- A simple model to detect the landing zone for drone.
- My team's term-end project for the course Image Processing INT3404 2 at UET.
- Members:
-
Clone the repository:
git clone https://github.com/huythai855/drone-landing-detector
-
Get into the folder:
cd drone-landing-detector -
Install dependencies:
pip install -r requirements.txt
-
Copy folder
test_imagesand filelabels.csvto the root folder of the project (as the example).⚠️ PLEASE NOTE that the folder structure should be:(root) │ ├── README.md ├── requirements.txt │ ├── models │ ├── trained │ | └── best.pt <- The trained model │ └── common.py │ └── experimental.py │ └── yolo.py │ ├── segment │ └── predict.py │ ├── utils/... │ ├── test_images <- Images folder │ └── img_example_001.jpg │ └── img_example_002.jpg │ ├── labels.csv <- Label of the images │ ├── detect.py ├── hubconf.py ├── trained.py ├── main.py ├── landing_detector.py <- Load trained model and predict ├── test.py <- Scoring file │ └── yolov5s.pt <- The pretrained model -
Run the scoring program:
python python test.py test_images labels.csv
- Demo result's scoring:
Python 3.10.9 (main, Mar 1 2023, 18:23:06) [GCC 11.2.0] on linux Run time in: 0.00 s Total test images: 5 filename: img_train_593.jpg Fusing layers... Adding AutoShape... {'x1': 239, 'y1': 300, 'x2': 381, 'y2': 420} filename: img_train_451.jpg Fusing layers... Adding AutoShape... {'x1': 330, 'y1': 370, 'x2': 424, 'y2': 478} filename: img_train_330.jpg Fusing layers... Adding AutoShape... {'x1': 284, 'y1': 238, 'x2': 370, 'y2': 264} filename: img_train_156.jpg Fusing layers... Adding AutoShape... {'x1': 256, 'y1': 270, 'x2': 508, 'y2': 414} filename: img_train_500.jpg Fusing layers... Adding AutoShape... {'x1': 212, 'y1': 282, 'x2': 242, 'y2': 382} [0.9501630181648812, 0.9255329318420801, 0.8578161822466615, 0.9271523178807947, 0.8490566037735849] Map score: 0.860000 Run time: 1.7516562938690186