screened to exclude tiles that do not contain buildings. Then, the generated rooftop photovoltaic sample dataset is used to train and fine-tune HRNet or other semantic segmentation pre-trained models. The trained semantic segmentation pre-trained models are used to identify rooftop photovoltaic panels based on tiles, thus classifying building rooftops into two categories: developed rooftops and undeveloped rooftops, with their areas calculated separately. By calculating the ratio of photovoltaic panel area to roof area on developed rooftops, the area utilization rate of that rooftop can be obtained. For each type of rooftop, the average area utilization rate of all developed rooftops in the region is taken as the utilization coefficient for that type of rooftop. Applying the utilization coefficient to the total area of ​​that type of rooftop allows for estimation of the total installable photovoltaic panel area; applying the utilization coefficient to the undeveloped rooftop area within that type allows for estimation of the installable photovoltaic panel area on undeveloped rooftops. After obtaining the installable photovoltaic panel area, the installed capacity potential can be calculated using the power generation per unit area of ​​commonly available photovoltaic modules. There are two ways to calculate annual power generation: one is to estimate by multiplying the installed capacity by the annual effective utilization hours in the region; the other is to estimate by multiplying the photovoltaic panel area by the solar radiation and system conversion efficiency in the region. The total photovoltaic potential of each type of roof can be calculated based on the total area of ​​installable photovoltaic panels for each type of roof; the undeveloped photovoltaic potential of each type of roof can be calculated based on the area of ​​installable photovoltaic panels for each type of roof.

[0051] This invention provides a device for estimating the potential of rooftop photovoltaic power of classified buildings based on remote sensing imagery, comprising:

[0052] The building classification module constructs a building classification system for photovoltaic potential assessment, which is used to obtain building classification information from remote sensing images. The classification system is specifically divided into four categories: industrial and commercial buildings with corrugated steel roofs, industrial and commercial buildings with concrete roofs, urban residential buildings, and rural residential buildings.

[0053] The building extraction module generates corresponding building samples for four types of buildings.

[0054] The building roof area calculation module, based on the corresponding building samples of four types of buildings, fine-tunes the instance segmentation pre-trained model to extract the total roof area of ​​the four types of buildings in the test area;

[0055] The photovoltaic potential calculation module calculates the developed roof area and the undeveloped roof area based on the total roof area of ​​various types of buildings, and evaluates the total photovoltaic potential and undeveloped photovoltaic potential of each type of roof.

[0056] The present invention also provides an electronic device, comprising: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement the method.

[0057] The present invention also provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, cause the processor to implement the method described thereon.

[0058] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of the present invention can be implemented using various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0059] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0060] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0061] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0062] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0063] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for estimating the photovoltaic potential of a classified building roof based on remote sensing imagery, characterized in that, The method comprises the following steps: Step 1, constructing a building classification system for photovoltaic potential evaluation, which is used to obtain building classification information on remote sensing images, and the classification system is specifically: the buildings are divided into four categories of color steel tile roof industrial and commercial buildings, concrete roof industrial and commercial buildings, urban residential buildings and rural residential buildings; Step 2, making building samples corresponding to the four types of buildings; Step 3, based on the building samples, fine-tuning the instance segmentation pre-trained model to extract the total roof area of the four types of buildings in the to-be-tested region; Step 4, calculating the developed roof area and the to-be-developed roof area according to the total roof area of each type of building, and evaluating the total photovoltaic potential of each type of roof; Step 2 comprises: based on building footprints, AOI data and visual interpretation, making building samples of the four types; The building footprint is the bottom contour of the building, which is the contour data of the building projected onto the ground in the top view, and the geographical position, boundary space range and floor area of the single building are determined through the building footprint; the AOI data is the spatial boundary and related attributes of the regional range stored in the form of polygon vector data, and the regional range includes residential, commercial, industrial and public buildings.

2. The method of claim 1, wherein the method further comprises: Step 2 comprises the following steps: first, superimpose the existing building footprint data and AOI data, visually select a high-quality area within the AOI data as a sample area, download the remote sensing image corresponding to the sample area, correct the sample area to obtain the building sample, and further add type information.

