ALPS

Public code repository for paper ALPS

• Linux OS (Ubuntu 14.04 preferred)
• Python 2.x
• OpenCV 2.4.x
• CUDA (if you want to accelerate with GPU)
• selenium for python

1. Specify the detection aera in config/query_area.info. The first GPS is southwest corner and the second GPS is northeast corner.
2. Specify other attributes in config/ files.
3. Run each python scripts one by one. Just use $python xx.py.
4. For step 06 and 08, please use YOLO for object detection. The input images can be found in data/.
5. The intermediate and final results will be stored in data/.

• The image downloading process (05 and 07) sometimes failed. What can I do?

  Google may break the link randomly (maybe due to security concern?). So you should mannually restart the script after updating the parameters in code with most recent downloaded image's filename.

• How to train ALPS to detect more objects?

  See YOLO.

• More questions?

  Leave your comments in 'Issue' and we'll reply ASAP.