Solar Irradiance Mapping From 2D Aerial Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current tools for determining solar distribution require costly three-dimensional data, which are not available for many parts of the world, making them inaccessible for widespread use.
Innovation Solution
A method using a computer-based approach that collects two-dimensional images from the sky, trains a model on a database of these images with associated irradiance data, and applies it to estimate solar distribution directly from two-dimensional images without needing three-dimensional data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If three-dimensional data are used to determine solar distribution, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent uses two-dimensional aerial images as simplified copies of the actual area instead of requiring complex three-dimensional data. The trained model learns to map 2D image features directly to solar distribution patterns, avoiding the need for expensive and complex 3D reconstructions while maintaining sufficient measurement precision for solar potential assessment
Solution Approach 2:
The patent replaces the mechanical/geometric approach of 3D data processing with a machine learning-based system. Instead of physically reconstructing three-dimensional space from multiple views or LiDAR data, the system uses a trained neural network to infer solar distribution directly from 2D images, substituting computational mechanics with statistical learning
2Measurement precision
If three-dimensional data are used to determine solar distribution, then measurement precision is improved, but ease of operation deteriorates
Solution Approach 1:
The system uses readily available two-dimensional aerial images from satellites or aerial photography as input, which are easily obtainable for most locations worldwide. This copying approach eliminates the need for users to acquire or process complex three-dimensional data, dramatically improving ease of operation while the trained model maintains sufficient precision for solar potential assessment
Solution Approach 2:
The trained model automatically processes the 2D images and generates solar distribution maps without requiring user intervention for 3D reconstruction or complex data preparation. The system performs self-service by handling all processing steps internally, making the tool accessible to users regardless of their technical expertise
3Measurement precision
If three-dimensional reconstruction is performed, then solar distribution accuracy is improved, but loss of time increases
Solution Approach 1:
The system performs preliminary action by training the model offline on a large dataset of 2D images with known solar distribution patterns. Once trained, the model can rapidly predict solar distribution for new areas without requiring time-consuming 3D reconstruction processes, significantly reducing the time loss for each new assessment while maintaining accuracy
Solution Approach 2:
By using 2D images as input copies instead of performing full 3D reconstructions, the system eliminates the time-consuming steps of multi-view geometry processing, point cloud generation, and mesh construction. The copied 2D representation is processed directly by the trained model, reducing processing time from hours or days to seconds or minutes
Data Source
AI summary
The present invention concerns a method for determining the solar distribution in an area, the including, a phase for collecting data to form a training database, a phase for training a model on the basis of the training database to obtain a trained model, the input of the trained model being an image of an area seen from the sky and the output being a global cartography of the irradiance projected on each surface of the area imaged on the input image, and a phase for operating the trained model. The phase for operating the trained models includes, a step of receiving an image of an area seen from the sky, and a step of determining by the trained model a global cartography of the irradiance projected on each surface of the area imaged on the received image.