Building roof photovoltaic resource evaluation method and system based on laser image fusion

By using laser image fusion technology, combining two-dimensional images with three-dimensional point clouds, roof feature parameters are segmented and calculated, solving the accuracy problem of multi-pane roof photovoltaic resource assessment. This enables rapid and accurate photovoltaic resource assessment, reduces workload and safety hazards, and provides decision support for photovoltaic development.

CN121438093BActive Publication Date: 2026-06-26HOHAI UNIV
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
HOHAI UNIV
Filing Date
2025-10-21
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

Existing technologies struggle to accurately assess rooftop photovoltaic resources, especially on multi-pane roofs. Furthermore, deep learning-based methods lack azimuth and tilt parameters, hindering the improvement of photovoltaic resource assessment quality. Traditional methods are time-consuming and pose safety risks.

Method used

A laser image fusion-based approach is adopted, combining two-dimensional images and three-dimensional point clouds. ResFAUnet is used to segment the outer contour of the building roof, and SAT-PointMamba is used to finely segment the roof surface. The roof feature parameters are calculated through image processing and point cloud processing algorithms, and photovoltaic resources are evaluated using the SHORTWAVE-C model.

Benefits of technology

It enables rapid and accurate photovoltaic resource assessment, reduces workload and safety hazards, improves assessment accuracy, and provides decision support for urban photovoltaic development.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121438093B_ABST
    Figure CN121438093B_ABST
Patent Text Reader

Abstract

The application discloses a kind of building roof photovoltaic resource evaluation method and system based on laser image fusion, belong to photovoltaic artificial intelligence technical field.The method is first by being equipped with laser radar and RGB camera unmanned aerial vehicle obtains the color point cloud data of target area and projects generation roof image dataset;Subsequently, using improved ResFAUnet image semantic segmentation network extracts building roof outer contour, and obtains building roof point cloud by rotating caliper algorithm and three-dimensional projection;Then, SAT-PointMamba point cloud refinement segmentation network of fusion self-attention mechanism is used to extract roof inner contour, and obtain roof panel point cloud;Then, based on random sampling consistency and principal component analysis algorithm, roof slope, orientation and available area and other characteristic parameters are calculated, and photovoltaic module arrangement mode is designed;Finally, combined with SHORTWAVE-C model, solar radiation is calculated, and roof photovoltaic power generation potential is evaluated.The application combines the advantages of two-dimensional image and three-dimensional point cloud, realizes the automatic, high-precision evaluation of building roof photovoltaic resource, and provides effective decision support for urban photovoltaic planning.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • CN118470438A

  • CN119600280A