Embedded Files
In the spring semester of 2025, computer science students Sanjaya Basyal, Prayesh Pandey and Pallav Regmi from the University of New Mexico set out to develop a software package that could take thermal and visual drone images from solar array inspections, classify those images and generate a .kml file allowing the user to plot a map on Google Earth Pro(TM) showing the location and type of each image captured.
This group, called The QuadSquad, made it much easier to take a folder of drone images and allow the user to quickly identify where a given defect photo was taken across a field of thousands of solar PV modules. To access this tool, visit https://7quadsquad7.pythonanywhere.com/
It is the responsibility of the drone operator to save their images in .jpg format that includes GPS coordinates in the file header. Right click on your file, select Properties, then Details and scroll down to find the GPS section with Latitude, Longitude and Altitude.
Because PythonAnywhere limits file uploads to 100MB, the user has a optional two-step upload where a second batch can be uploaded if the initial group of files is larger than 100MB.
Once the images are uploaded, the user can then classify each image using the drop down lists on the right side of the screen.
Once image classification has been completed, hit the Update Classifications & Generate KML button.
The software application will regenerate a .csv and .kml file with the user's classifications. *Note that the drop down list will reflect "1-No issue - Normal image" but the .csv and .kml files have stored your updates.) Once this is complete, download both files to your computer.
Open the .csv file to see the raw data including image file name, latitude and longitude coordinates, altitude in meters and your classification for each image. Don't convert large numbers to scientific notation.
Open the .kml file and this will launch Google Earth Pro navigating to your solar array.
All images of normal/good PV modules will have a green star icon. All issues identified will have icons and colors according to the legend below. Classes 1 through 12 are for thermal IR images while 13 through 25 are for visible wavelength images.
To reach out to the software developers, contact them on LinkedIn:
https://www.linkedin.com/in/sanjaya-basyal-56953427a/
Page updated
Report abuse