search

Glutamat42 / Ultra-Fast-Lane-Detection

This is part of our master project. The projects goal is to adapt https://github.com/cfzd/Ultra-Fast-Lane-Detection for the model car of our university. Our results (including datasets and models) are available on our website: https://glutamat42.github.io/Ultra-Fast-Lane-Detection/
Other
4 stars 1 forks source link

Carla related query #3

Open k-nayak opened 2 years ago

k-nayak commented 2 years ago

Hello @Glutamat42, i have a doubt on a problem i am working on. I am basically doing the same as your project to detect lanes in Carla in real time. My question is if it is possible to get the map (X,Y) coordinates and get a trajectory of the list of detected lane points for the car to drive on. I would offset the detected lane list to the middle of the lane to keep the car in the lane boundary. But i couldn't find any ways to transform the pixel coordinates to map coordinates yet. Any advice on this case would be very helpful. I am searching the carla forum as well, hoping to find some thing there. Thanks in advance.

Glutamat42 commented 2 years ago

If i understand you correctly you want to convert a position on the camera image (x,y coordinate) to coordinates in the carla map. In the other project related to this one we did it the other way around: https://github.com/Glutamat42/Carla-Lane-Detection-Dataset-Generation Also maybe this issue might help you: https://github.com/cfzd/Ultra-Fast-Lane-Detection/issues/301#issuecomment-1114970734

k-nayak commented 2 years ago

cam

basically, i want to get the XY map coordinates of the detected lane points so the car can use it as a trajectory. Is that the same type of thing you also did? In my case just using the center lane the green points and offsetting it to the center of the lane (this is the idea that i currently have)

Glutamat42 commented 2 years ago

yep, exactly the opposite of what my partners did in the dataset generation project. You should be able to use it as a reference. Also because it wasn't my part of the project i can't help you that much with it. But as i also said in the issue i linked its a relatively common computer vision problem and you will find the formula for that conversion.

k-nayak commented 2 years ago

Thanks i will take a look into it and see how i can use it, if i am not wrong it's the fast_lane_detection.py where the world coordinate is converted to image coordinate. So when we say world coordinate in carla its the map coordinate right? the exact x y position on the map.

Glutamat42 commented 2 years ago

i think so. i don't currently have the time to look it up myself

k-nayak commented 2 years ago

No worries, thanks for the input. I will check how I can do it for my case.