Detecting Crop Circles in Google Earth Images with Mask R-CNN and YOLOv3

Automatic detection and counting of crop circles in the desert can be great use for large-scale farming as it enables easy timely management land. However, so far, literature remains short relevant contributions this regard. This letter frames problem within a deep learning framework. In particular, accounting their outstanding performance object detection, we investigate Mask R-CNN (Region Based Convolutional Neural Networks) well YOLOv3 (You Only Look Once) models circle desert. order to quantify performance, build dataset from images extracted via Google Earth over area East Oweinat South-Western Desert Egypt. The totals 2511 samples. With small training set relatively large test set, plausible rates were obtained, scoring precision 1 recall about 0.82 0.88 0.94 regarding YOLOv3.