Semantic mapping of lidar data using machine learning methods for autonomous vehicles
Budget: €12 – €18 EUR
I am looking for a freelancer to develop a solution to a challenging problem in autonomous vehicle research. The task is to create a semantic mapping of lidar data using machine learning methods for autonomous vehicles. The data I am working with is point cloud data, and I prefer to use supervised learning to accomplish the goal. The main objective of the project is to create a road mapping solution. With this mapping, autonomous cars would be able to map their environment and create disambiguated maps more accurately. The technology developed should successfully capture complex motions of dynamic objects on roads and accurately relay them to the autonomous car.
The tasks include:
1. semantic annotation of LiDAR scans of an Ouster OS1-128 (Point Label Tool- CloudCompare )
2. fine-tuning a model for semantic segmentation on the new data (Detectron2 (for example)Open Source Bib for object detection and instance segmentation - from Facebook AI Research based on PyTorch)
3. application of the model to further LiDAR scans
4. generation of semantic maps with SLAM & Octomap
The ideal freelancer should have experience working with lidar point cloud data. Knowledge of different machine learning approaches is also important. Furthermore, they should have familiarity with perception algorithms used to interact with the environment. The right freelancer should apply this knowledge to create a detailed mapping of the environment for autonomous vehicles.
This project entails sophisticated and creative problem solving. The final output should be of a high quality, accurate, and well-structured. It should embody best practices for engineering, performance optimization, and coding. The freelancer should take time and attention to detail when developing an answer.
I am looking forward to hear some offers from you.
The tasks include:
1. semantic annotation of LiDAR scans of an Ouster OS1-128 (Point Label Tool- CloudCompare )
2. fine-tuning a model for semantic segmentation on the new data (Detectron2 (for example)Open Source Bib for object detection and instance segmentation - from Facebook AI Research based on PyTorch)
3. application of the model to further LiDAR scans
4. generation of semantic maps with SLAM & Octomap
The ideal freelancer should have experience working with lidar point cloud data. Knowledge of different machine learning approaches is also important. Furthermore, they should have familiarity with perception algorithms used to interact with the environment. The right freelancer should apply this knowledge to create a detailed mapping of the environment for autonomous vehicles.
This project entails sophisticated and creative problem solving. The final output should be of a high quality, accurate, and well-structured. It should embody best practices for engineering, performance optimization, and coding. The freelancer should take time and attention to detail when developing an answer.
I am looking forward to hear some offers from you.