EP4616317A1 - A method for determining installation constraints for the installation of photovoltaic panels on a roof of a building - Google Patents

A method for determining installation constraints for the installation of photovoltaic panels on a roof of a building

Info

Publication number
EP4616317A1
EP4616317A1 EP23804979.5A EP23804979A EP4616317A1 EP 4616317 A1 EP4616317 A1 EP 4616317A1 EP 23804979 A EP23804979 A EP 23804979A EP 4616317 A1 EP4616317 A1 EP 4616317A1
Authority
EP
European Patent Office
Prior art keywords
roof
building
image
category
interest
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
EP23804979.5A
Other languages
German (de)
French (fr)
Inventor
Simon ALBERGEL
Emmanuel LE BORGNE
David BENHAIEM
Thibault PERIER
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
TotalEnergies Onetech SAS
Original Assignee
TotalEnergies Onetech SAS
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by TotalEnergies Onetech SAS filed Critical TotalEnergies Onetech SAS
Publication of EP4616317A1 publication Critical patent/EP4616317A1/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/176—Urban or other man-made structures
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/10—Geometric CAD
    • G06F30/13—Architectural design, e.g. computer-aided architectural design [CAAD] related to design of buildings, bridges, landscapes, production plants or roads
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model

Definitions

  • rooftops represent an impressive potential. Therefore, prospectors are looking for additional information that will allow them to target the most interesting sites.
  • the type of roof can have an impact on the conditions of installation of photovoltaic panels and therefore on the feasibility and cost-effectiveness of a project.
  • the invention relates to a method for determining installation constraints for the installation of photovoltaic panels on a roof of a building, the method comprising the following phases which are computer-implemented:
  • the training database associating an image seen from the sky of an area comprising at least one building having a roof, called building image, with a roof mask for each building of the building image and a roof category for each building of the building image, the roof category being chosen among a predetermined list of categories, each roof category implying specific installation constraints for the installation of photovoltaic panels,
  • the trained model having as input an image of an area seen from the sky and being configured for determining a roof mask and a roof category among the predetermined list of categories for each building imaged on the input image,
  • the method according to the invention may comprise one or more of the following features considered alone or in any combination that is technically possible:
  • the specific installation constraints for each roof category of the predetermined list of categories comprise at least a constraint relative to:
  • the method comprises a phase of displaying, on a display, the image of interest on which is superimposed the determined roof mask(s) and a data indicative of the roof category of each roof ;
  • the trained model associates a confidence metric to each determined roof category
  • phase of building the training database comprises the following steps :
  • - the roof category for each building image is obtained using a labelling tool, the labelling tool being configured to display the building image on a display as well as the predetermined list of categories and to acquire the roof category selected by a user among the predetermined list of categories;
  • the predetermined list of categories comprises at least the following categories: concrete-gravel flat roof, steel roof, roof made in fiber cement without asbestos, roof made in fiber cement with asbestos, tiles roof, and flat roofs;
  • the method comprises a phase of determining an arrangement of photovoltaic panels on each roof imaged on the image of interest , the determination phase comprising:
  • the phase of determining an arrangement of photovoltaic panels comprises displaying, on a display, the image of interest on which is superimposed the determined arrangement of photovoltaic panels;
  • the method comprises a phase of determining solar data relative to the arrangement of photovoltaic panels determined for each roof of the image of interest;
  • the building images and the images of interest are images in the visible range acquired by a camera mounted on an aircraft;
  • the detection and classification model is a convolutional neural network
  • the detection and classification model has been pretrained on a database different from the training database so as to achieve transfer learning.
  • the invention also relates to a computer program product comprising a readable information carrier having stored thereon a computer program comprising program instructions, the computer program being loadable onto a data processing unit and causing a method as previously described to be carried out when the computer program is carried out on the data processing unit.
  • the invention also relates to a readable information carrier on which is stored a computer program product as previously described.
  • Figure 1 is a schematic view of an example of a computer for implementing a method for determining installation constraints for the installation of photovoltaic panels on a roof of a building
