PV Panel And Fault Detection

Hi, I need a project about detecting solar PV panels from drone images and creates KML file. After detection of panels if this panels has faults the software from you should detect this faults. Sample images will delivere to you from us from train data. The project is clearly; - Should detect the solar panels from given image (images is .tif file and opencv/tensorlflow can open this file like windows image viewer - this file stores the gps points of the pv panel farm) - Should detect the faults and types of faults . ( There are 8 types of faults and we give you nearly 100 images per faults to train) - Should create the KML file from detecting panels as rectangles. - And finally this program should runnable from commandline. If there will API it is big plus KML file is stores the shapes, polylines data with coordinates. This is basic XML file. After detection of PV panels this is an easy part of project.