3. The method of claim 1, wherein the method further comprises: Step 3 comprises: Step 3.1, based on the building samples, fine-tuning the instance segmentation pre-trained model; Step 3.2, traversing the remote sensing image of the to-be-tested region, using a sliding overlapping window, using the trained model to extract window by window, and recording the information of the bounding box and the mask in each window and the effective area within the window; Step 3.3, processing each window as follows: reading the pixel coordinates of the sliding window, converting them to latitude and longitude coordinates; calculating the pixel coordinates of the effective area, converting them to latitude and longitude coordinates; discarding the edges, selecting the effective bounding box, and converting the pixel coordinates to latitude and longitude coordinates; reading the mask within the effective bounding box; Step 3.4, splicing the results of each window, storing the sliding window, effective area and bounding box in shpfile format, and storing the mask in tif format; Step 3.5, using the community AOI and industrial and commercial AOI to correct the mask in two steps; Step 3.6, converting the corrected mask from tif format to shpfile format, setting the projection coordinate system and calculating the area, using the township administrative boundary as the mask, and calculating the total area of the roofs of the buildings in the to-be-tested region.

4. The method of claim 3, wherein the method further comprises: The above two-step correction refers to that the building falls within the community AOI but is identified as a concrete roof industrial and commercial building, and is corrected to an urban residential building; the building falls within the industrial and commercial AOI but is identified as an urban residential building, and is corrected to a concrete roof industrial and commercial building.

5. The method of claim 1, wherein, In step 4, the buildings in the to-be-tested area are classified into developed roofs and to-be-developed roofs, the area utilization rate of the developed roofs is obtained by calculating the ratio of the area of the photovoltaic panels on the developed roofs to the area of the roofs, for each type of roof, the average value of the area utilization rates of all the developed roofs in the area is taken as the available coefficient of the type of roof, the available coefficient is applied to the total area of the type of roof, and the total installable photovoltaic panel area on the type of roof is estimated; The available coefficient is applied to the area of the to-be-developed roofs in the type of roof, the installable photovoltaic panel area on the to-be-developed roofs is estimated, the total photovoltaic potential of each type of roof is calculated according to the total installable photovoltaic panel area of each type of roof, and the to-be-developed photovoltaic potential of each type of roof is calculated according to the installable photovoltaic panel area of the to-be-developed roofs of each type of roof.

6. A method of estimating the photovoltaic potential of a building roof based on remote sensing imagery according to claim 5, characterized in that, The buildings in the to-be-tested area are classified into developed roofs and to-be-developed roofs, including: based on the obtained building roof vector, the tiles in the remote sensing image are screened to exclude tiles that do not contain buildings, a semantic segmentation model is used to identify the roof photovoltaic panels based on the tiles, and then the building roofs are classified into developed roofs and to-be-developed roofs, the areas are calculated respectively, and the area of the photovoltaic panels on the developed roofs is calculated.

7. A device for estimating the photovoltaic potential of a building roof based on remote sensing imagery according to any one of claims 1-6, characterized in that, Including: The building classification module constructs a building classification system for photovoltaic potential evaluation, which is used to obtain building classification information on the remote sensing image, and the classification system is specifically: the buildings are classified into four categories of color steel tile roof industrial and commercial buildings, concrete roof industrial and commercial buildings, urban residential buildings, and rural residential buildings; The building extraction module makes building samples corresponding to the four types of buildings; The building roof area calculation module adjusts the instance segmentation pre-training model based on the building samples corresponding to the four types of buildings to extract the total roof area of the four types of buildings in the to-be-tested area; The photovoltaic potential calculation module calculates the developed roof area and the to-be-developed roof area according to the total roof area of each type of building, and evaluates the total photovoltaic potential and the to-be-developed photovoltaic potential of each type of roof.

8. An electronic device, comprising: Including: One or more processors; A memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method of any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, Executable instructions are stored thereon, which are executed by a processor to make the processor implement the method of any one of claims 1 to 6.

Images