  • Figure 2 is a schematic representation illustrating the implementation of phases of a method for determining installation constraints for the installation of photovoltaic panels on a roof of a building
  • Figure 3 is a schematic representation of an example of the building of a training database
  • Figure 4 is a schematic representation of an example of some phases of the method for determining installation constraints for the installation of photovoltaic panels on a roof of a building.
  • a computer 20 and a computer program product 22 are illustrated in figure 1 .
  • the computer 20 is preferably a computer.
  • the calculator 20 is a computer or computing system, or similar electronic computing device adapted to manipulate and/or transform data represented as physical, such as electronic, quantities within the computing system's registers and/or memories into other data similarly represented as physical quantities within the computing system's memories, registers or other such information storage, transmission or display devices.
  • the calculator 20 interacts with the computer program product 22.
  • the calculator 20 comprises a processor 24 comprising a data processing unit 26, memories 28 and a reader 30 for information media.
  • the calculator 20 comprises a human machine interface 32, such as a keyboard, and a display 34.
  • the computer program product 22 comprises an information medium 36.
  • the information medium 36 is a medium readable by the calculator 20, usually by the data processing unit 26.
  • the readable information medium 36 is a medium suitable for storing electronic instructions and capable of being coupled to a computer system bus.
  • the information medium 36 is a USB key, a floppy disk or flexible disk (of the English name "Floppy disc"), an optical disk, a CD-ROM, a magneto-optical disk, a ROM memory, a memory RAM, EPROM memory, EEPROM memory, magnetic card or optical card.
  • the computer program 22 comprising program instructions.
  • the computer program 22 is loadable on the data processing unit 26 and is adapted to entail the implementation of a method for determining installation constraints for the installation of photovoltaic panels on a roof of a building, when the computer program 22 is loaded on the processing unit 26 of the calculator 20.
  • Obstacle means an element present on the surface of the roof and incompatible with the placement of a photovoltaic panel on this element. Obstacles O are, for example, chimneys, windows (velux), air conditioning systems or vegetation (trees, plants).
  • the determination method comprises a phase 100 of building a training database DB.
  • the building phase 100 is, for example, implemented by the calculator 20 interacting with the computer program product 22, that is to say is computer-implemented.
  • the training database DB associates an image seen from the sky of an area comprising at least one building having a roof R, called building image IM B , with a roof mask MR for each building of the building image IM B and a roof category CR for each building of the building image IM B .
  • the buildings images IM B have been acquired by a sensor, such as a camera mounted on an aircraft.
  • the building images IM B are satellite images.
  • the building images IM B are images in the visible range (400 nanometers to 700 nanometers).
  • each building image IM B represents only one building. In another example, at least one building image IM B represents several buildings.
  • the roof category CR is chosen among a predetermined list of categories.
  • Each roof category CR implies specific installation constraints for the installation of photovoltaic panels on a roof R belonging to the roof category CR.
  • the predetermined list of categories comprises at least the following categories: concrete-gravel flat roof, steel roof, roof made in fiber cement without asbestos, roof made in fiber cement with asbestos, tiles roof, and flat roofs.
  • the list of categories might also comprises only some of the above categories, or other additional categories.
  • the specific installation constraints for each roof category CR of the predetermined list of categories comprise at least a constraint relative to:
  • the installation constraints comprise a constraint of using fixing clamps to fix the photovoltaic panels.
  • the installation constraints comprise a constraint of removing the asbestos from the roof R before installing the photovoltaic panels.
  • the building phase 100 comprises the following steps 100A to 100D which are illustrated on figure 3.
  • Step 100A is a step of collecting input data.
  • the input data comprise input images IM seen from the sky of areas comprising buildings having roofs R, and a roof mask MR for each building of each input image IM.
  • the roof mask(s) M are considered as being the ground surface of the building, the ground surface being a known data.
  • the roof mask(s) MR are obtained by a segmentation performed on the input image IM to highlight the roofs R.
  • the segmentation is, for example, performed by an edge detection algorithm.
  • the roof category CR for each building image IM B is obtained using a labelling tool T L .
  • the labelling tool T is an algorithm configured to display the building images IM B on a display as well as the predetermined list of categories (for example superimposed on the building images IM B ), and to acquire the roof category CR selected by a user among the predetermined list of categories.
  • the labelling tool T renders easier the labelling of the building image IM B and reduces the risks of errors.
