Unmanned aerial vehicles, or drones, are technological tools with diverse application fields whose growth is a reality. For the inspection of solar panels, these drones can transport high-precision and wide-range thermographic cameras, allowing a wider field of view. One of the main issues in the development of aerial inspection is the positioning of the aircraft. The drones generate the positioning due to various systems, mainly the GPS. The information obtained is based on a thermographic image format, where the thermal variations presented by each inspected surface can be appreciated. The amount of data generated by this system is huge, due to the extension of the solar plants and the capacity of the system to capture video and thermograms. The post-processing of the data is an unfeasible question regarding time and cost since it would be necessary to review each of the thermograms manually. The project proposes a data processing system that allows the processing from an online platform. The user only has to upload the data and extract the results automatically, which is challenging for this project’s development.
The online platform employs Python for advanced analytics, mainly RCNN, but the users can use other techniques. The interface with the user is done using PHP, JavaScript, CSS and HTML. Data, results, users and models are stored in MySQL. RabbitMQ is employed to manage the messages. The system can run in Apache or Nginx. Please access it at the following demo link, which uses a fake model to analyse the photovoltaic panels. It provides solutions as the panels detected and the hotspots together with the probability and marked by a box in the photo. The data details are also provided, and all can be downloaded in zip format. You can upload a photo or a set of images located in the same directory, you need just to select this directory. Please pulse the following icon to access the application:
If you wish to collaborate with us, please fill out the following form:
References