Authors:

Yifan Sun

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

Current, most 3D semantic segmentation models for autonomous driving are mainly trained on spinning Light Detection And Ranging (LiDAR) data because spinning LiDAR sensors have been one the most popular sensors for autonomous driving vehicles and there is an abundance of spinning LiDAR dataset available to the public. However, spinning LiDAR sensors are costly and requires large amounts of energy to operate. The newly emerged quasi solid-state LiDAR sensors are more cost efficient and require lower amount of energy to operate on autonomous driving vehicles. If we reuse the current models pretrained with spinning LiDAR data on quasi solid-state LiDAR data, its performance is below expectation. Currently there are not enough quasi solid-state LiDAR data to train 3D semantic segmentation deep learning models effectively, and the data pattern for quasi solid-state LiDAR is mostly different from the spinning LiDAR data. This research will study the similarities and differences of the data patterns from spinning and quasi solid-state LiDAR's and develop an algorithm to transform large spinning LiDAR datasets into quasi solid-state LiDAR datasets in order to synthesize enough training data for deep learning models of 3D semantic segmentation. Then a visualization tool will be developed to visualize the training and testing results produced by the deep learning models so that their performances can be evaluated based on comparing the labels of objects in 3D spaces to the truth. After that, part of the newly generated quasi solid-state LiDAR data can be fed into various deep learning models for evaluation, and the best model with high performance and accuracy will be chosen to be trained on the entire dataset. The Labor and Environmental Dimensions of the Oil And Gas Industry in Qatar Zaina Aloudeh Texas A&M University at Qatar

Source:

Texas A&M University / 2023

Topics:

No topics listed

Co-authors:

Yifan Sun