  • the labelling tool T displays three categories C1 , C2, C3, and the selected category is C2.
  • Step 100D is a step of building the training database DB by associating each building image IM B with the corresponding roof mask MR and the obtained roof category CR.
  • At least one building image IM B represents several buildings.
  • the labelling tool T is configured to display the building images IM B on a display as well as the predetermined list of categories, and to acquire the roof categories CR, selected by a user for each roof, among the predetermined list of categories.
  • the determination method comprises a phase 200 of training a detection and classification model M on the basis of the training database DB so as to obtain a trained model M’.
  • the training phase 200 is, for example, implemented by the calculator 20 interacting with the computer program product 22, that is to say is computer-implemented.
  • the input of the trained model M’ is an image of an area seen from the sky.
  • the trained model M’ is configured for determining a roof mask MR and a roof category C among the predetermined list of categories for each building imaged on the input image IM.
  • the detection and classification model M is a deep learning model.
  • the detection and classification model M is a convolutional neural network (CNN).
  • CNN convolutional neural network
  • the detection and classification model M has been pretrained on a database different from the training database DB so as to achieve transfer learning.
  • the training technique is typically based on supervised learning, with training and validation substeps (performed by dividing the training database DB, for example 80% of the data for the training substep and 20% of the data for the validation substep).
  • the detection and classification model M is a mask Region-based Convolutional Neural Network (mask RCNN) with a Feature Pyramid Network (FPN) as backbone.
  • the detection and classification model M is for example obtained from the Facebook Al computer vision tool Detectron 2.
  • the trained model M’ associates a confidence metric to each determined roof category CR.
  • the determination method comprises a phase 300 of operating the trained model M’.
  • the operating phase 300 is, for example, implemented by the calculator 20 interacting with the computer program product 22, that is to say is computer-implemented.
  • the operating phase 300 comprises the following steps:
  • the determination method also comprises a phase 400 of displaying, on a display (for example the display 34 of the calculator 20), the image of interest I Mi on which is superimposed the determined roof mask(s) MR and a data indicative of the roof category CR of each roof R.
  • the confidence metric is also displayed on the display.
  • Figure 3 illustrates an example of implementing phase 300.
  • the determination method comprises a phase 500 of determining an arrangement A of photovoltaic panels on each roof R imaged on the image of interest IMi.
  • the determination phase 500 is, for example, implemented by the calculator 20 interacting with the computer program product 22, that is to say is computer-implemented.
  • a target image IM T is obtained by applying the corresponding roof mask(s) M on the image of interest I Mi so that only the roofs R are represented on the target image IM T .
  • a contour detection is performed on the target image IM T to obtain a contour mask M c for each roof R.
  • the detected contours materialize obstacles O on the roof R.
  • the contour detection is performed by applying a Canny filter on the target image IM T .
  • the obstacles O are highlighted on each roof R of the image of interest I Mi by combining the roof masks MR and the contour masks Me in order to obtain a mask of the obstacles O present on each roof R of the image of interest IMi, called obstacle mask Mo.
  • the determination phase 500 comprises a step 500B of determining an arrangement A of photovoltaic panels on each roof R of the image of interest I Mi as a function of the corresponding obstacle mask M o so that each photovoltaic panel covers a portion of the roof R that is free of obstacles O.
  • step 500B comprises the positioning of representative shapes of photovoltaic panels on each roof R of the obstacle mask Mo, followed by the elimination of the shapes at least partially superimposed on an obstacle O.
  • the determined arrangement A of photovoltaic panels matches the remaining shapes.
  • the steps 500A and 500B are materialized by the model M2 on figure 4.
  • the method comprises a phase of designing and/or manufacturing an arrangement A of photovoltaic panels corresponding to the panels of the determined arrangement A, possibly followed by the actual positioning of the photovoltaic panels on the roof(s) R of the studied area.
  • the above method enables to determine both the surface area of roofs R on an image as well as the category of the roof R. This enables to determine in a more precise way installation constraints for the installation of photovoltaic panels on a roof R of a building.
  • the prediction is fast, about one second to make a prediction on an image.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Software Systems (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Computational Linguistics (AREA)
  • Mathematical Physics (AREA)
  • Biophysics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Databases & Information Systems (AREA)
  • Medical Informatics (AREA)
  • Photovoltaic Devices (AREA)

