https://github.com/haishaw/lanes_detection_objects_fusion

Lanes detection (Deep Learning), Objects detection and Object-Lane fusion
https://github.com/haishaw/lanes_detection_objects_fusion

Last synced: 19 days ago
JSON representation

Lanes detection (Deep Learning), Objects detection and Object-Lane fusion

• Host: GitHub
• URL: https://github.com/haishaw/lanes_detection_objects_fusion
• Owner: HaiShaw
• License: mit
• Created: 2019-11-06T17:03:21.000Z (about 5 years ago)
• Default Branch: master
• Last Pushed: 2023-03-24T22:14:23.000Z (almost 2 years ago)
• Last Synced: 2024-11-30T03:30:30.826Z (26 days ago)
• Language: Python
• Size: 4.29 MB
• Stars: 4
• Watchers: 2
• Forks: 0
• Open Issues: 2
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# lanes_detection_objects_fusion

Lanes detection (Deep Learning), Objects detection and Object-Lane fusion

### Lane detection uses Deep Learning vs. CV methods by default

  - Lanenet DL detection model

  - Use polyfit (^2) by default

  - Curvature, CenterDeviation, Heading Angle Error

### Objects detection uses autoware.ai perception methods by default

  - DetectedObjectArray

  - Object Label, Distance Vector, Velocity Vector (TODO)

### Object-Lane fusion

  - Detected Objects to lanes relative to ego lane

  - Register lane id according to ADAS map and localization (TODO GNSS and Visual Cue)

### Software configuration:

  - Host: Ubuntu 16.04/18.04 LTS

  - Docker CE 19.03 (--gpus=all) or 18.09 (--runtime=nvidia)

  - CUDA: Ubuntu 16.04 (xenial): 10.0, Driver: 410.48

          Ubuntu 18.04 (bionic): 10.1, Driver: 418.87

  - CuDNN: 7.6.x

  - OpenCV: 4.0.0 (3.4+) CUDA enabled

  - TensorFlow: tensorflow-gpu 1.13.1

  - Python: 2.7.16

  - Autoware: 1.12

  - ROS:  melodic

  - CARLA: 0.9.5+

### Performance:

  - i7-8750H CPU @ 2.20GHz, 1 GTX1070 MaxQ, 32GB RAM: 10~15 FPS overall @800x600:

	- Object detections [DLCV '+' LiDAR]

	- Lane detections

	- Lane fitting

	- Lane fusion

### Future Improvement:

  - Photo geometry

  - Radar support 2 DetectedObject

  - Spline, etc. advanced fittings

  - Stereo/Monostereo/OFlow vision

  - Localization: GNSS/Vision/ADAS MAP

  - Add Qualcomm SNPE/DLC

  - TensorRT optimization

### Disclaimer:

  - Work is compiled together with OSS and reinvented code, free of use.

  - Contribution and improvement is encouraged.