YOLOv7 planes object detection
The following example shows how to use the YOLOv7 model for object detection on satellite images.
Dataset
The example is based on the Airbus Aircraft Detection dataset. It provides satellite images with 50 cm/px resolution. Annotation bounding boxes for the planes are provided.
YOLOv7
We built our pipeline based on the YOLOv7 repository and using the cfg/training/yolov7-tiny.yaml
config.
Converting to onnx
When model training is completed, export the model using the command below:
python export.py --weights yolov7-tiny.pt --grid --simplify --img-size 256 256
Example inference
Run QGIS, next add Google Eart map using QuickMapServices
plugin.
Then run our plugin and set parameters like in the screenshot below. You can find the pre-trained onnx model at examples/yolov7_planes_detection_google_earth/model_yolov7_tiny_planes_256_1c.onnx
path. Push the Run button to start processing.
After a few seconds, the results are available:
stats
output layers
predicted mask
predicted mask with Google Earth background