Abstract

The invention relates to a method for determining installation constraints for the installation of photovoltaic panels on a roof (R) of a building, the method comprising: - a phase of building a training database (DB) associating an image seen from the sky of an area comprising, called building image (IMB), with a roof mask (MR) for each building of the building image (IMB) and a roof category for each building of the building image (IMB), and - a phase of training a detection and classification model on the basis of the training database (DB) so as to obtain a trained model configured for determining a roof mask (MR) and a roof category for each building imaged on an input image (IM).

Description

A method for determining installation constraints for the installation of photovoltaic panels on a roof of a building
TECHNICAL FIELD OF THE INVENTION
The present invention concerns a method for determining installation constraints for the installation of photovoltaic panels on a roof of a building. The present invention also concerns an associated computer program product. The present invention also relates to an associated readable information carrier.
BACKGROUND OF THE INVENTION
The scarcity of areas for the installation of photovoltaic sites pushes prospectors to look for entropized areas on which to deploy production assets.
In this context, rooftops represent an impressive potential. Therefore, prospectors are looking for additional information that will allow them to target the most interesting sites.
Among the indicators of interest, the type of roof can have an impact on the conditions of installation of photovoltaic panels and therefore on the feasibility and cost-effectiveness of a project.
However, such indicators are difficult to obtain and are not systematically taken into account.
SUMMARY OF THE INVENTION
Hence, there exists a need for a method enabling to determine in a more precise way installation constraints for the installation of photovoltaic panels on a roof of building.
To this end, the invention relates to a method for determining installation constraints for the installation of photovoltaic panels on a roof of a building, the method comprising the following phases which are computer-implemented:
- a phase of building a training database, the training database associating an image seen from the sky of an area comprising at least one building having a roof, called building image, with a roof mask for each building of the building image and a roof category for each building of the building image, the roof category being chosen among a predetermined list of categories, each roof category implying specific installation constraints for the installation of photovoltaic panels,
- a phase of training a detection and classification model on the basis of the training database so as to obtain a trained model, the trained model having as input an image of an area seen from the sky and being configured for determining a roof mask and a roof category among the predetermined list of categories for each building imaged on the input image,
- a phase of operating the trained model comprising the following steps:
• receiving an image of interest of an area seen from the sky, and
• determining, by the trained model, a roof mask and a roof category for each building imaged on the image of interest, the determined roof category(ies) indicating installation constraints for the installation of photovoltaic panels on the corresponding roof(s).
The method according to the invention may comprise one or more of the following features considered alone or in any combination that is technically possible:
- the specific installation constraints for each roof category of the predetermined list of categories comprise at least a constraint relative to:
- the hanging system of the photovoltaic panel on the roof, and/or
- a treatment to be carried out on the roof before the installation of the photovoltaic panels;
- the method comprises a phase of displaying, on a display, the image of interest on which is superimposed the determined roof mask(s) and a data indicative of the roof category of each roof ;
- the trained model associates a confidence metric to each determined roof category;
- the phase of building the training database comprises the following steps :
• collecting input data comprising: input images seen from the sky of areas comprising buildings having roofs, and a roof mask for each building of each input image,
• cutting out the input images around each building, on the basis of the roof mask(s) of the input images, so as to obtain an image for each building, each obtained image forming a building image,
• labelling each building image with a roof category, the roof category being selected among a predetermined list of categories, and
• building the training database by associating each building image with the corresponding roof mask and the obtained roof category;
- the roof category for each building image is obtained using a labelling tool, the labelling tool being configured to display the building image on a display as well as the predetermined list of categories and to acquire the roof category selected by a user among the predetermined list of categories; - the predetermined list of categories comprises at least the following categories: concrete-gravel flat roof, steel roof, roof made in fiber cement without asbestos, roof made in fiber cement with asbestos, tiles roof, and flat roofs;
- the method comprises a phase of determining an arrangement of photovoltaic panels on each roof imaged on the image of interest , the determination phase comprising:
- a step of detecting obstacles on each roof of the image of interest by processing the image of interest using the corresponding roof mask(s), to obtain an obstacle mask for each roof of the image of interest , and
- a step of determining an arrangement of photovoltaic panels on each roof of the image of interest as a function of the corresponding obstacle mask so that each photovoltaic panel covers a portion of the roof that is free of obstacles;
- the phase of determining an arrangement of photovoltaic panels comprises displaying, on a display, the image of interest on which is superimposed the determined arrangement of photovoltaic panels;
- the method comprises a phase of determining solar data relative to the arrangement of photovoltaic panels determined for each roof of the image of interest;
- the building images and the images of interest are images in the visible range acquired by a camera mounted on an aircraft;
- the detection and classification model is a convolutional neural network;
- the detection and classification model has been pretrained on a database different from the training database so as to achieve transfer learning.
The invention also relates to a computer program product comprising a readable information carrier having stored thereon a computer program comprising program instructions, the computer program being loadable onto a data processing unit and causing a method as previously described to be carried out when the computer program is carried out on the data processing unit.
The invention also relates to a readable information carrier on which is stored a computer program product as previously described.
BRIEF DESCRIPTION OF THE DRAWINGS
The invention will be easier to understand in view of the following description, provided solely as an example and with reference to the appended drawings in which:
Figure 1 is a schematic view of an example of a computer for implementing a method for determining installation constraints for the installation of photovoltaic panels on a roof of a building, Figure 2 is a schematic representation illustrating the implementation of phases of a method for determining installation constraints for the installation of photovoltaic panels on a roof of a building,
Figure 3 is a schematic representation of an example of the building of a training database, and
Figure 4 is a schematic representation of an example of some phases of the method for determining installation constraints for the installation of photovoltaic panels on a roof of a building.
DETAILED DESCRIPTION OF SOME EMBODIMENTS
A computer 20 and a computer program product 22 are illustrated in figure 1 .
The computer 20 is preferably a computer.
More generally, the calculator 20 is a computer or computing system, or similar electronic computing device adapted to manipulate and/or transform data represented as physical, such as electronic, quantities within the computing system's registers and/or memories into other data similarly represented as physical quantities within the computing system's memories, registers or other such information storage, transmission or display devices.
The calculator 20 interacts with the computer program product 22.
As illustrated on figure 1 , the calculator 20 comprises a processor 24 comprising a data processing unit 26, memories 28 and a reader 30 for information media. In the example illustrated on figure 1 , the calculator 20 comprises a human machine interface 32, such as a keyboard, and a display 34.
The computer program product 22 comprises an information medium 36.
The information medium 36 is a medium readable by the calculator 20, usually by the data processing unit 26. The readable information medium 36 is a medium suitable for storing electronic instructions and capable of being coupled to a computer system bus.
By way of example, the information medium 36 is a USB key, a floppy disk or flexible disk (of the English name "Floppy disc"), an optical disk, a CD-ROM, a magneto-optical disk, a ROM memory, a memory RAM, EPROM memory, EEPROM memory, magnetic card or optical card.
On the information medium 36 is stored the computer program 22 comprising program instructions.
The computer program 22 is loadable on the data processing unit 26 and is adapted to entail the implementation of a method for determining installation constraints for the installation of photovoltaic panels on a roof of a building, when the computer program 22 is loaded on the processing unit 26 of the calculator 20.
A method for determining installation constraints for the installation of photovoltaic panels on a roof R of a building, will now be described with reference to the organigram of figure 2, and to figures 3 and 4, which illustrate examples of some phases of this method.
Some obstacles O might be present on the roofs R. The term "obstacle" means an element present on the surface of the roof and incompatible with the placement of a photovoltaic panel on this element. Obstacles O are, for example, chimneys, windows (velux), air conditioning systems or vegetation (trees, plants).
The determination method comprises a phase 100 of building a training database DB. The building phase 100 is, for example, implemented by the calculator 20 interacting with the computer program product 22, that is to say is computer-implemented.
The training database DB associates an image seen from the sky of an area comprising at least one building having a roof R, called building image IMB, with a roof mask MR for each building of the building image IMB and a roof category CR for each building of the building image IMB.
By the term “seen from the sky”, it is understood that the images have been taken from an elevated view. For example, the building images IMB have been acquired by a sensor, such as a camera mounted on an aircraft. In a variant, the building images IMB are satellite images.
Preferably, the building images IMB are images in the visible range (400 nanometers to 700 nanometers).
In an example, each building image IMB represents only one building. In another example, at least one building image IMB represents several buildings.
The roof category CR is chosen among a predetermined list of categories. Each roof category CR implies specific installation constraints for the installation of photovoltaic panels on a roof R belonging to the roof category CR.
Preferably, the predetermined list of categories comprises at least the following categories: concrete-gravel flat roof, steel roof, roof made in fiber cement without asbestos, roof made in fiber cement with asbestos, tiles roof, and flat roofs. The list of categories might also comprises only some of the above categories, or other additional categories.
In an example, the specific installation constraints for each roof category CR of the predetermined list of categories comprise at least a constraint relative to:
- the hanging system of the photovoltaic panel on the roof R, and/or
- a treatment to be carried out on the roof R before the installation of the photovoltaic panels. For example, in the case of tiles roofs, the installation constraints comprise a constraint of using fixing clamps to fix the photovoltaic panels.
For example, in the case of roofs made in fiber cement with asbestos, the installation constraints comprise a constraint of removing the asbestos from the roof R before installing the photovoltaic panels.
In an example of embodiment, the building phase 100 comprises the following steps 100A to 100D which are illustrated on figure 3.
Step 100A is a step of collecting input data. The input data comprise input images IM seen from the sky of areas comprising buildings having roofs R, and a roof mask MR for each building of each input image IM.
In an example, the roof mask(s) M are considered as being the ground surface of the building, the ground surface being a known data.
In another example, the roof mask(s) MR are obtained by a segmentation performed on the input image IM to highlight the roofs R. The segmentation is, for example, performed by an edge detection algorithm.
Step 100B is a step of cutting out the input images IM around each building, on the basis of the roof mask(s) MR of the input images IM, so as to obtain an image for each building. Each obtained image forms a building image IMB. Step 100B is possibly an optional step when the input images IM represent only one building.
Step 100C is a step of labelling each building image IMB with a roof category CR, the roof category CR being selected among the predetermined list of categories.
In an example, the roof category CR for each building image IMB is obtained using a labelling tool TL. The labelling tool T is an algorithm configured to display the building images IMB on a display as well as the predetermined list of categories (for example superimposed on the building images IMB), and to acquire the roof category CR selected by a user among the predetermined list of categories. The labelling tool T renders easier the labelling of the building image IMB and reduces the risks of errors.
In the example of figure 3, the labelling tool T displays three categories C1 , C2, C3, and the selected category is C2.
Step 100D is a step of building the training database DB by associating each building image IMB with the corresponding roof mask MR and the obtained roof category CR.
In a variant, at least one building image IMB represents several buildings. The labelling tool T is configured to display the building images IMB on a display as well as the predetermined list of categories, and to acquire the roof categories CR, selected by a user for each roof, among the predetermined list of categories. The determination method comprises a phase 200 of training a detection and classification model M on the basis of the training database DB so as to obtain a trained model M’. The training phase 200 is, for example, implemented by the calculator 20 interacting with the computer program product 22, that is to say is computer-implemented.
The input of the trained model M’ is an image of an area seen from the sky. The trained model M’ is configured for determining a roof mask MR and a roof category C among the predetermined list of categories for each building imaged on the input image IM.
The detection and classification model M is a deep learning model.
Preferably, the detection and classification model M is a convolutional neural network (CNN).
Preferably, the detection and classification model M has been pretrained on a database different from the training database DB so as to achieve transfer learning.
The training technique is typically based on supervised learning, with training and validation substeps (performed by dividing the training database DB, for example 80% of the data for the training substep and 20% of the data for the validation substep).
In an example, the detection and classification model M is a mask Region-based Convolutional Neural Network (mask RCNN) with a Feature Pyramid Network (FPN) as backbone. The detection and classification model M is for example obtained from the Facebook Al computer vision tool Detectron 2.
Preferably, the trained model M’ associates a confidence metric to each determined roof category CR.
The determination method comprises a phase 300 of operating the trained model M’. The operating phase 300 is, for example, implemented by the calculator 20 interacting with the computer program product 22, that is to say is computer-implemented.
The operating phase 300 comprises the following steps:
• receiving an image of interest I Mi of an area seen from the sky, and
• determining, by the trained model M’, a roof mask MR and a roof category CR for each building imaged on the image of interest IMi, the determined roof categories CR indicating installation constraints for the installation of photovoltaic panels on the corresponding roofs R.
Preferably, the determination method also comprises a phase 400 of displaying, on a display (for example the display 34 of the calculator 20), the image of interest I Mi on which is superimposed the determined roof mask(s) MR and a data indicative of the roof category CR of each roof R. Preferably, if applicable, the confidence metric is also displayed on the display. Figure 3 illustrates an example of implementing phase 300. Optionally, the determination method comprises a phase 500 of determining an arrangement A of photovoltaic panels on each roof R imaged on the image of interest IMi. The determination phase 500 is, for example, implemented by the calculator 20 interacting with the computer program product 22, that is to say is computer-implemented.
The determination phase 500 comprises a step 500A of detecting obstacles O on each roof R of the image of interest IMi by processing the image of interest IMi using the corresponding roof mask(s) MR. This enables to obtain an obstacle mask Mo for each roof R of the image of interest IMi.
In an example of implementation, a target image IMT is obtained by applying the corresponding roof mask(s) M on the image of interest I Mi so that only the roofs R are represented on the target image IMT.
Then, a contour detection is performed on the target image IMT to obtain a contour mask Mc for each roof R. The detected contours materialize obstacles O on the roof R. For example, the contour detection is performed by applying a Canny filter on the target image IMT.
Thereafter, the obstacles O are highlighted on each roof R of the image of interest I Mi by combining the roof masks MR and the contour masks Me in order to obtain a mask of the obstacles O present on each roof R of the image of interest IMi, called obstacle mask Mo.
The determination phase 500 comprises a step 500B of determining an arrangement A of photovoltaic panels on each roof R of the image of interest I Mi as a function of the corresponding obstacle mask Mo so that each photovoltaic panel covers a portion of the roof R that is free of obstacles O. In an example, step 500B comprises the positioning of representative shapes of photovoltaic panels on each roof R of the obstacle mask Mo, followed by the elimination of the shapes at least partially superimposed on an obstacle O. The determined arrangement A of photovoltaic panels matches the remaining shapes.
The steps 500A and 500B are materialized by the model M2 on figure 4.
Optionally, the determination phase 500 comprises a step 400C of displaying, on a display, the image of interest I Mi on which is superimposed the determined arrangement A of photovoltaic panels.
Optionally, the determination method comprises a phase 600 of determining solar data relative to the arrangement A of photovoltaic panels determined for each roof R of the image of interest IMi. The determination phase 600 is, for example, implemented by the calculator 20 interacting with the computer program product 22, that is to say is computer- implemented. In an embodiment, the solar data comprises at least one data among the solar potential of the arrangement A and the yield of the arrangement A. The solar potential is the solar power received on a zone in kilowatt per hour per year (kW/hours/year). The yield is the ratio between the energy produced (kWh) and the theoretical power (KWp) of the installation.
Optionally, the method comprises a phase of designing and/or manufacturing an arrangement A of photovoltaic panels corresponding to the panels of the determined arrangement A, possibly followed by the actual positioning of the photovoltaic panels on the roof(s) R of the studied area. Hence, the above method enables to determine both the surface area of roofs R on an image as well as the category of the roof R. This enables to determine in a more precise way installation constraints for the installation of photovoltaic panels on a roof R of a building.
In addition, the prediction is fast, about one second to make a prediction on an image. The person skilled in the art will understand that the embodiments and variants described above can be combined to form new embodiments provided that they are technically compatible.

Claims

1 A method for determining installation constraints for the installation of photovoltaic panels on a roof (R) of a building, the method comprising the following phases which are computer-implemented:
- a phase of building a training database (DB), the training database (DB) associating an image seen from the sky of an area comprising at least one building having a roof (R), called building image (IMB), with a roof mask (MR) for each building of the building image (IMB) and a roof category (CR) for each building of the building image (IMB), the roof category (C ) being chosen among a predetermined list of categories, each roof category (CR) implying specific installation constraints for the installation of photovoltaic panels,
- a phase of training a detection and classification model (M) on the basis of the training database (DB) so as to obtain a trained model (M’), the trained model (M) having as input an image of an area seen from the sky and being configured for determining a roof mask (MR) and a roof category (CR) among the predetermined list of categories for each building imaged on the input image (IM),
- a phase of operating the trained model (M’) comprising the following steps:
• receiving an image of interest ( I Mi) of an area seen from the sky, and
• determining, by the trained model (M’), a roof mask (MR) and a roof category (CR) for each building imaged on the image of interest (IMi), the determined roof category(ies) (CR) indicating installation constraints for the installation of photovoltaic panels on the corresponding roof(s).
2.- A method according to claim 1 , wherein the specific installation constraints for each roof category (CR) of the predetermined list of categories comprise at least a constraint relative to:
- the hanging system of the photovoltaic panel on the roof (R), and/or
- a treatment to be carried out on the roof (R) before the installation of the photovoltaic panels.
3.- A method according to claim 1 or 2, wherein the method comprises a phase of displaying, on a display, the image of interest (IMi) on which is superimposed the determined roof mask(s) (MR) and a data indicative of the roof category (CR) of each roof (R).
4.- A method according to any one of claims 1 to 3, wherein the trained model (M’) associates a confidence metric to each determined roof category (CR).
5.- A method according to any one of claims 1 to 4, wherein the phase of building the training database (DB) comprises the following steps :
• collecting input data comprising: input images (IM) seen from the sky of areas comprising buildings having roofs (R), and a roof mask (MR) for each building of each input image (IM),
• cutting out the input images (IM) around each building, on the basis of the roof mask(s) (M ) of the input images (IM), so as to obtain an image for each building, each obtained image forming a building image (IMB),
• labelling each building image (IMB) with a roof category (CR), the roof category (CR) being selected among a predetermined list of categories, and
• building the training database (DB) by associating each building image (IMB) with the corresponding roof mask (MR) and the obtained roof category (CR).
6.- A method according to any one of claims 1 to 5, wherein the roof category (CR) for each building image (IMB) is obtained using a labelling tool (TL), the labelling tool (TL) being configured to display the building image (IMB) on a display as well as the predetermined list of categories and to acquire the roof category (CR) selected by a user among the predetermined list of categories.
7.- A method according to any one of claims 1 to 6, wherein the predetermined list of categories comprises at least the following categories: concrete-gravel flat roof, steel roof, roof made in fiber cement without asbestos, roof made in fiber cement with asbestos, tiles roof, and flat roofs.
8.- A method according to any one of claims 1 to 7, wherein the method comprises a phase of determining an arrangement (A) of photovoltaic panels on each roof (R) imaged on the image of interest (IMi), the determination phase comprising:
- a step of detecting obstacles (O) on each roof (R) of the image of interest (IMi) by processing the image of interest (IMi) using the corresponding roof mask(s) (MR), to obtain an obstacle mask (Mo) for each roof (R) of the image of interest (IMi), and
- a step of determining an arrangement (A) of photovoltaic panels on each roof (R) of the image of interest (IMi) as a function of the corresponding obstacle mask (Mo) so that each photovoltaic panel covers a portion of the roof (R) that is free of obstacles (O).
9.- A method according to claim 8, wherein the phase of determining an arrangement (A) of photovoltaic panels comprises displaying, on a display, the image of interest (IMi) on which is superimposed the determined arrangement (A) of photovoltaic panels.
10.- A method according to claim 8 or 9, wherein the method comprises a phase of determining solar data relative to the arrangement (A) of photovoltaic panels determined for each roof (R) of the image of interest ( IMi).
1 1.- A method according to any one of claims 1 to 10, wherein the building images (IMB) and the images of interest (IMi) are images in the visible range acquired by a camera mounted on an aircraft.
12.- A method according to any one of claims 1 to 11 , wherein the detection and classification model (M) is a convolutional neural network.
13.- A method according to any one of claims 1 to 12, wherein the detection and classification model (M) has been pretrained on a database different from the training database (DB) so as to achieve transfer learning.
14.- A computer program product comprising a readable information carrier having stored thereon a computer program comprising program instructions, the computer program being loadable onto a data processing unit and causing a method according to any one of claims 1 to 13 to be carried out when the computer program is carried out on the data processing unit.
15.- A readable information carrier on which a computer program product according to claim 14 is stored.
EP23804979.5A 2022-11-08 2023-11-08 A method for determining installation constraints for the installation of photovoltaic panels on a roof of a building Pending EP4616317A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
EP22290058.1A EP4369237A1 (en) 2022-11-08 2022-11-08 A method for determining installation constraints for the installation of photovoltaic panels on a roof of a building
PCT/EP2023/081207 WO2024100152A1 (en) 2022-11-08 2023-11-08 A method for determining installation constraints for the installation of photovoltaic panels on a roof of a building

Publications (1)

Publication Number Publication Date
EP4616317A1 true EP4616317A1 (en) 2025-09-17

Family

ID=84463171

Family Applications (2)

Application Number Title Priority Date Filing Date
EP22290058.1A Withdrawn EP4369237A1 (en) 2022-11-08 2022-11-08 A method for determining installation constraints for the installation of photovoltaic panels on a roof of a building
EP23804979.5A Pending EP4616317A1 (en) 2022-11-08 2023-11-08 A method for determining installation constraints for the installation of photovoltaic panels on a roof of a building

Family Applications Before (1)

Application Number Title Priority Date Filing Date
EP22290058.1A Withdrawn EP4369237A1 (en) 2022-11-08 2022-11-08 A method for determining installation constraints for the installation of photovoltaic panels on a roof of a building

Country Status (2)

Country Link
EP (2) EP4369237A1 (en)
WO (1) WO2024100152A1 (en)

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10754999B1 (en) * 2019-04-05 2020-08-25 Sunpower Corporation Generating a photovoltaic system design for a building
CN111191500A (en) * 2019-11-12 2020-05-22 广东融合通信股份有限公司 Photovoltaic roof resource identification method based on deep learning image segmentation
WO2021144273A1 (en) * 2020-01-13 2021-07-22 Markus Gisler Photovoltaic system creation

Also Published As

Publication number Publication date
WO2024100152A1 (en) 2024-05-16
EP4369237A1 (en) 2024-05-15

Similar Documents

Publication Publication Date Title
AU2021250975B2 (en) Statistical point pattern matching technique
EP3629282B1 (en) Photovoltaic power generation facility design assistance device, design assistance method, design assistance program, and design assistance learned model creation device
AU2025230742A1 (en) System and method for detecting features in aerial images using disparity mapping and segmentation techniques
CN118887767B (en) Forest fire early warning method, system and equipment based on satellite Internet of things
US7551783B2 (en) Methods, functional data, and systems for image feature translation
EP2003601A2 (en) Systems and methods for fingerprint matching
Wakabayashi et al. Flooded rice paddy detection using sentinel-1 and planetscope data: a case study of the 2018 spring flood in West Java, Indonesia
CN112349150A (en) Video acquisition method and system for airport flight guarantee time node
US20250015749A1 (en) Method for determining an arrangement of photovoltaic panels on areas of a territory
EP4369237A1 (en) A method for determining installation constraints for the installation of photovoltaic panels on a roof of a building
JP5523914B2 (en) Power generation amount prediction system and program, power sale amount prediction system and program
Divine et al. Regional melt-pond fraction and albedo of thin Arctic first-year drift ice in late summer
CN115077489B (en) Unmanned aerial vehicle oblique photography earthwork volume calculation method
US12449274B2 (en) Method and apparatus for generating map data, and non-transitory computer-readable storage medium
CN118898793B (en) A landmark rapid positioning and recognition method and system based on drone vision technology
CN110415130A (en) Agricultural insurance Claims Resolution method, apparatus, equipment and computer readable storage medium
CN118482672B (en) InSAR deformation intelligent monitoring method, system and storage medium for focusing nuclear power region
JP6829753B1 (en) Trained model generator and program
CN114662299B (en) Space environment incentive graph evaluation method and system for spacecraft orbit-related affairs
EP4229550B1 (en) A method for determining the solar distribution in an area
JP7863551B2 (en) A method for determining the distribution of sunlight in a region.
CN121259621B (en) A Method and System for Automatic Identification of Non-Deployable Areas in Wind Farms Based on AI Visual Recognition
CN121640360A (en) An online safety monitoring system for wind turbine blades based on image recognition
JP2025097152A (en) AI model learning data generation device, AI model learning data generation method, AI model learning data generation system, program, and data structure of AI model learning data
Ferchichi et al. Using evidence theory in land cover change prediction to model imperfection propagation with correlated inputs parameters

Legal Events

Date Code Title Description
STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: UNKNOWN

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE

PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE

17P Request for examination filed

Effective date: 20250604

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC ME MK MT NL NO PL PT RO RS SE SI SK SM TR

DAV Request for validation of the european patent (deleted)
DAX Request for extension of the european patent (deleted)
STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: EXAMINATION IS IN PROGRESS